THE DAZZLE FARM HYPOTHESIS


FOREWORD: HOW BLIND PREJUDICE IS FORMED

A year after I found five euros at a bus stop, whenever I ride past it on the other side of a busy road some neurons make me wonder if there could be some money over there again.

I have developed a superstition.

Then some higher order thought tells me it's not worth going over to look, and a bit weird even thinking about that.

Clearly, each repeat of a stimulus associated with that single reward is spoofing my rational self. I tell myself I'm hallucinating. The seeming irrationality of it annoys me. But what if my find was not so unusual?

After all, some statistical probability exists of finding money at bus stops. People get their money out, sometimes in a hurry...

But I never think about it near other bus stops. Only this one. Clearly there is something unreliable about my processing of the data.

I get the evolutionary advantage. But superstitions cause a lot of problems. They act like an addiction and do not seem to require voluntary input.

Withdrawal from superstition is as difficult as coming off antidepressants or cigarettes.

Could the hypothesis "People are more likely to find money at bus stops than other places" be tested? Yes. Is this research likely to happen? No.

You couldn't ask everybody in the world. Some of them don't have bus stops.

So you would test a sample. But the effect of bus stops on finding money could be very small. You'd need a very big sample to reliably detect such an effect.

Will the study actually cost more than you might ultimately make if it confirms scouring bus stops is a worthwhile idea?

In earlier times, if you wanted proof of something, you just asked a priest. Who would have said something like "If it is God's Will you will find money at a wagon stop. Now pay your tithes." Or for more specifics you'd try a fortune teller, who might advise "I will predict lucky bus stops if you cross my palm with silver."

Thanks to the Enlightenment and progress in statistics, methods have changed. The definition of reality has been democratized.

What is there to learn? Here, RFK Jnr. shares how a troubled son encountered ayahuasca, opened up, and began taking the bins out and doing the dishes.
https://x.com/0xQuasark/status/2007133536243839072 [5807]

Commodification and enclosure of the psychedelic experience are live issues and as usual it's the latest arrivals at a party who cause the most trouble.
https://www.sciencedirect.com/science/article/pii/S0955395925003706 [5808]

As a defendant with my own bins, dishes, and legal trouble I've been watching worldwide legal prejudice against cannabis and psychedelics (CaPs) in full strategic disintegration in real time during my own individual process.

My interest in the history of power analysis arose by wanting to know what it was possible for the persons behind these laws to have known or understood about CaPs at the time, during the various iterations of prohibition legislation [3500].

Evidence testing came along with the evidence, the vast majority of both of which unfortunately followed the lawmaking and did not precede it.

This is the story of an attempt at proof, as hallucinated by "anti-drug" warriors in a part of Slovenia famous for its alcoholism!

VILLAGE HEROES

One night in 2020, Ptuj Police were watching some people taking drugs in the park. It was after midnight when their informant Mr T, who they testify lives in the park, emerged and agreed to go to the police station.

Mr T claimed the Defendant had a "laboratory". The Police say they then began noticing a light coming from the Defendant's house and with these observations, plus a claim furnished by Elektro Maribor that he was buying too much electricity, obtained a search warrant for the Defendant's home.

We shall examine the logic flow in this Application. Mr T can be said to have provided from one third to one half of the hypothesis, consisting of "He sold weed because I crept in his back door one night to steal it, and then together with my accomplices sold it."

Maybe Mr T didn't say all that to the Police at that time. But that's what happened. They stole some binoculars too.

However Mr T and his fellow home invaders did not find any laboratories during their visit. The Defendant was asleep.

The Defendant was never socialising in the park, which is not nearby. Mr T and the Defendant had never met until proceedings befell the latter.

After several failed attempts to get him there, Mr T finally turned up to Ptuj Court staggering drunk and confirmed it was their first encounter.

And there, from the witness stand, the Prosecution's star witness tried to throw furniture about, then threatened to smuggle in a ceramic knife to defeat the metal detector and stab the Defendant at a future hearing.

The uncharacteristically designed-sounding item of Mr T's evidence, his death threat was an entirely avoidable prohibition-related crime punishable by up to one year, for which he (ergo the justice system which hired him) had to be quietly let off.

Mr T also boasted about creeping into the Defendant's home with two others to steal, both in Court and to other witnesses. The Defendant has never met these either.

Officer L has claimed in Court that all reports are treated equally. Yet no Police complaint has been pursued.

For the sake of context we have gotten ahead of this story, which otherwise concentrates on how the warrant was obtained.

After considering his statement, either Mr T's burglary, Mr T himself, or his hearsay, were respectively considered too compromising, unconvincing, or uncorroborated by Police as a basis to obtain a search warrant.

Mr T later asserted to another witness that he would receive 3000 euros for this information. He may have sincerely believed this. It is not known how Mr T felt about the people hanging around in his alleged home after midnight.

And, pre-primed by an unknown associate with a made-up story about "selling white powder" motivating him to target the Defendant, Mr T was somehow egged on to embellish his statement with an imaginary visit to the outside of the Defendant's home with another witness, to buy weed.

That witness testifies it never happened.

But, as the Police must have surely realised from Mr T's demeanour, all these carefully orchestrated tales might not stand up to routine skepticism, and could lead back to a finding of entrapment by them. Thus was motivated a mission to construct other evidence, a study of a first floor light.

This lighting story would not rely on the labile informant or his accomplices, who are from the village of Kicar and potential awardees in a corporate neighbourhood fluorosis plus metal poisoning, with bureaucratic overtones [5761].

The Court is invited to adjudicate on the points marked with a *.

ORIGIN OF THE LIGHTING STUDY:

The decision of the Police to try to reinforce their warrant Application with some lighting research evidences an opinion that Mr T's evidence wasn't good enough.*

The only prior connection between the informant and the Defendant was that the former had burgled the latter's house in the company of two others.*

As a basis for the search warrant, "Mr T hears, from someone he wants paying to identify, that the Englishman sells white powder and has a laboratory because we stole his weed" was legally problematical for the non-foreigners.*

The Police are unhelpfully not interested in (or are) that someone.*

We shall see this matches a pattern of turning away from discomforting facts in what followed.

This Defence opinion is that the inadequacy of Mr T's portrait of the burglar as a young drug warrior was the trigger for the lighting research.

Both the Police and Mr T could discern from outside the Defendant's home that one of the upstairs blinds was broken and light came out at the top.*

Being the sort of thing a burglar might take notice of, the matter of the light must have arisen for Mr T before the Police zoomed in on it.*

For what, other than the light, could have inspired his "laboratory" story?*

Would the Police waste time researching lights and electricity if they could rely on Mr T for the warrant?*

Evidently, and evidentially, this mentally malleable issue of the Defendant openly using his light right in the middle of Ptuj without a care in the world originated not with the Police, but burglars - but could have been relayed to either by any passer-by.*

Can it be claimed the Police were trying to protect their source with their lighting research?* As his name ultimately appeared in the warrant, the Defence excludes this as a motivation.

How could these lighting observations corroborate the statement of Mr T, or evidence a crime by themselves?* 

ERRORS IN PTUJ POLICE'S NON-SURVEILLANCE OF THE DEFENDANT'S LIGHT:

As the Police decided to try to build a separate strand we are free to investigate the remaining one half to two thirds of the Police's warrant material, consisting of "He has a bright light on a lot and a big electricity bill" - independently of Mr T's contribution.

The Defence disputes all the contents of the warrant Application and its relevance to the topic of cannabis cultivation, proposing to examine the process as a phenomenon in the arenas of Aristotelian logic, objective scientific enquiry, and power analysis.

These, and not crowd hysterias, are the most successful methods by which the human race has figured out the truth about reality. Or more importantly, in view of the problem of superstition, what is NOT the truth.*

Can the Prosecution offer any reason to abandon objective scientific strategies at the peak of their global sophistication and popularity - assumed to be now?

In general, what is NOT the truth does not interest drug warriors at all.*

Yet a Defendant is entitled to the best available analysis of reality.*

Can a power analysis, typically seen in scientific investigations, be applicable to a surveillance?

Yes, 100%. Why not? It is easily applicable to a bivalent true/false effect like a light - an observable, directly measurable entity somewhat impervious to sophistry.

Drug warriors are steeped in skewed, often delusional beliefs based on verbal rather than scientific thinking, rooted in mob instinct and its kid brother, groupthink.

In one instance, the Court has seen the witness Officer G freeze at the suggestion that alcohol is a "drug", although he thinks "all drugs are the same".

This popular concept of "drugs" flattens all drugs in the illegal set into a thought-terminating cliche, making them "all the same". This is strongly appealing to political stigmatizers and nuance-free thinkers alike.

But as the Court also heard correctly from Officer T, all drugs are not the same. The ZPPPD does not think so either. Not even the all-drugs-are-the-same-named UNODC thinks so. Officer G resolutely volunteered that he "hates drugs".

So whatever effect his hating them had on his clientele, it was disproportionately applied throughout his career thanks to this flattening - because he simply could not separate one category of drug, or user, from another.

Some users have the same problem. Many dumb choices about which to take are entirely caused by this flattening. Once "drugs" are established as a black box, the crowd is ready for more unhelpful adjuncts, like "narcotics", "psychoactive", "clean", "abuser".

The Defence suggests this is too large and important a subject to be reduced to a litany of epithets, without the assistance of which prohibition will quickly fold.

Nutraceutical users are long accustomed to vilification by loaded language, which blocks progress and understanding - and which the Court should repudiate.

Drug warriors may have drunk the Kool-Aid of linguistic determinism as if their living depends on it, but the true mathematics of reality are not affected by semiotics.

Much better than those words, power analysis is devoted to proving what is NOT true.

Power analysis is a tried and tested approach to approximating reality via a study design; unlike those words, its outputs are unaffected by whether it was used or not!*

Rejecting the relevance of power analysis is opting for policing and the law itself to be decided by superstition and not rationality.*


The best laid plans

Once you have some idea of what you want to study, formulation of a hypothesis and power analysis are fundamental first steps in any methodological evaluation of some kind of association, such as broken blinds at night and laboratories or cannabis growing.*

How do we approach this problem fairly? The police officers want a simple answer to a simple question: does this particular light on all the time prove the Defendant is growing cannabis?* The answer they want is yes.*

Yet here they have made their first mistake, by seeking to verify and not to falsify a hypothesis,* and in doing so they are disagreeing with Karl Popper. 

Back to him shortly. At some point, the Police decided to commit to the number of observations they would say verifies their hypothesis,* without knowing what it was.*

Whatever it was, they reasoned, the unknown total was "enough"* - according to the design of the experiment they never designed.*


Finding sample size by reverse engineering a lighting study using an AI conversation

Are we comparing two independent samples, requiring a z-test for two independent proportions, or comparing a single sample proportion with a fixed hypothesised value, using a binomial one sample test?

Ptuj Police did not have a control sample or a fixed hypothetical value.*

Using the first method, Grok very generously crunched darkness data with the appropriate test, type of power analysis, desired Cohen's h, proportion under alternative, and other input parameters to solve for an n = 35.*

In this first iteration time of day did not influence the calculation.

In a second iteration Grok was asked to repeat the calculation assuming the light is imperceptible during daylight hours. Grok thought this was

"Great — this changes the problem fundamentally and makes the statistical modeling much more realistic."*

But the anti-hypothesis idea that people might use lights at night for reasons other than growing weed was bad news for the Police budget:

"Because the [August] dark period is only ~6.3 h long, an 18/6 or 24/0 schedule means the light is almost certainly on whenever it is dark outside.
→ Expected proportion when observable ≈ 1.00*

"With these values the difference is tiny → hundreds or thousands of night-time observations would be needed to prove anything, which is not useful in practice."*

Grok never questioned the "light = laboratory" sub-hypothesis baked into the main hypothesis. But despite ignoring that, came up with a smart alternative strategy - comparing "normal" and "non-normal" light use in a third iteration, switching to a difference from constant binomal one sample, one-tailed test, shown here by G*Power in person.

With a 90% power and 0.9 constant, representing the Defendant's "normal", we find an effect size g of 0.0800001 (see note on G*Power's g in the Resources - it does not calculate h for this test, but it can be) and to achieve a power of 60-77% the Police would need 46 observations, up to 87% 61, up to 93% n=76, while 88 are needed for a power of 94-95%.

With a more "normal normal" of 75% the requirement for power above 80% is reduced to 18. But this 75% does not accurately portray the Defendant's real life profile.

A fourth iteration asked Grok to suppose the light was completely imperceptible during daylight + civil twilight hours. Grok answered:

"If the light is on in ≥ 34 out of 35 observations → statistically overwhelming evidence of artificial lighting far beyond normal household use."

In the fifth iteration the "normal" homeowner has his lights on for 75% of the dark period, the "non-normal" 98%.

Grok was asked to ignore the number of times the light is off.

This elicited a very interesting answer about consecutive hits:

"If you make 37 passes during real darkness and see the light ON every single time (37 out of 37), you have ≥90 % power to reject the null that it's a normal household."*

n rose to 108 consecutive passes by the time the normal light was on 90%.*

As it had turned out judgements about this lifestyle-based variable exert such a profound effect on sample size the fifth iteration asked Grok to factor in night-owls, for whom "a common balanced figure is ~25%" of the population.*

This raised n by 15-20% via a weighted "normal" proportion.*

But people are not the average of all people: the actual class of night-owls would be more accurately represented by the 90%-ers of the fourth iteration with n=108.*

Grok's adventures with G*Power and the problem of night owls in urban areas.
https://x.com/i/grok/share/MQHpEfrl9kCpdIkbM7BDOxWm3 [5711]

Here's G*Power again, calculating with alpha at 0.05 for 90% power.

The Police can achieve almost 100% power if on average, for each "guilty" supporter of the Benedictions, they are happy to risk expending equal resources raiding one "innocent" opponent of them.

You can reduce sample size by accepting more mistakes. In the exact test G*Power indicated 76 observations would be necessary to meet 90% power.

But the two methods show as few as 10-17 random observations are required if you are prepared to accept a false positive 50% of the time, which means rejecting the true null hypothesis and visiting a Benedictions-free household to stamp out illegal health when it is not there:*

Crappy performance is therefore a master tweak* and the Police can p-hack night and day if not held to any standard. This outlandish alpha offers a decent achieved power of 84% with n = 6.

No more tedious lighting investigations! Accepting the risk of incorrectly rejecting the null hypothesis half of the time is not even the same as getting it wrong half the times [5875]. We have reached a point here where words can get a bit tricky.

But with an alpha of 0.5, you might as well simply flip a coin to decide whether to bust or not, with a fighting chance of stopping Benedictions in about 16-25% of households anyway, according to the referendum result.

The Prosecution may agree that the anti-pot minority might modify their views, if an alpha of 0.5 was the benchmark, with paramilitary teams routinely crashing into their allegedly abnormally lit homes.

With lighting ubiquitous and interdiction below 1%, who gets swatted would remain a question of favouritism and factionalism.*

Operational cockup rates are a village policy matter and the Courts should not interfere in how random or successful the Police are required to be.*  

The next G*Power models a scenario where the normal light is on more than the test light. The owner of normal light p1 is a night-owl with a dark time usage of 90%.

But the abnormal light is on only 12 hours a day because the plants are flowering.

Here the graph is used to show the effect of increasing p2 on sample size, and the Court can see that with a normal 18/6 growth cycle (i.e. 0.75) the Police would need a sample of 216.*

Ptuj Police have claimed no such n.*

And shockingly, what could a light actually tell them in this inverted scenario?* As with the other one, nothing.* And even less in winter.*

As the lighting duration of unproven horticulture p2 approaches the lighting duration of unproven normalcy p1, G*Power's patience is exhausted.

In these circumstances, the policeman's lot will not be a happy one: as G*Power reaches its limit, 97282 non-surveillance drive-pasts are needed to achieve 90% power with an alpha of 0.05 when p2 reaches 0.8943.

When p2 = p1 n becomes infinite. When p2 > p1 "you are looking in the wrong direction".

Brain teaser: With p1 night-owling at 0.9 and p2 flowering at 0.5, on which dates should the policeman should be investigating suspiciously unusual amounts of darkness visible through broken blinds at this latitude?*

Oi! Put that light on!
https://x.com/i/grok/share/1fc8e8f334474c9fa77c3e91b63fee3a [5880]

In 2020, Slovenia had no laws about what time you should go to bed.* Residents are not required to turn off all the lights when they retire for the night (or day).*

Is this line of evidence absurd? Not really. After all, Ptuj's judges will be used to hearing Slovenian policemen testifying the Defendant's place is untidy. This is used because the majority of judges are women.

Can there be a decent, responsible Slovenian lighting owner out there today who does not feel betrayed by a nation which cannot judicially order his foes to beddybyes because of a lacuna in the area of land-of-nod laws?

Proof and evidence are two different words and two different things in English, which suggests a big set theory problem in Slovenia's jurisdiction.

Because the Ptuj Police inserted superstition and bluff in place of a rational Application, which the investigating judge neglected to bring into check, the trial should stop. But let's plough on and find out what this proof business is about...


Types of error

In "Sample size, power and effect size revisited: simplified and practical approaches in pre-clinical, clinical and laboratory studies" Ceran Serdar et al (2020) list the "main statistical errors frequently encountered in scientific studies". Ptuj Police have managed to incorporate most of these.*

"Flawed and inadequate hypothesis;

Improper study design;

Lack of adequate control condition/group;

Spectrum bias [N/A];

Overstatement of the analysis results;

Spurious correlations;

Inadequate sample size;

Circular analysis (creating bias by selecting the properties of the data retrospectively);

Utilization of inappropriate statistical studies and fallacious bending of the analyses [N/A - they have no statistics and didn't do any analyses];

p-hacking (i.e. addition of new covariates post hoc to make P values significant) [N/A - no significance chosen; all confounders were ignored];

Excessive interpretation of limited or insignificant results (subjectivism);

Confusion (intentionally or not) of correlations, relationships, and causations;

Faulty multiple regression models [N/A];

Confusion between P value and clinical significance [N/A]; and

Inappropriate presentation of the results and effects (erroneous tables, graphics, and figures)." [N/A - no data in evidence]
https://pmc.ncbi.nlm.nih.gov/articles/PMC7745163/ [5806]

The Prosecution may well claim that laboratory standards are not applicable to Police work. However the above errors and ways of avoiding them remain real whether or not they are known about or used.*

For convenience, the Defence summarizes the errors in Ptuj Police's study of the Defendant's light broadly chronologically - into malicious purpose, fallacious hypothesis, lack of design and power analysis, vague methodology and metrics, non-collection of the data, and unattainable interpretation.


MALICIOUS PURPOSE:

Armed with the bulletproof assurances of a burglar who allegedly lives in the park, presumably dependent on what lighting he can find, what is the question the Ptuj Police are trying to answer, in the course of their night patrols, about laboratories and lights?*

Their mission became a one-sided* pursuit of as many unusual features about this light as they could big up: its odd timing,* its extended duration,* its intensity,* its high colour temperature,* its "noticeability"*. Whatever.

This "separate strand" of Ptuj Police's investigation was not so separate at all. Prejudice against this particular light was all they had to tunnel in upon. This cheaply and informally allowed their suspicions and superstitions free rein to blossom into a solution to their fantasies.*

As if to prove this, their results are couched in elastic verbal descriptions, prose adequate for unquantifiable subject matter, not the mathematical terms suited to the measuring of physical processes.*

Watch out for animism: it is in no way the light which is "suspicious".*

In the officers' paranoid conceptualization of their mission, the Defendant's light did not need to be compared with anything except itself.*

A fundamental oddity of their study is that it does not seek to measure an association between lighting and cannabis growing.*

That association is assumed to have a Pearson's r of 1 baked into the hypothesis,* and cannot be tested by it.* Assume makes an ass of u and me.

What then, is being measured? How should the Police measure the criminality of the lighting of foreigners with broken blinds who report the theft of their binoculars by Slovenian burglars to the Police to help solve an arson?

A second round of burglars had been less lucky, and took it out on the neighbour's outbuilding, as well as a dumpster in Ptuj town centre.*

What is the technique? The officers were trying to cook up a narrative that would look more technical, objective, sciencey and immutable than Mr T.*

Offering some clues, a video on experimental design, determining sample size, power analysis equations, and using R code, with some binary outcome examples.
https://www.youtube.com/watch?v=yd-ILiF1PS0 [5794]

Because the Ptuj Police had a one-sided "guilty-by-default" view and did not do any experimental design, the lessons and formulas from the video quickly become impossible to apply to what they did. For this faulty framing, the trial should stop. But let's press ahead with their proof step by step.

FALLACIOUS HYPOTHESIS:

What is the Police's hypothesis? For sure, there was no formal hypothesis.*

Good research practice demands that the hypothesis precedes the experiment.*

Testing a non-existent hypothesis isn't easy.*

Working without a prior hypothesis opens the door to the creation of false positives via p-hacking.*

Hypothesizing After Results are Known (HARKing) is almost inevitable.*

p-values become uninterpretable.*

What are they comparing with what? Outputs dependent upon views about normal lighting demand input values.*

For disregarding all of this the Court should dismiss the warrant and end the proceedings. Post hoc, we must make what we can of the Prosecution's informal hypothesis, which the Defence now attempts to reconstruct from the known circumstances.

This informal post hoc hypothesis of the Slovenian Police is something like:

"Due to a broken blind, whenever we drove past the house belonging to an Englishman constructed by persons we are not interested in as the enemy of Mr T, who we will say lives in the park, we saw a light like we see in the courtroom. And because we noticed it a number of times which we didn't count and can't remember, it was so unusual it independently verified Mr T's information about the Englishman having a laboratory, which we received after he successfully identified and reported Mr T to the Police for burglarising his home, which we don't care about. And his electricity bill proves we are right."

The Defence baptizes this the Dazzle Farm Hypothesis (DFH).

The Prosecution implies an unusual incidence of "light on" signals were determined by the Police, without comparison of this foreigner and his broken blind with a global sample of visible and non-visible lights.*

No global sample existed of comparable people with broken blinds, nor of the light-tight households of those who can safely be presumed innocent.*

Also not represented were the supposedly lighting-dependent electricity bills of farmers with window shutters or no windows.*

Consequently the investigation was unable to make such comparisons.*

Some number of households, unlike the Defendant's, may have blinds that aren't broken - and thus illuminate their homes without Police knowledge.*

The Police position is woolly-minded. We need a clear picture of what they believed, and why.

Is the exact hypothesis that "lights mean laboratories", or "broken blinds mean cannabis", or both?

Both together mean 100% observability, while without one or the other, 0% detection could be made. Both are equal contributors to the outside Effect to which the Police are alluding.* We can present the results in a matrix:



                                Light Off            Light On

    Blind is lightproof         No light             No light

    Blind not lightproof        No Light              LIGHT



There's nowhere to put cannabis or a laboratory in this table. In logic terms, the blind is half of the AND gate "light on + blind not lightproof", the only pair that permits the signal.*

In other words, were the Police testing for a lighting condition, or a blind condition? Or only for both?

The answer was decided afterwards in their claimed results, all of which inescapably required both. The Defendant, the Police may say, is simply unlucky to have both a broken blind and to have reported a burglary to help the fire investigators. They imply these have nothing to do with their observations.* Is that true?*

Objectively, they detected the Defendant's broken blind with as much verity as his light. They are doing as much blind policing as light policing.*

And yet the Police make no mention at all of this bothersome input, the blind, in their warrant.*

In Grok's estimation, 3.03% to 11.11% of blinds could be out of action at any time. So it is not so uncommon, and should be something the Police can live with.*
https://x.com/i/grok/share/c2fa9000d74d4a39b978071edea92c30 [5892]

It's obvious it's broken...or what are they looking at? A deliberate "blind eye" to the obvious evidences their intentional disingenuousness, and a mens rea in their actions to the detriment to the Defendant.*

The Application did not disclose a connection between the shady world of broken blinds and underworld goings on.* It is the first demonstration of their positivity bias, also known as the Pollyanna Principle:

"The name derives from the 1913 novel Pollyanna by Eleanor H. Porter describing a girl who plays the 'glad game' — trying to find something to be glad about in every situation."
https://en.wikipedia.org/wiki/Pollyanna_principle [5783]

Was there scotomization? Which is "the mental blocking of unwanted perceptions, analogous to the visual blindness of an actual scotoma."

Here "no light" is an unpleasant absence of a stimulus, given the Police's ignorant and misguided missionary position against the Benedictions.

Ignoring null signals is a second opportunity for pollyannaism.*

What measures could the experiment have included to prevent detection bias from null signals going unnoticed?
https://en.wikipedia.org/wiki/Scotomization [5784]

Greenland et al note "a serious problem of defining the scope of a model, in that it should allow not only for a good representation of the observed data but also of hypothetical alternative data that might have been observed." [5706]

Nothing can stop people correlating ice cream sales with shark attacks, or broken blinds with laboratories. But is it a valid question?* What percentage of lights are laboratories?* For more superstitions from Slovenia and the UK see [5690].

Before considering power, significance, sample size, or effect size, it is desirable to show some reasoning why the dependent variable could be dependent. It is for the Prosecution to show how ≈100% of farms or labs are betrayed by a broken blind and a light,* as well as how the illumination of illegal farms and labs is distinguished from the legal,* and good luck with that.

Google offers the following advice on bad hypotheses. Commentary on the Prosecution's examples follows each point, illustrated or in square brackets:

"Bad hypotheses in research are typically untestable (vague, opinion-based, not falsifiable), biased/subjective, lack clearly defined variables, are too broad/narrow, or simply state a fact rather than a relationship, making them useless for scientific inquiry, unlike good ones which are specific, measurable, and predict a relationship between variables."

"Non-Falsifiable/Untestable Hypotheses:
Cannot be proven wrong or right through observation or experiment.
Example: "Garlic repels vampires" (No way to test for vampires).
Example: "It doesn't matter if you do your homework or not" (Lacks measurable variables)."

[Broken blinds and strength and/or colour of lights at night predict cannabis farming]*

"Vague or Subjective Hypotheses:
Uses unclear terms or expresses personal beliefs rather than objective relationships.
Example: "Chocolate is better than vanilla" (No defined variables for "better")."

[A few no-tech observations will "prove" what we wanted to "prove" beforehand with a story too implausible to use on its own]*

"Too Broad or Too Narrow Hypotheses:
Too general to be tested effectively or too narrow to be meaningful."

[Too broad: see examples 1 and 4 at [5809] for "light = lab", and example 10 for post hoc reasoning and escape clauses; applicable quick rules of thumb for spotting these: "Can explain any possible observation equally well" and "Contains open-ended 'except when' clauses that can be invoked forever" and "Makes the absence of evidence count as positive evidence" and "Pushes all disconfirming evidence into an unfalsifiable metaphysical realm" [5809]].

[Too narrow due to framing error*, chance error*, and response error*, absence of a control*, untested confounding variables: e.g. "they apply to such a tiny, idiosyncratic slice of reality that even if true, they don't explain or predict much"* and "have very little theoretical or practical importance"* - see examples 1-10 at [5809]..."Is it those foreigners who do not speak Slovene, don't like the Town Smell, live in a bigger place than us in the centre with a broken blind, who report homeless burglars we are paying but whose crimes we are ignoring, that have their cannabis lights on at night for all the world to see?"*]

"Lacks clearly defined Variables:
Failure to identify the independent and dependent variables clearly.
Example: "Studying affects test scores" (Needs to specify how studying, e.g., hours studied, affects scores)."

[Intensity, colour temperature and illusions; the hyperbolic "laboratory" is a fake dependent variable,* a Big Lie technique proposing an all-in-one concept: "dazzling light"; the blind is an AND gate* - had it been intact would the Police have rejected the DFH? What would the Police have done? The logic says

(broken blind + light we can see when it's on) x times we drive past at night = laboratory

∴ no broken blind = no laboratory;*

There was no attempt to model for vampires, night-owls etc.*]

"Biased Language:
Loaded words sway the expected outcome, making it non-objective.
Example: "The amazing new drug will cure the disease" (Bias is built-in)."

["This terrible drug will make everybody ill"; this terrible drug and not perigestational metal poisoning and hooch is causing burglaries and arson by the Police's employee and his jealous drunk coked-up associates*; lowbrow conflatory narrative about "drugs" as a shorthand for criminality and as a justification to harass minorities*; subjective judgement passed down through folk myths vs [852]*; thought-terminating cliches as motivators for "research" into lights*; unreliable weighting of harms attempts to justify the ultimate exploitation of Mr T* and the Defendant in gladiatorial combat for the entertainment and profit of a gawking bourgeousie.]

Does the "study" show a relationship or state a fact?

Via the formally non-existent, informally imprecise Dazzle Farm Hypothesis the Police claim to be proving a fact based on the fact they, a priori, incorrectly claim it proves. Without any facts.*

The only two variables with a known relationship - the blind still being broken and the light still being visible - were not recorded separately.* They ended up described as a single event, baked into a bad hypothesis.

Observations of the signal therefore depended as much upon ignoring one thing as observing the other.* A key variable was ignored.*

Meanwhile officer T (no relation) has admitted (Dec 4 2025 hearing) that all of the 860,000 households in Slovenia (2021 figure) probably used lighting.

From this it follows all the pro- and anti-cannabis legalisation voters in the 2024 referendum have lights, making lighting a non-useful marker of cannabis growing.*

Nothing is reported on their relative incidence of blind malfunction.*

To understand the informal hypothesis, the public would need an explanation of why the lights of Ptuj police station burn all night, and why 6500k lights and open blinds in courtrooms do not mean cannabis.* If all weed has 6500k lights over it, is everything under 6500k lights weed?* Are you weed right now?*

Most would not consider it fruitful to compare the police station's electricity bill with that of the average sized household either.

693,016 valid votes were cast in the 2024 medical cannabis referendum.

It seems there are 2.45 persons per household in Slovenia. With a ±5% error level, 269,394 to 297,751 households took an interest on both sides.

Working from the central estimate, 188,698 households were in favour, while 94,165 sought to inflict their unwanted health advice and garbled morality on the other households via statutory deprivation of the Benedictions [5638].
https://x.com/i/grok/share/UvulkfJclH2VymJDSSLKm5zgF [5803]

There is nothing in the Police evidence to distinguish their lighting apart.*

The Police could never have shown a causal relationship or association between a broken blind and cannabis farming, to any power or significance by these or other observations.*

Crude data from Canada and Uruguay suggest around 1-3% of households grow cannabis, giving a general probability that any random single light per household = farm of ~0.02.

The Defence estimates about 803,987 or 38% of Slovenia's population are males over 25 [5920]. In Canada, 30% of men over 25 reported daily or almost daily use [250]. The mean proportion of days of using cannabis supposedly increased by 0.35% (95% CI, 0.19% to 0.51%) per year (P  < .001), 1.75% over 5 years from 2018-23 [4951]. We hear the rate of home cultivation among Canadian users rose from 4.8% before to 5.2% after legalization [5919].

Of being male, over 25, and using electric lighting, the Defendant is guilty on all counts. This much Ptuj Police could have ascertained from their researches.*

Putting it together, the Court can determine with no difficulty that, applied to this demographic in Slovenia, 5% of 30% of 38% gave Ptuj's Police and burglars an expected 1 in 175 prospect of "success"* at decreasing glycemic control by stamping out inhibition of di-peptidyl peptidase-IV (for example).

Should the Court determine that the suspicions of the Police and investigating judge must not apply only to people with broken blinds, we should also factor in Grok's high estimate of 11.11% for final odds of 0.0057 x 0.1111, which being 0.00063327 reduces their chances of guessing correctly to 1 in 1579.*

But if Slovenians are on top of their game when it comes to blind maintenance, the Police's odds of rescuing GLP-1 drug manufacturers (for example, [5687]) fall to 0.0057 x 0.0303, i.e. 0.00017271 or 1 in 5791.*

The Court may apply Slovenia's own statistics directly. Slovenia's answers would depend upon anonymity. People would not want it spreading far and wide that their blind is broken. In any event the true odds are infinitesimal beyond anything that could be considered prima facie evidence in any other victimless transgression.*

"A bad hypothesis lacks specificity, objectivity, and a clear path for empirical testing, making it impossible to draw meaningful scientific conclusions."*
https://tinyurl.com/badhypotheses [5758]
https://x.com/i/grok/share/5aWnwcLB3ryvnVMHmc7fd5E01 [5809]

As Mendenhall et al say in "Mathematical Statistics With Applications" (2nd ed. 1981):

"It is essential that you grasp the difference between theory and reality. Theories are ideas proposed to explain phenomena in the real world and, as such, are approximations or models for reality. These models, or explanations of reality, are presented in verbal forms in some less quantitative fields and as mathematical relationships in others. Whereas a theory of social change might be expressed verbally in sociology, describing the motion of a vibrating string is presented in a precise mathematical manner in physics. When we choose a mathematical model for a physical process, we hope that the model reflects faithfully, in mathematical terms, the attributes of the physical process so that mathematical methods can be used to arrive at conclusions about the process itself. If we could develop an equation to predict the position of a vibrating string, the quality of the prediction would depend upon how well the equation fit the motion of the string. The process of finding a good equation is not necessarily simple and usually requires several simplifying assumptions (uniform string mass, no air resistance, etc.). The final criterion for deciding whether a model is good is whether it yields good and useful information. The motivation for using mathematical models lies primarily in their utility....The objective of statistics is to make an inference about a population based on information contained in a sample."
https://archive.org/details/mathematicalstat0000mend_s4b3/page/12/mode/2up?q=%22the+objective%22 [5713]


Evidence of the old bill

Perhaps dimly aware their lighting findings were no less underwhelming than what they had before, Ptuj Police tried to beef up their hokum by confirming some dream about the Defendant's electricity bill,* only to repeat the same mistakes - tunnel vision, no room for falsification, blind faith in an unreliable authority, pollyannaism, an irrational hypothesis, ignoring confounders, inadequate power, unknown error bounds, and sampling error - all over again.*

Following the success of "all lights in this study mean labs or farms whichever it is", the Police launched the next strand "all electricity use in this study means lighting, labs or farms whichever it is", fitting their view of the inside of the tunnel.

Of course both lights and electricity usage can mean lots of other things: a long list of alternatives they were determined to ignore.*

Just as all laboratories have lights but not all lights illuminate labs, all lighting uses electricity but not all electricity is for lighting.*

Despite never being computed, "broken blind" is also the ultimate superset for "big electricity bill" which, via the invalid syllogism "all electricity use is lighting" could not have seemed confirmatory of "all lights = what we want them to be" had the blind been intact and the light therefore impossible to see.*

The provider, in its expertise, did not counter these assumptions*. It did nothing to refine the data to provide an accurate and fair comparison with "normal" customers.* Instead it helped the Police in their mission to exaggerate the story Mr T was helped to make up by someone they are not interested in.*

The specifications provided by the Police in their communication with Elektro Maribor incorporated these first two invalid syllogisms into their informal hypothesis which set out to prove (not test): all broken blinds = lights = oversized bill = farm/lab.*

The Police's invalid syllogisms are a third pollyannaism or scotomization.*

The fourth pollyannaism was about the Defendant's living space, which at 125 m2 is 4.2 times the national average. As here all sympathy among Slovenians for the foreign Defendant evaporates, it must be noted for international readers that whiter immigrants are viewed as a a source of limitless wealth and knowledge; while its offendedness is sometimes dressed as a political or moral dogma, the preeminent Slovenian notion of success is yoked to the elimination of another's, regardless of left or right. "Slovenians forgive everything, except success", a rich one explains [5935]. The narrow valleys! Naj soseda krava crkne. It's just a normal sized house, though.

"According to official data from the Statistical Office of the Republic of Slovenia (SURS), the average usable floor space per person in occupied dwellings was 29.6 m2 as of the beginning of 2021 (up from 29.0 m2 in 2018)."
https://x.com/i/grok/share/HWhdSJOyezCnvUEFkMloBG7HD [5793]

An apparently homeless burglar. A light. Electricity bills of families of four.* What do any of these three stories add to any of the others?*

They inspire little confidence in the Prosecution's ability to explain why CaPs are illegal and being pursued by the Police in the first place. But it shows what they do.

DESIGN AND POWER ANALYSIS:

There's no point testing a bad hypothesis, and Ptuj Police didn't.*

But even if they had, would the successful detection of a broken blind have added anything to Mr T's story, upon which the Police already felt unable to rely? Was misrepresenting a broken blind as evidence of guilt necessary for any other reason?*

In the opinion of the Defence, non-testing of the Ptuj Police's bad hypothesis leading to the unjustifiable issue of a search warrant is a metric of the Courts' lack of interest in the measurement of reality generally, and of the benefits of a consciousness upgrade in particular. [5871]

The inability or unwillingness of the investigating judge to challenge the Police application is a mistake and the trial should stop.* Let's see why.

Type I error: rejecting the null hypothesis when it is actually true
Type II error: accepting the null hypothesis when it is actually false
Statistical power: the probability that a significance test will detect an effect that truly exists
https://jamanetwork.com/journals/jama/fullarticle/2503156 [2419]

Power analysis is vital to help frame any test in advance, focus on what exactly is to be tested, to see if the study is worth doing at all, what resources will be needed to produce a meaningful result, and to discourage cheating.


Study power is not p

p is a significance test in terms of the Type I error.

The power of a statistical test is the probability that you correctly detect an effect when it truly exists (1 - β, where β is Type II error).

How does sample size affect power? Is there a rule of thumb?

One hundred times as many subjects or trials only gives you 10x the power.

3.1622776601683793319988935444327^2 = 10

https://x.com/i/grok/share/wS1TcMux1pttPNYl3wb8UmyF0 [5696]


Relationships between the APES


https://www.statisticssolutions.com/components-of-power-analysis/ [5789]


Error and confidence level - a common confusion

People very often conflate "95% confidence" and "95% power"

These are not the same thing!

95% confidence (1 − α) = low Type I error risk (α = 5%)

95% power (1 − β) = very low Type II error risk (β = 5%, quite ambitious)

Grok was asked to "Distinguish error and confidence level in power analysis"
https://x.com/i/grok/share/WbsM8EdeOH7EB9Vkt1v0Vx2fq [5795]


Sample size - a helpful shortcut

Though not a substitute for a thorough power analysis using G*Power, R, a calculator, or just formulae and pen and paper, the "Rule of 400" offers a rough-and-ready way to work out how big a sample you need. [5696]

It is specifically used as a benchmark for achieving a margin of error of approximately ±5% at a 95% confidence level.

For measuring the imaginarily suspicious parameters of lighting owned by foreigners with broken blinds who report their burglaries to Ptuj Police, the required sample size for a ±5% margin of error can be approximated by n=1/B^2 where B is the error bound.

For B=0.05 (5%), the calculation yields n=1/(0.05^2)=1/0.0025=400.

So 400 is a useful level. As for going larger, Google AI reports...

"Diminishing Returns: Beyond 400, increasing the sample size yields significantly smaller improvements in accuracy for the added cost and effort; for example, reducing the error to ±3% requires increasing the sample to roughly 1,000.

"Detecting Small Effects: In broader power analysis, samples in the range of 400 to 800 are typically required to detect small effect sizes (e.g., Cohen's d=0.2 or r=.1) with sufficient statistical power."
https://tinyurl.com/ruleof400 [5803]


How does the Rule of 400 achieve a 95% confidence level?

Grok reveals the Rule of 400 is really the Rule of 384, where Z = 1.96, leaving 2.5% area under the curve (AUC) at each tail.

"Z is the Z-score from the standard normal distribution that corresponds to the desired confidence level....The result is approximately 384, but it's often rounded up to 400 in practice for simplicity and to provide a slight buffer (or approximated using Z ≈ 2, which gives exactly n = 400)."
https://x.com/i/grok/share/4iCGLl2puLkUdyWQfNd6uwPnT [5804]


Two centuries after Hume, and a quarter century after Fisher and Popper...

During 1960–2 Jacob Cohen examined low power in psychology research. His 1962 survey of studies in the Journal of Abnormal and Social Psychology showed average power was only 48% for medium effects, revealing widespread underpowering. 

Cohen was led to conclude that

"...much research is resulting in spuriously 'negative' results. One can only speculate on the number of potentially fruitful lines of investigation which have been abandoned because Type II errors were made, a situation which is substantially remediable by using double or triple the original sample size."

and

"Since power is a direct monotonic function of sample size, it is recommended that investigators use larger sample sizes than they customarily do. It is further recommended that research plans be routinely subjected to power analysis."
https://replicationindex.com/wp-content/uploads/2025/09/Cohen.1962.The_statistical_power_of_abnor.pdf [5756]

Why power analysis must be done before, not after, the study:
https://x.com/i/grok/share/kmH6fdbrfmEIihvs0VSOlhBYF [5705]


Why the hypothesis is whatever the Police and Prosecution want it to have been afterwards

The likely impracticality of assembling and processing the amount of data that would be needed to prove such revolutionary hypotheses as "the broken blinds of people with lights who report burglars mean drugs" demonstrates that humans are evolutionarily biased towards decisions based on low-sample, biased, and superstitious assessments, in day-to-day opinion- and decision-making.*

In contrast the parameters used in power analysis are real. The five components of study power are Statistical Power, Effect Size, Sample Size, Significance Level (Alpha, α), and Variability (or Variance).

With any three you can calculate the others.

You can see what you might achieve by that in the tables at:
https://x.com/i/grok/share/0CpAZR8sfVJfgIg1aBFmMOkon [5786]

Ptuj Police did not specify or calculate any of these parameters.*

Along with a 25-strong list of "What P values, confidence intervals, and power calculations don't tell us" - the last two are fallacies about power analysis - Greenland et al (2016) in "Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations" join others in regarding the artificial dichotomy of "significant" versus "nonsignificant" as "an especially pernicious statistical practice".

This is a problem that might beset observations of lights being switched on at certain times of the day or night by urban nightowls with damaged blinds, and their comparison with lights further from the policeman's car in remote villages whose occupants rise at dawn and bed down at dusk; between those who do or don't believe in the Benedictions and/or do or don't grow cannabis, being lighting users of different ages, with different work or sleep patterns, varying tastes in colour temperature and wattage, from light sources at different distances from the window internally, through different thicknesses of windows and windscreens at different distances with a range of optical qualities, being observations made, using vision of unequal capacity, of the lightings of different demographics, in varying ambient lighting conditions...with the aim of establishing an effect size for marijuana-growing light ownership, including presumably those whose blinds work ok.*

Ptuj Police did not set out to tackle any of that. They didn't worry about any of these variables because they were no use to them. For these reasons the trial should stop. But we can't without mentioning that...

"Despite its shortcomings for interpreting current data, power can be useful for designing studies and for understanding why replication of 'statistical significance' will often fail even under ideal conditions."
https://pmc.ncbi.nlm.nih.gov/articles/PMC4877414/ [5706]

Even with a biased but sciencey-looking title - perhaps "Substance abuse lighting and blind maintenance incompetence of a foreigner in an urban Slovenian patriotic-burglar-reporting context" - many conditions in the Ptuj Police's case report were far from ideal from a statistical standpoint.

Ptuj Police had no planned design, and not even rudimentary data collection.*

The experimenters merely set out to provide a desired "answer" to a fake enquiry, based on sloppy premises that don't make sense.*

With the blessings of Popper and Cohen the Defence says that their pretense of 100% certainty, in what could only be a binary choice, Ptuj Police were guaranteed to get it 100% wrong. They could have achieved a known level of certainty, but as usual they chose not to.

Transitioning from this amotivational syndrome to a hyperactive state, the Police found no farm or lab.

With a busted flush in the 174-1 stakes, no evidential hindsight is required to confirm there was never any reason to suppose otherwise.

Because their investigatory technique opposed the modern philosophy of science the trial should stop.

But let's just ignore that, and dig in on the topic of magical thinking - common in OCD. The drug warriors are too busy obsessing with their words to stop stopping anyone stopping OCD [5871].


What type of test?

In most scientific contexts (especially in psychology, medicine, and social sciences), two-tailed tests are the default and recommended approach unless there is strong theoretical justification for a directional prediction.

Using a two-tailed test with a directional hypothesis would be overly conservative (you lose statistical power), while using a one-tailed test with a non-directional hypothesis would be inappropriate (it inflates Type I error risk by ignoring one possible direction).

In the present case the Prosecution do rely on a very unspecific inference that some combination of the appearance of this light and its repeated presence, during what the Police claim was an unplanned and disorganised series of observations which they needed to bolster the manipulated and malicious allegations of a thief, evidenced the growth of marijuana or the operation of a laboratory to some unstated statistical power and level of significance.*

As it does not, and their thinking is magical, the evidence is worthless and the warrant should never have been granted. Because of this inexplicable design of a flimsy excuse, the trial should stop. But let's continue.

Omitting other variables which would make them look bad, the Effect the Police were pretending to seek - "this particular light = farm" - is a unidirectional effect (according to them) and therefore a one-tailed test would have been more appropriate.*

As mentioned, the relationship cannot in fact be tested because it is baked into the DFH.*

No type of test was chosen, because for the Police their result had already been decided by Mr T from the park, muddying their neutrality as lighting researchers.*

The Police show no evidence of any other component of an a priori power analysis. No significance level was chosen. There was no calculation for sample size. There was no control group or base value. There was no effect size. There were no units of measurement. And no measurements.*

Rather, realising that Mr T's evidence was inadequate to obtain a warrant, and facing a deadlock in the investigation into obesity prevention, the officers convinced themselves how they would, with the assistance of further flim-flam from the electricity company, be able to shove this one over the line.*

All they needed to confirm Mr T's claims, they felt for some reason, was to drive past a few times and tell the judge they noticed a weird light in a window like the ones they have seen many times in the courtroom.

And, sitting under a stronger version of just such a light, the investigating judge in Ptuj did indeed issue a search warrant based on this childish reasoning. Many a child could explain that you can't tell what something is that you can't see, by looking at something else.

A broken blind obviously increases the probability of light the Police can perceive on their patrols from none to some.*

A normal observer from the outside would tell you the light is visible because something's wrong with the blind.*

As with the lighting observations, the irrelevant claims of the utility firm do not define a control sample of "normal" light emissions and usage patterns of non-cultivators and folks without broken blinds.*

Preferences, and various factors like the size of a dwelling, influence the proportion of the electricity used by the occupier on lighting,* further complicating any guess at what is the "normal" amount the latest victim of Mr T's vengeance might be exceeding. Barring strenuous efforts on private property, lighting's part in overall consumption is invisible to us all.*

You might do more sewing in the winter. You might keep the blinds closed in summer. Both increase the lighting requirement. But all else being equal, average lighting use might be expected to increase either side of the summer solstice with a maximum at the winter solstice, around 21 December.*

The Defendant is cost-conscious and adapted to off-peak use. The length of Elektro Maribor's MT period is close to the shortest night (summer solstice) value of 8 h 11m.

This January 2021 electricity bill shows 56% of 2020's usage was between 2200 and 0600 weekdays and all day weekends and holidays.



There was no difference in the ratio before and after the Police's visit: the last three months' ratio of MT:VT in 2020 is 56:44%, the same as the first nine months.* The Police interest in electricity consumption ignored the equal rights of the differently sleep-oriented.*

Adequate power - consisting chiefly of sample size and alpha - is only part of the construction of a good study.

Is the hypothesis falsifiable?* It is possible to pass Ptuj police station every night and see a light on. Its occupants may well have enemies who are criminals.

To deal with the threat of such comparisons, the officers chose a single-subject design.* A methodical approach would have considered what distinguishes these examples, from the viewpoint of the experimental plan.

They did not have one of those: we are held at bay by the Police dogma. How many observations would be required to establish the presence of a weed farm in the Defendant's building, or the police station? According to Karl Popper:

"'In so far as a scientific statement speaks about reality, it must be falsifiable: and in so far as it is not falsifiable, it does not speak about reality."

The criterion is that "statements or systems of statements, in order to be ranked as scientific, must be capable of conflicting with possible, or conceivable observations."

Why rank falsifiability over verifiability? Popper was peripheral to but never participated in the Vienna Circle. In "The logic of scientific discovery" ("Logik der Forschung", 1934, English translation 1959) he diverged from the logical positivists over their "verifiability" criterion: the idea that a statement is meaningful/scientific only if it can be empirically verified:

"But I shall certainly admit a system as empirical or scientific only if it is capable of being tested by experience. These considerations suggest that not the verifiability but the falsifiability of a system is to be taken as a criterion of demarcation. In other words: I shall not require of a scientific system that it shall be capable of being singled out, once and for all, in a positive sense; but I shall require that its logical form shall be such that it can be singled out, by means of empirical tests, in a negative sense: it must be possible for an empirical scientific system to be refuted by experience."
https://archive.org/details/logicofscientifi00popp/page/40/mode/2up?q=1934 [5759]

Popper argued that no number of positive observations can ever conclusively verify a universal theory, due to the problem of induction, as noted by David Hume in 1748. According to whom, we assume that the future will resemble the past (e.g., the laws of nature remain constant).

But this assumption itself is based on past experience. So justifying induction requires an inductive argument. We don't really know the future. Popper concludes that science progresses by falsification, not inductive confirmation.

After seeing millions of white swans and concluding all swans are white, a single black swan can logically falsify that hypothesis. Logic is asymmetric: confirmation is weak, disconfirmation is strong.*

We establish that the Police hypothesis was a single-subject, directional one, in which they would see only white swans, at night.

With no guarantee of recording null observations,* and no other subjects with healthy blinds on good terms with the park's inhabitants,* there were no distractions from the argument the Police set out to verify.*

The Police are fully familiar with the issue of falsification. The suspect's DNA is found and there is a 5 billion to 1 chance that it was not him. Falsification is the 1. Without it DNA identification would not work. Is that what you want?

In the absence of DNA, much identification of unknown perpetrators depends on set theory of one sort or another. It might be sifting through vehicles or communications. We expect the force to be intimate with Boolean logic.

What have we learned? Despite these tentative definitions of their study, the Police's hypothesis and design were very indefinite,* and their execution downright sloppy.* Both were seriously partisan.*

Their methodology is not just pre-Cohen or pre-Fisher but, appropriately enough, pre-Enlightenment.*


The methodology is circular

Ptuj Police want to decide their methodology not before starting observations, to ensure a fair test - but at the end, to suit the answer they wanted to get.*

Another example can be seen in their insistence that they did not organise a surveillance, but just happened to notice the light (but either did not count or did not notice that they were not noticing it when it wasn't on) at times that "proved" what they wanted to believe to an unknown degree of certainty, via a number of observations yet to be announced.

How is the visibility of this particular light and this particular broken blind at certain times credible proof of what the Police were looking for?

The problem is the Ptuj Police disagree with Popper's theory of falsification, and with it the underlying methodologies that have guided all credible data collectors and statisticians since the publication of Logik der Forschung, closely followed by Fisher's "The Design of Experiments", which paired a new philosophy of science with the widespread adoption of p into the algebraic pantheon.

In their scientific heresy the cops were trying to verify, not falsify, their very flexible hypothesis.* Their p was 0 because they saw only a 100% certain outcome.* Indeed, was their study really a question, or only a statement of that outcome?*

This they achieved at the expense of being wrong, because there was no grow op, no laboratory, just two domestic 8w lights.

"But," they may wail, how could we know in advance?"

By designing the experiment, they could have "known" according to parameters chosen in advance.* Since no legal parameters were specified, their acts in respect of the lights were rudderless and ultra vires.*

Officer T was not even curious to see if she was right, because she knew before she got there that her warrant material was 100% fictitious.

Such was her psychological lock, officer T could not even go upstairs to confront the error she knew she was foisting upon the Court.*

The Police had no reason to believe their theory was, for instance, at most 5% fictitious. They did not make 400 observations.*

While not making them, they were able to confuse the animate and inanimate in a mystical "transferance" effect.

The Prosecution's informal hypothesis was malformed by confusion within and between inanimate and animate elements of the study.*


Errors about the inanimate

The Police and the light are not the independent and dependent variables; these are what the light/blind/both and the farm/lab/both are supposed to be.*

But the Police have adduced no evidence of measuring either, or of interaction between them.*

The hypothesis excluded all other possible explanations for the signal being measured, besides the broken blind.*

These alternatives include all other reasons for having lighting: to enable basic visibility and movement, as task lighting for reading, cooking, working, or studying, feature lighting for the emphasis of objects, artwork, architectural features, and to add style, ambiance, or mood to spaces, for safety and security, to prevent accidents and deter crime, for disinfection (ask Dr Trump), as light therapy for mood and sleep disorders, and to light up the endocannabinoid system in a defined anatomical region [5937]*

There are no highly technical issues here. Ptuj Police are as aware as anyone else that not all mammals are bats.

The statement "All A is B, therefore all B is A" is a classic example of a formal logical fallacy, specifically an invalid syllogism.* There are 256 types of syllogism, of which only 24 are valid.
https://en.wikipedia.org/wiki/Syllogism [5769]

"All A are B" means the set of A is entirely contained within the set of B (A ⊆ B).
Incorrect Reversal: The conclusion "All B are A" claims the set B is contained within A (B ⊆ A), which is only true if A and B are the exact same set (A = B).

Just because everything in one group belongs to a larger group doesn't mean everything in the larger group belongs to the smaller one. So not all things under lights are weed even if all weed grows have lights over them* - here's one that doesn't:
https://windsorstar.com/news/local-news/no-light-pollution-no-pricy-hydro-at-sunshiny-outdoor-pot-farm [5781]

"Broken blind/light = farm" is also a faulty generalization,* an informal fallacy wherein a conclusion is drawn about all or many instances of a phenomenon on the basis of one or a few instances of that phenomenon....It is an example of jumping to conclusions.

"Expressed in more precise philosophical language, a fallacy of defective induction is a conclusion that has been made on the basis of weak premises, or one which is not justified by sufficient or unbiased evidence. Unlike fallacies of relevance, in fallacies of defective induction, the premises are related to the conclusions, yet only weakly buttress the conclusions, hence a faulty generalization is produced. The essence of this inductive fallacy lies on the overestimation of an argument based on insufficiently large samples under an implied margin of error."
https://en.wikipedia.org/wiki/Faulty_generalization [5770]

No attempt was made to define "normal" lighting, against which the Defendant's could be adjudged extreme.* No adjustments were made to compensate for darkness, a constantly shifting phenomenon [5856] requiring different solutions in different times of the year and weather conditions.*

Night, day, civil, nautical and astronomical twilight in Ptuj through the year 2020:
https://www.timeanddate.com/sun/@3192241?month=8&year=2020 [5708]

Civil, nautical, and astronomical twilight occur when the centre of the Sun is <=6 degrees, >6 and <=12 degrees, and >12 and <=18 degrees below the horizon, respectively.
https://www.timeanddate.com/astronomy/different-types-twilight.html [5709]


Errors about the animate

The informal hypothesis did not consider the work, health, or economic profile of the experimental cohort (n=1).*

Shift workers, vampires, and other nocturnal syndromes were not excluded from the study sample or grouped separately.*

Some people work nights and habitually sleep in the day. So their lighting schedules do not equate with the officers' idea of normal,* even those who work nights themselves.

With this they confuse animate and inanimate by flagging the target as different. A symbol of "the person" is depersonalised via the thought-terminating cliche "drugs" into "the thing", i.e. a light.

For aggression and childhood trauma to be sublimated, the target must be othered, and his behaviour differentiated from that of the owner of the ensuing saviour complex. Thus under no circumstances can the victim simply be someone who is also up at night.

Some householders, including the Defendant, consider lighting and other appliances part of their heating output. The Police DFH did not weigh this.*

This light did not evidence any particular human activity.* In 2020 Slovenia, we could forget to turn the light off without legal consequences.*

Subjects were not counted into blind-competent and non-blind-competent groups.* The sampling was biased: both cultivators and non-cultivators with working blinds were counted as "light off" because their lights could not be seen and/or were uninteresting to the investigators due to having no enemies in the dark, dark park.*

Now, the imagined measurement of the inanimate thing was "transferred" to the character of the animate Defendant.*

Via this transfer, the Police hypothesis effortlessly asserts that people with self-appointed enemies living in the park, and who do not mend their blinds for some unknown number of nights, surely should not be surprised when officers come knocking about their fancy lights.

What was really being monitored? Perceived light, or the Defendant's property maintenance skills?* Why not both equally? Why was the blind so ignored and yet the Defendant so condemned by his blind...?

PARAMETERS AND METHODOLOGY:

So we have some hypothesis "broken blind and non-grotty lighting = cannabis farm". As with anything, if this were remotely likely, the "proof" would grow exponentially with the quantity of observations.*

For large n, by the law of large numbers (LLN), if the proportion of "light on" samples is 1 (all on), you can conclude with high confidence (e.g., 99%) that the true "on" proportion is > 1 - ε, where ε shrinks as 1/sqrt(n).

"The LLN can be used to optimize sample sizes as well as approximate calculations that could otherwise be troublesome."
https://www.lakeheadu.ca/sites/default/files/uploads/77/images/Sedor%20Kelly.pdf [4536]

The smaller n becomes, the larger the expected error. To define the standard of their proof, the researcher must decide in advance what error they are willing to tolerate.*

At this the non-surveilling officers failed.*

As part of the power analysis process it is customary to set alpha, a value in this case representing the proportion of light-emitting households that Slovenia feels could be raided for labs or farms in error.*

By choosing an alpha of 0.05 we accept a 1 in 20 risk of falsely concluding something extreme (Type I error rate).

Because there was no thought or action aimed at falsifying the informal hypothesis no alpha was set, and because of this excess of logical positivism, the trial should stop.

The Police have not testified to any specific number of observations that would qualify as "large numbers". For which Grok accurately points out there is no such threshold, but for n informally suggests "a few hundred to a few thousand" and "n = 1,000 is already pretty good, and n = 10,000 is usually excellent.".
https://x.com/i/grok/share/TCDIZTiRe8TfFlcI7eHdZUx2H [5792]

But if you really want to suffer, try the Wilson Score Interval for binary data. If your acceptable level of raids performed in error on completely innocent lighting owners was 1 in 20, and you observed a light 19 out of 20 times, your Wilson score interval would be 0.8775 ± 0.1136.
https://www.statskingdom.com/proportion-confidence-interval-calculator.html [5760]

The Wilson Interval shows we are 95% confident (in the frequentist sense) that the true "lights on" and-therefore-somehow "weed farm" proportion lies between 76.4% and 99.1%.

This means the data are consistent with the light being on almost continuously. The proportion could be as high as the 99.1% upper bound.

But equally importantly evidentially, the same data are also consistent with the light being off quite a bit: up to 23.6% of the time, at the lower bound of 0.764.

The Police have set no parameters.* They have not optimized their sample size.* There is no confidence interval for their data because they don't have any.*


Effect size was not known

Officer T testified there was no Cohen's d.

Understanding effect size is vital because a statistically significant result (low p-value) doesn't automatically imply a large or practically important effect, especially with very large sample sizes.

Why especially with large n? This is counterintuitive.

"A low p-value with a very large sample size frequently tells you about precision, not about importance. Always ask for (and look at) the effect size and its confidence interval before deciding whether a result actually matters in the real world."
https://x.com/i/grok/share/WtsbfQtiqCWl0XDJXaC9b8xNs [5819]

You cannot measure an effect size without collecting data.*

Lighting profiles of farms with and without lights and lights with and without farms were not collected and counted.

Without this, the effect "farm" baked into the brick-like hypothesis "light = farm" cannot be measured.*

The electricity bills of those with intact blinds, who just shrug off home invasions and don't bother reporting crime to the Police, were ignored.* Thus, the Effect "broken blind + vengeful burglar = lab" cannot be measured.*

This shows the Police were not really interested in the characteristics of this light disturbing no one in Ptuj village centre.

It was merely all there was available that could be deliberately misinterpreted to target the Defendant.*

There is no Pearson's r showing the degree of linear relationship between lights shining out onto the street and cannabis growth.* There is no significance test for the absent r.*

No base values for lighting with and without farms were established with which to compare the sole experimental subject.*

Which statistical tests should be chosen to measure effect size?

Given this void, the answer depends on which type of measurement the Police will be claiming to make after deciding which fits their post-hoc theory best.

The Defence shows two choices:

If all we are concerned about is a binary answer, i.e. "is any kind of light on or off" Cohen's h is the appropriate measure for effect size, comparing the proportions of light present between the experimental and non-existent control group in a one-tailed z-test.

But the Police have alluded to a particularly attention-grabbing quality of this Defendant's light, to its PERCEIVED brightness, a function of watts power, efficiency, colour temperature, and distance, adjusted to account for varying sensitivity at different spectra of a standard human eye.

If analog quantities (time on and brightness) and/or qualities (spectral distribution, colour temperature - due to metamerism not the same!) are what the Police are trying to express, yes or no cannot be the answer.

Rather, "light off" can be zero on a continuous scale of intensity and/or colour to produce a perceived light in candela. Cohen's d would then be appropriate for the effect size, comparing means in a one-tailed z-test.

The difference between Cohen's d and Cohen's h.
https://x.com/i/grok/share/XCnfrukj5ZIWgJskW7LUrmG9x [5714]

Is Cohen's h for means OR proportions and Cohen's d only for means?
https://x.com/i/grok/share/8H4WBxRphVRMqGNEhVNB2ZVix [5850]


Cohen's d and Cohen's h both standardize variance but in different ways:

"Cohen's d explicitly standardizes by dividing the mean difference by the standard deviation (√variance), setting the denominator's variance to 1 in a standardized metric.

"Cohen's h achieves a similar goal implicitly for proportions: the arcsine transform normalizes the variance, allowing h to represent the effect in a way that's comparable across different baseline proportions and sample sizes. Without it, effect sizes would be confounded by the inherent variance instability of binomial data."
https://x.com/i/grok/share/ph6b2JXkknGxSSb93XSDVS0Qa [5791]


Putative alternative legal design and sampling options

If, perhaps to simplify the problem for experimental design and legal purposes, lights of an intensity and colour beyond specified values were counted as "guilty" when on too often, and only these observations were recorded as on/off, the Police could say they were plotting activity against time and comparing means. This would enable use of Cohen's d.

However in reality there were no lighting laws proscribing high colour temperatures, lumen levels, wattages, or excessive domestic lighting times in Slovenia in 2020.* And the Police didn't do any of that kind of plotting.

Color is a property of perception, not a property of light. To the naked eye, the actual characteristics of grow lighting lie within domestic ranges.* Technical assistance would be essential to exclude metameric effects from irradiated spectra,* obviating any use for superstitious guesses. Claude has made a lovely web page explaining why reliable human detection is a fool's errand [5940] and see the Resources for more on metamerism, why colours don't exist, and the brain doesn't care about colour much.

There was no public outreach by health or law enforcement to advise or warn the public to limit these illumination values, or face the consequences of being set up and having their property stolen.*

Incorporating analogue variants such as lux, colour temperature, or both into this study could necessitate the use of Cohen's d. But Officer T testified on 4 Dec 2025 that no objective assessments were made. Cohen's h is the correct choice for binary-answer problems that compare proportions.

Hers, however, was a single-subject case study - there were no control group or base values against which to compare the Defendant's allegedly errant lighting cycles.* Indeed any other salient light source was deliberately ignored in this rigged experiment.* No effect size is shown.*

We don't really know which of these types of measurement of effect size Ptuj Police want theirs to have been, as once again they want to shape their design as late as possible to fit the rest of their story.

No specific sampling method was chosen

Sampling was not random.* Police were focussed on the target of the story angry burglar Mr T had been told to give them by other operatives. The Police do not drive down every street equally with the Defendant's street.

Broken blind-owning residents of other, perhaps more brightly lit streets may have enemies in the park but they do not count.

Sampling was not systematic.* Testing of the informal hypothesis "enemy in the park + broken blind + light = farm" was not done every nth time they went past, nor at some specific hour. You would expect systematic sampling in any study concerned with timing and duration.

Sampling was not stratified.* There was no control group.

Sampling was not clustered.* The Defendant was not sampled from some number of city blocks, postcode, or GPS area and compared with a sample from others.

There was no random assignment.* It was a single-subject, convenience-sample study,* without a control. [2756]

Sampling methods (p70) and commonly used symbols in parameters and statistics (p72) can be found in Basic and Clinical Biostatistics by Dawson-Saunders and Trapp (1994)

https://archive.org/details/isbn_9780838505427/mode/2up?view=theater [2756]

Most tellingly in the light of the surveillance regulations, the observers' sample of 1 was selected because of - not just for - the observation, guaranteeing a salience and outcome the officers thought should occur.*



Sampling errors can be subdivided into framing error, chance error, and response error. All three types are represented in the Prosecutor's model.*


Sampling vs. continuity

Just because you see a light twice hours apart doesn't prove it is on continuously all the time in between.* Proving otherwise to a known standard would require close observations minutes or seconds apart.*

One or two observations in a night would be the lowest standard of proof: none at all.


Sampling density

In fact the theoretical number of observations required to get a different result is 3. This is known as the coupon collector problem. You are collecting a set of n different coupons, animal stickers, or baseball cards. All their variants are equally prevalent, but you have to buy to find out.

Probabilistically, how many purchases will you need to make to get every different coupon? Here are some more results, with a graphic from Wikipedia:



n = 1: E[T] = 1
n = 2: E[T] = 2 x (1 + 1/2) = 3
n = 6 (dice faces): E[T] ≈ 14.7
n = 50 (U.S. states): E[T] ≈ 224.96
n = 365 (birthdays, for full set): E[T] ≈ 7145

where n = coupons T = trials E[T] = expected value of T ≈ approximately equals

Let T_i be the number of additional trials needed to get a new coupon when you already have exactly i−1 distinct coupons.

T = T₁ + T₂ + ... + Tₙ

Each T_i is geometrically distributed with success probability p_i = (n - (i-1))/n = (n - i + 1)/n.

The expected value of a geometric random variable with success probability p is 1/p.

Thus:

E[T_i] = 1 / p_i = n / (n - i + 1)

Expected Total Trials is

E[T] = E[T₁] + E[T₂] + ... + E[Tₙ] = n/ n + n/(n-1) + n/(n-2) + ... + n/1= n x (1/n + 1/(n-1) + ... + 1/2 + 1/1)= n x Hₙ

where Hₙ is the nth harmonic number: Hₙ = 1 + 1/2 + 1/3 + ... + 1/n

In case you never learned how to add fractions, we live in an age of blind trust. There is a calculator for the coupon collector problem:
https://demonstrations.wolfram.com/CouponCollectorProblem/ [5849]

Though it demonstrates a relationship between sampling density and reliability, the coupon problem is not the ideal analogy. How does it illuminate the light problem?

In the coupon case, sampling density is achieved by repetition, and time is not counted.

In the case of the unruly light, density is only achieved by concentrating the observations in time until the probability of the light being off when unobserved is a certain level of small, relative to the size of the sample.

The Police's actual sampling density fails to support the conclusion reached by the investigating judge.*


Postscript to the methodological analysis: Expert opinion or case study?

The Defence proposes that sampling density and power correspond with the pyramid of evidence proposed by the Canadian Task Force on the Periodic Health Examination in their landmark 1979 report, and later expanded and refined by David Sackett in 1989:

This figure from that report shows Systematic Review of Randomized Controlled Trials at the summit.
https://media.springernature.com/full/springer-static/image/art%3A10.1038%2Fs41587-021-00834-6/MediaObjects/41587_2021_834_Fig1_HTML.png [5766]

Sackett's improvements:
https://x.com/i/grok/share/in6sZldfSzjCbLGHpweBSJ4o7 [5767]:

A Dr Joe Zundell has a simplified version, which adds some useful thoughts but not very good punctuation:



And there are lots of other versions. Due to the lack of a lighting study design the Defence proposes that the Police lighting evidence fails to meet evidentiary standards because it is at best expert opinion or a case study.*

NON-COLLECTION OF THE DATA:

The observations were not counted or timed.* No actual data was recorded.*

The Defence groups the use of elastic verbal descriptions, rather than data, with the specific rejection or avoidance of technical means of light measurement as a fifth pollyannaism.

We baptize the five instances of Police pollyannaism

1) Blind pollyannaism: the Police are blind blind, and cannot see the 50% contribution of the blind to the signal. A major bias passes by unnoticed.* The Defendant has been abnormally criminalized for his blind.*

2) Null pollyannaism: when a patrol passed and there was no signal. No black swans were seen at night. Are they accusing the Defendant of having his light on only when they noticed it?* Do they specify that this light was never off?* Can the Police witnesses guarantee a consecutive number of hits?*

3) Dim pollyannaism: "All A is B, therefore all B is A", invalid syllogism, faulty generalization. Banal, with major loading of the framing in favour of the informal hypothesis.* Plainly inane.*

4) Jealous pollyannaism: having 4.2 times the mean floor area in Slovenia is probably enough grounds for a warrant without any drugs. Ignores other motives for denunciation. Together with central location bias and a park-based enemy, this is a case of convenience sampling not mitigated in any of the eight ways suggested by Grok:*
https://x.com/i/grok/share/zA2zoXJBEGVKomhyYV4rj8RNY [5793]

5) Fuzzy pollyannaism: rejection of tech for objective measurements; words instead of numbers; wilful avoidance of precision and arguable factoids.


How cherry-picking and confirmation bias prevent fair data collection, and how to avoid them

"These fallacies are interconnected and often overlap with cognitive biases like the frequency illusion (perceiving something as more common after noticing it once) or availability heuristic (judging frequency by memorable instances). To avoid them in experiments, one should systematically log both positive and negative outcomes,* use blinded methods,* or apply statistical tests for base rates.*
https://x.com/i/grok/share/4guFNHHr6hzWUu9qIkrtOxIxF [5785]

Being nonexistent, the Police and investigating judge's methodologies neatly sidestepped any such problems by not counting anything.* Superstition took over, in pursuit of a warrant whose issuance was lax and, er, unwarranted.*

Lest we underestimate the power of a flexible construction of a story, the Police are not supposed to be like journalists, picking and choosing a scandal angle which suits them best. Are they?* Today's newspapers are tomorrow's waste paper. 

It will be argued it is the Police's job to be suspicious. Does superstition make people more suspicious, or vice versa? Oracle Grok advises that superstition is a coping mechanism for dealing with missing information:

"When people feel uncertain, powerless, or lack control over events, they are more likely to engage in both superstitious thinking and paranoid/suspicious interpretations of the world. Studies demonstrate that inducing feelings of lost control experimentally leads people to:

"Perceive patterns in random data
See conspiracies in innocent situations
Embrace superstitions as a way to regain a sense of agency"

And AI advises:

"In extreme or rigid forms, it correlates with anxiety, obsessive thinking, and interpersonal distrust.

"It shares features with conditions like schizotypal traits (odd beliefs, magical thinking, ideas of reference) where suspicion of others is prominent.

"Cultural or religious contexts that emphasize superstition can amplify generalized mistrust." [5785]

What could better describe the situation of the Police and the CaPs myths?

Although not strictly a part of this motion, a Defence which must be remembered in a set theory overview is that studies of the lighting problem exist inside further syllogisms such as "all drugs are bad", which depends upon flattening: "all drugs are the same" (upon which the Prosecution witnesses disagree), therefore since "cannabis is a drug", "cannabis is as bad as heroin".

Miscellaneous volatile factors in the lighting environment, and detective tech

Did the Police ever see a light during the day?* How do the time of year and traditional outdoor flowering periods influence their concept of normal lighting?* Is cannabis interdiction, as it seems anecdotally, a seasonal enthusiasm for the Police?* Did the study consider extra lighting might be needed because the blind was broken, being equally as shut during the day as it is open during the night?* How might not considering these influences affect their results?*

The warrant application contained no such information,* and with no record of any results the answers to these questions can never be known.*

Ptuj Police made no objective measures.* There were no light meters or handheld spectroradiometers such as the Sekonic C-800 or C-7000, UPRtek MK350, Gigahertz-Optik MSC15, or Konica Minolta CL-70F, all capable of offering affordable empirical alternatives to superstition as grounds for suspicion, or to "copper's hunches", or even "intuition".
https://x.com/i/grok/share/z7qSd9E8n5hiVaCGzJmrDHvtg [5768]

Moreover, none of those are considered reliable diagnostic methods of collecting data for driving over the alcohol or speed limits.

It is because such conclusions can become habituated, biased, throwaway, and unreliable that we invented the breathalyser, speed radar, and polygraph.

Compare these with the most subjective and conflatory tool in the law enforcement armory: "smell of cannabis", in which the conflated bogeyman word "drugs" has become a superstitious shorthand for "criminal types".

If true, Slovenia is in big trouble, judging by the referendum result.

Via an olfactory stimulus criminality is magically transferred from the inanimate to the animate. You can't prosecute a plant. But with some Pavlovian work you can train a policeman to think "Crime? I can smell it!"

Oddly, Slovenia's normally gadget-loving Police's measurements of the potentially guilty light were non-automated.* Assessments were not blinded.*
https://algorithmwatch.org/en/slovenia-police-face-recognition/ [5863]

The Police have made no effort to standardize a power analysis-type model defining unusual lights in Slovenia.*

This matches a historic pattern of outrageous rule-stretching in areas where there are no rules - here, about what makes a light "suspicious".*

Suspiciousness does not have an SI unit. The Defence is unaware of a formula for suspiciousness. It might have something to do with outliers.* But how can the Police identify outliers, exhibiting too much light or too little, without a median?*

The Police model took no medical, optical or ophthalmological confounding variable, Bouma factor, miosis, mydriasis, apparent brightness due to ambient contrast, lateral inhibition, the Gelb Effect [5864, 5865], or crowding into account.* It did not consider relevant white hole, Asahi, Kinasza, Chevreul, or other illusions.*

Regarding the easily overlooked influence of contrast in the visual field, it is not the Defendant's fault if he lives in an underlit area where his light seems exceptional. Nearby street lights periodically do not work for weeks on end; either something is broken, or the Council is saving money, or not paying its bill.

The Police characterization of what they saw mentions none of these factors.* Somehow their repeated exposure to the same thing was mind-altering, even at its retelling.*

Yet the Defendant's alleged contribution to street lighting has not been sufficiently outstanding to be compensated by the town's burghers, or reported for interfering with air traffic.

Yet again the law has let down the Ptujčani, lacking provisions to deal with foreigners causing too much contrast in Slovenia. Here the Police would need a gadget to apply Weber's formula:

Weber's Contrast can be used to measure it and the formula is not so hard:

C_W = (L - L_b) / L_b

where:

L = luminance of the target/feature (in cd/m2)
L_b = luminance of the immediate/large surrounding background

Grok was therefore asked to check out the contrast in this photo using Weber value. Its response in part:

"The illuminated upper window (bright rectangle above the door)
Estimated L ≈ 40-50 cd/m2
→ Weber contrast ≈ (45 - 8) / 8 ≈ +4.6 (460%)
→ Extremely strong positive contrast — this is why the window immediately jumps out as the dominant bright feature."

But for the street lamp crowding the passing observers' increasingly peripheral view of the guiltier light...

"Estimated L ≈ 100-150 cd/m2 (direct view)
→ Weber contrast ≈ +10 to +18 (1000-1800%)
→ Extremely high"

Despite its lower Weber value Grok ranked the rectangular window top in "noticeability", partly because it is "bigger" than the "point source" of the lamp. Challenged to explain, it opined at length, see link. In brief, the lamplight is filtered out because it is too extreme, does not contain any information, and is peripheral to the scene, but of course less peripheral than the window to the observers - unless they were driving against traffic. Unlike them, Grok sees things quite romantically:

"In your photo, the warm rectangular glow of the window (and door) creates a stronger sense of invitation and focus than the sharp, point-like street lamp — even though the lamp wins on pure Weber contrast. That's classic nighttime photography drama: high-contrast tiny sources add sparkle, but extended lit areas steal the show. Great catch!"

The Defence of course concedes that the glare of the streetlamp as represented in the photo is somewhat more exaggerated and featureless than reality, due to the dynamic range of the medium. The legal team or Court may decide to contact the relevant authority to obtain objective data on the streetlamp.*

Exif data

Grok warns of a 30-50% error margin in its W_C estimates but makes sense by declaring the values true relative to each other. And critically for whatever the officers were feeling based upon these inputs to their eyeballs, what Grok tells us is all too human:

"Glare and bloom from point sources
Very bright small sources often produce optical glare, halos, or bloom in our eyes/camera (especially at night when pupils are large). This spreads light but can make the lamp feel more like a distracting 'hot spot' or source of discomfort rather than an inviting focal point. The window's light is softer/spread out, so it feels more naturally prominent without overwhelming."

This seemed like a good time to ask Grok about tunnel vision. Not Ptuj's alco algo, but actual physiologic tunnel vision that occurs when adrenaline (epinephrine) and cortisol are released to prepare you for immediate action in stressful situations.

Anything the Police noticed about this light when they were doing anything other than just beetling around would be of a degraded evidentiary value.*

Racing to an emergency or in pursuit the officers' pupils would be dilated, letting in more light and sharpening central focus, with narrowing of the visual field. The Defence proposes this would increase the Bouma ranges.*

Physiologic tunnel vision can also happen due to panic attacks or chronic hyperarousal, all things the Defence assumes the Police talk about a lot before going on a really expensive psychedelic therapy session for their PTSD in other countries.
https://x.com/i/grok/share/f7ac0888c71549ea944d9b4046fc29c7 [5874]

The investigators didn't measure contrast any more than brightness, and didn't count their observations or anything else. Overall, the sophistication of their data collection is inferior to the Mesopotamian (3500-3000 BCE) and Lebombo bone (44000 BCE) markers of arithmetical development.*
https://www.historyofinformation.com/detail.php?entryid=2338 [5780]

Because of the non-collection of data according to the non-design of their non-surveillance (non-surveillance is a legal position the Police are forced to adopt because they didn't get the right kind of warrant for surveillance) and because the Police want to say they surveilled this light without surveilling it, this prosecution should be shuddering to a halt.

But wait! Can we interpret the results their lighting investigation didn't collect?* And how about those of its reverse engineered mirror?

WHAT THIS DOES NOT ADD UP TO - INTERPRETING THE RESULTS:

Unlike the original, the reverse engineered study produced some results.*

If the Court seeks to know how many observations it would take per night to establish lighting continuity or eccentricity, the answer is...more than 2!

Sadly, even if they drove past every two minutes and it was on, or every night for a year and it was on...well so what??

Thus the Defence distinguishes between the banality of the nebulous informal hypothesis, and the mess made by not testing it.

Common sense tells us that the only concrete facts these observations could really have produced was that 1) the blind was broken,* 2) lights are more noticeable in the dark,* 3) they will be the only one visible if everyone else's blinds are not broken,* or if these others (whose households did not interest the Police) have their lights off at the moment the Police pass.*

Using standard parameters common to all measurements of physical phenomena Ptuj Police could not have predicted a level of certainty from their casual observations without predetermined known values for n and alpha, and some fixed ideas about the dimensions of "innocent" and "incriminating" lighting.*

Mohammed et al (2024) had an n of 2503: "a major constraint of our analyses is the nature of cross-sectional study in which temporal causality cannot be established." [4702]

On the other hand, a strict hypothesis, an a priori power analysis, considered design, systematic execution, and diligent record-keeping by the Police describe a degree of planning which would by definition qualify their activities as surveillance, as distinct from patrol status.*

In sum: an application was stitched from the hyperbole of Police agitprop, a vague to non-existent prior hypothesis with a baked-in non sequiteur, an unknown n, a meaningless design, convenience sampling suited to patrols of an atypical road at unknown times, an n<400 for an alpha of 0.05, the lack of a planned experimental model that actually occurred, the absence of results data, failure to prevent experimental error, the five blind, null, dim, jealous, and fuzzy pollyannaisms, the evolved tendency to underpowered decision-taking, inattention to the wider problems of induction vs. sampling density - all pointing to the statistical invalidity and impossible legality of the decision reached by the investigating judge on the relevance of the Defendant's 16 watts and 1120 lumens at 6500k whenever.

In general Ptuj Police's interpretation of their non-existent data and its peripheral information is whack-a-mole. When Mr T's word is worthless, it's the light. When the light is hogwash, it's the electricity bill.

And when it's neither of those, it's Mr T again, whose implausibility is the beginning and end of this futile exercise. In the cloudy original, each element is the excuse for the other two. In reality nothing has started.

This reminds us of the escape clauses [5809] typical of bad hypotheses. Even worse, this is a post hoc bad hypothesis, one so absent during the original study that it required reverse engineering by the Defence.

The Police investigation began with a conclusion and worked backwards. The reverse engineered study design starts at the beginning and works forwards.

It hardly needs to be said that the Prosecutor is free to argue with the Defence model of the informal hypothesis and power analysis inputs, with the proposition that the three elements of the warrant do not corroborate each other, and even with the existence of pleitropic effects.

Apart from the inanimate:animate transference phenomenon, the DFH does not quite "push all disconfirming evidence into an unfalsifiable metaphysical realm" [5809] with its crazy logic. Perhaps into a realm of short attention spans.

In this motion the lighting/electricity components are isolated from Mr T's component. But lest we forget, it was "all these things together" in the warrant application with which the Police set out to flood the zone of the misbranded "investigating" judge.* The only thing connecting these three elements are the Police and the burglars.*

For lights alone, the Court is reminded that in the Prosecution interpretation the faulty blind has a 0% effect. In reality it's 50% of, and essential to, the Effect "light through broken blind".*

An AND gate can be thought of as two switches in series. Unlike a real gate, you can only go through this logic gate when it's closed.

The partly open blind is the first switch in an AND gate, and it is (in the logic sense) stuck in a closed condition.

Were it not for this would the Defendant be innocent according to the investigators' criteria, unless it was opened deliberately (impossible)?*

Slovenia is prosecuting you for your big bill because of your blind now...

This brings us full circle to the bad hypothesis. We now see two informal post hoc hypotheses have emerged with the same results: light = farm/lab, and broken blind = farm/lab.*

This tells us that each hypothesis is weak due to its failure to exclude the results of the other;* both are too flabby to be tested without additional information which the Prosecution is too late to provide.*

In a great demonstration of the trouble with post hoc hypotheses, the two informal post hoc hypotheses are then "confirmed" with a third, effectively "broken blind = big electricity bill because cannabis".*

An alternative, more reliable but far less desirable Occam's Razor which might have intruded into the electricity story is "big house = big bills".*

But for unknown reasons the Police evaded this more obvious hypothesis;* the Defence's hunch is that only role for house size was already occupied by Jealous Pollyannaism. Does the Court agree?*

With this unfalsifiable hypothesis, the Police went to sea in a colander. Whatever the motivations for their bad hypothesis turn out to be (spoiler: these include Mr T and Ptuj Police being upset at Mr T being reported to the Police for burglary and encouraging arson, dour small-town hatred, envy, ignorance, low-risk career advancement, plunder, and Slovenia's limited sense of humour) the problem for the Prosecution is that the Dazzle Farm Hypothesis itself is drowning in superstition, illogical leaps, and the workings of the unconscious mind.

With watching things being their job, the Police can reasonably be expected to have an at least rudimentary knowledge of how eyes work.

The Police can reasonably expect their methods, observations, and inferences therefrom to be presented and questioned in a court trial.* They have removed the Benedictions from hundreds of people, typically by exploiting junkies and alcoholics, because the word "drugs" allows them to compress all their fans together into a mendacious mush.

Here, in the late Roman Empire of their war against evolution, the poor building standards of their edifice are laid bare.

Here, the Police manufactured evidence* via a manufactured crisis,* in a superstitious and suspicious investigation* fueled by a superstitious law.*

As such, their warrant application was a monument and hopefully an epitaph to the limitations of the CaPs prohibitionists' world view.

Baked into their informal hypothesis, the assumption "All A are B therefore All B are A" is what has been known since medieval times as illicit conversion (conversio illicita) of an A-proposition.

The distinction of valid vs. invalid syllogisms are traceable to their formalisation by Aristotle around 350 BCE, but bad syllogisms themselves are "as old as human arguing".
https://x.com/i/grok?conversation=2007936293821420027 [5824]

Reverse engineering of the Police options shows their failure was due to poor theory, preparation, and execution, whether defined individually or collectively as surveillance or not.*

The Police conducted their shapeless investigation of the Defendant's light in a non-organised way and collected no relevant data.*

There were no base values, or results to interpret. Neither null nor positive observations were recorded.*

No laws were being broken on this evidence.*

There was no answer to the question because there was no question.* The Police and investigating judge did not question the absence of a question, about this light, or know what the question was or could have been.*

Via reverse engineering this has now been clarified: Ptuj's Police and Court colluded in a circular story which makes no sense.*

More broadly this behaviour demonstrates the religious nature of CaPs prosecutions, and the state into which states have gotten themselves with the nutraceuticals' misrepresentation. From a besieged catholic doctor's perspective, 462,292 alleged barbarians are at the alleged gate.

The Police displayed blind prejudice both figuratively and literally.*

In further overreach from Planet Ptuj, a side hypothesis "broken blinds probably signify explosives" was fabricated out of nothing, so Officer G and his dog could come. He testified this dog does not detect drugs.

If CaPs (or blinds) and explosives are connected, why are there so few explosions in Slovenia? 2025 was a quiet year for explosions in Slovenia with one piece of WW2 ordnance, one gas grill, and one boat engine, with no mention of cannabis or psychedelics.

CaPs-related explosions are rarer than finding money at a bus stop.* The Police will save some, by avoiding this superstition too.

Zero cannabis-related explosions have been reported in Slovenia since independence in 1991.*
https://x.com/i/grok/share/FfkSdhmX00vdBY0CSOSGS9tVi [5849]

By contrast, whatever the Police have been doing is reckless and dangerous.

By steering away from logic and reality, towards bullying and plunder, full of overinflated self-righteousness, and with a law with no credible grounds - ultimately the witchfinders did more magical thinking than the witches. They had an excuse back then: ignorance. But not now.*

Power analysis is an antidote to proponents of kooky religious or superstitious ideas who believe what they believe because they believe that.

The Court may support kooky ideas and reject the findings of Aristotle et al. As we show, it already has. If things remain so, we must go higher.

These power analyses show mathematically that evidence proving the Police's already unworkable hypothesis was absent in the warrant application.* The Court may wish to examine the solutions cited above or with different parameters: for links to G*Power and tutorials see the Resources.

Politicians will not be crusading against, nor will Ptuj Police be racing to investigate, this 3.74 upper vs. lower tertile alcohol intake odds ratio for high-grade prostate cancer. Michael et al (2024) obtained these results of high significance from an n of 650, but admit that "given that the vast majority of men reporting heavier alcohol intake at ages 15-19 were also in the upper tertile of cumulative lifetime alcohol intake, we were unable to definitively separate the potential effects of early-life exposure from cumulative lifetime exposure to alcohol in this study." [2022]

MOTION TO DISMISS

It is shown that "several" observations of commonplace domestic lighting do not constitute any more grounds for suspicion than one or none. Several is not a number.

At the same time, several is not a huge number, a big number, or as many as 13, 37, 108, or 400. Even several million is just a few of something.

The Police account seems designed for the Pirahã language, spoken by a tribe in the Amazon, which famously lacks words for exact numbers, colours, and past or future tenses [5891]. This is all too convenient and slipshod.

Suspicion does not have to be based on something 100% concrete. That would be certainty, not suspicion. But nor can it be based on a dream, or conjured out of wish fulfilment or speculation.*

Why this light?* Because Mr T says so?* Why investigate it, then?* Was the broken blind probable cause?* Did it increase the suspiciousness of the light?* If so, how?*

Did the characteristics of these items somehow change the officers' opinion over time?* How was that?*

Or did they just build up steam by convincing each other of something empirically no more apparent at the end than it was at the beginning?* Or how was it more apparent?*

Are the odds of success by random raiding of lit households of over-25 males themselves sufficient?*

Was the suspicion well-constructed,* or fabricated?* Was there a thought process?* Did the thoughts exhibit valid reasoning* or contain false lemmata* in their sequence?*

From the information supplied, do the conclusions reached by the investigating judge follow?*

And speaking of reality more generally, what is the answer to Ms Rookmaker's query?*

What do your Benedictions biomarkers and mine have to do with the law anyway?*

What does Slovenia know about CaPs that other territories do not?*

The sleuths should be fully aware how important "I don't know" is in sleuthing.*

Empirically the officers' evidence to the investigating judge is no more weighty than "an angel spoke to us in a dream". Or perhaps a belief in beelzebulbs.

The public needs protection from tainted, underinformed, and simplistic drivers, and from superstitious manias and psychic epidemics.* It enjoys the rights to the Benedictions.* It has the right to feel sexy, musical, and connected to its environment.* Your opinion has no role to play in these rights.*

Walk the line all you like: prohibition is the problem. And it should butt out with its ignorant nonsense.*

Comparing the outcomes for CaPs users and drunk pharma consumers, what exactly is it you are trying to achieve?*

Here the observations were insufficiently numerous or organised to justify suspicion to any specified degree of certainty. A level which the law does not need to supply, as laws about certainty and the parameters thereto already exist in freestanding mathematical conventions, applicable to the lighting topic.*

Ignoring probabilistic realities is consistent with ignoring the unpleasant outcomes of a propaganda-defined "drugs"-chasing career. The hypothesis is mute.* What does it mean?* The logic, reconstructed in the reverse engineered model, is flawed and unrealistic in the cornucopia of ways already described.*

That the observations were not relevant and did not satisfy any criteria are mathematical but not yet legal facts.* The latter must proceed from the former.* That "several" detections of a light escaping in a central Ptuj street bore no grounds for a ransack should be news to no one.* If they do, shouldn't everyone know?*

What happens next depends on personal preferences.

A preference for folk myth correlates with a desire for unconstrained Police powers lacking definition, and contrary to rational aims.* This case shows how their distaste for rules about rudimentary empirical proof has made them popular with a tiny fringe element and unncessarily unpopular with a large natural and normal group.*

A preference for empirical enquiry correlates with an interest in hearing why CaPs are illegal.* To date the Prosecution has made no reply to the Benedictions. However it is really for them to show why the alleged offence is without right per Article 2(1) of Framework Decision 2004/757/JHA [2311].

Just because you don't like someone or their truth doesn't mean it isn't a right.

Meanwhile, after depriving the Defendant of his PTSD treatment Ptuj's Court expects him to get up close and personal with the guy they lined up to threaten him. For the benefit of its microphones! Think again!

With inadequate security and without justifiable excuse the Court is continuing to objectively elevate the risk of harm and anxiety to the Defendant and other participants by requiring our interaction with the unpredictable and dangerous operators of a vacant casus belli, whose collapse now resembles the last days of the Stasi [5638].

The Defence considers this a violation of the principle of proportionality.

To the question of proportionality, it is proposed that the Prosecution has zero reasons to oppose the Benedictions. Rather it is proposed that it is those who oppose them who should be opposed. And we do.

The proportion of some number to zero is infinity, and it is with infinite skepticism that the Court should weigh the Prosecution's zero argument. Wait until Ptuj hears it is going to be overrun with Niagarans [5942].

The Court in its reply of 31 January 2025 "emphasizes that the issuance of a warrant for a house search requires the existence of a relatively low standard of proof that the suspect has committed a certain criminal offence".

So it has to be low? Here the question of how low you can go has been explored, and no standard was found of any kind.* Can the Police hallucinate something out of nothing and present it to the investigating judge?*

Compared to lighting's weak association with suspiciousness, the associations of anti-CaPs laws with suspicions of racism, eugenics, and political medicine are far better chronicled, far more reliable.

The Police witness insists to the Court that the Defendant's two 8 watt lights, separated from her by a distance of over 14 meters and four layers of glass, were brighter than the many unobstructed banks of LED battens perhaps 1.4 meters above her eyes, which will be 30 to 50 watts per strip, 60-100w for a pair of 120cm, up to 280w for four 180cm strips, and therefore if we take a massive underestimate of 100w for the courtroom...

By the time of the preliminary hearing, a myth had been reinforced to the point where the Police were prepared to testify it was brighter than the courtroom. Fortunately these matters are calculable* and show the evidence provided to obtain the warrant was very untrue.*

A lie went around the world before the truth got its boots on. But a fully-booted analysis shows that for the brightness of the window's light to agree with the officers' report would have cost the Defendant from at least €2767.03 to €4611.68 weekly in lighting costs alone at 2020 prices.

So no, Officer T was more than a little mistaken: the Defendant's lights were not brighter than that.* Is any wild, dishonest tale justified by the thought-terminating cliche?* Will any story do?* In Slovenia? Cui bono?*

Indeed, as an either an expert in cannabis lighting detection or merely an everyday end user of the now widely appreciated laws of physics, the witness cannot be unaware her testimony could not possibly be true.*

Only a true examination of the lumens per watt and wattage of the Court's lighting will expose the scale of the witness's hyperbolic tendencies. 100 watts was a massive underestimate for the spaces you work in.

Some estimates of your normality - lighting standards which must not be afforded to the Defendant for ideological and discriminatory reasons - have been prepared: [5927].

Our faith in physics over her imagination is further strengthened by the cast-iron guarantee of the Court's adherence to EU standard EN 12464-1 (and its Slovenian implementation for a giveaway price of €130.68) which recommends 300–500 lux at desk/eye level for your reading and writing tasks.

https://ecommerce-sist-si.translate.goog/catalog/standards/sist/a3f2647e-41c6-4bef-adbd-711b024c6244/sist-en-12464-1-2021?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=sl [5949]

This actual giveaway explains without invoking the endocannabinoid system how "Our eyesight has an exceptional ability to adapt to brightness variations from a few lux to more than 100000 lux."

https://ec.europa.eu/research/participants/documents/downloadPublic?documentIds=080166e5b6d1bf85&appId=PPGMS [5950]

Now we see the officer's argument in evidence is indeed being pushed into a metaphysical realm, as following her discovery of mydriasis and miosis there isn't really anywhere else for it to land. What is the Court to make of the investigators' divinatory occult powers and post hoc augury?*

No more believable were the "results" of the "several" observations, which were not statistically significant, but only garnish for a preconceived, prejudiced opinion lacking any objective or relevant inputs.*

These "several" observations were fewer than all possibly valid values of n using reasonable parameters,* even were the hypothesis less dubious and experimental error less probable.*

Within the Defendant's demographic, can a raw probability of correct farm-guessing between 1 in 175 and 1 in 5791 be enough, absent other criteria?* In this manner the basics of 85 years of modern science were abandoned in favour of folk myth and support of hate speech, in the issuance of the warrant under a pointless hate law.*

As to sample adequacy, neither the value of n, nor whether or not to even have an n, can be a matter of individual or collective whim.* Collection facilities for an adequate standard for suspicion according to the power analysis would have required a surveillance warrant, using credible metrics and parameters.*

Not having an n doesn't make the reality of n go away.*

In what resembles a vindictive action, the odds that the five Pollyannaisms exerted zero influence is 1 in 32.

They show not a series of accidental mistakes - but a repeated turning away from each blatant flaw in the reasoning the officers presumably allege took place.*

Superstitions are apparitions. The Defence therefore asks the Court to answer that the impossible perceptions alleged by the observers, the impossibility of obtaining any meaningful interpretation of the lighting and electricity elements, and the lack of a falsifiable attempt to do so, affirm in toto the impossibility and illegality of a search warrant based partly or wholly upon said interpretation.*

In closing, the Defendant asks the Court and Police to treat him equally, by recovering his stolen binoculars. The lenses and their coatings have great apochromatic, anti-flare, anti-ghosting, and sentimental values.

By returning to square one, maybe Mr T and the team will catch some Slovene-speaking arsonists then too - whose motives they have inexplicably found less suspicious than their non-Slovene-speaking guest's light.

RESOURCES:

Induction and falsification

From Hume to Popper, the philosophy of science and practical statistics since the 1930s, and the Bayesian response. Grok and the Defendant put it all together:
https://x.com/i/grok/share/g1xw1Xuj7sSetCRmVwTvM4duG [5765]

The eyes have it: why we don't notice changes in lighting, Weber's law, and crowding

"The law states that the smallest change in a stimulus that can be consciously detected — called the just noticeable difference (JND) or difference threshold (ΔI) — is proportional to the magnitude of the original stimulus (I). In other words, the relative (not absolute) change matters."

Asked to analyse this photo in the context of Weber's law, optical or ophthalmological confounding variables, miosis, mydriasis, apparent brightness due to ambient contrast, lateral inhibition, crowding, and illusions, Grok advises that:

Metamerism and ocular sciences

https://youtu.be/FTKP0Y9MVus [5943]

http://hyperphysics.phy-astr.gsu.edu/hbase/vision/colcon.html [5944]

http://www.12v.si/metamerism [5959]

http://www.12v.si/metamflow [5961]

http://www.12v.si/metaminstructions [5962]

Sample size and power (clinical trials oriented)
https://www.youtube.com/watch?v=tQ5Ua7cT1To [5702]

Power analysis (general introduction, parametric tests)
https://www.youtube.com/watch?v=OkbXBWuzikM [5703]

G*Power power analysis software tool
https://www.psychologie.hhu.de/arbeitsgruppen/allgemeine-psychologie-und-arbeitspsychologie/gpower [5701]

G*Power tutorial with t test
https://www.youtube.com/watch?v=L5LJIm-h9bc

G*Power tutorial with exact test
https://www.youtube.com/watch?v=MQ2qs4WKgT4 [5704]

G*Power manual

https://www.psychologie.hhu.de/fileadmin/redaktion/Fakultaeten/Mathematisch-Naturwissenschaftliche_Fakultaet/Psychologie/AAP/gpower/GPowerManual.pdf#page=16.05 [5712]

What is critical z?
https://x.com/i/grok/share/IKfhhmiyKw3vF8iF6HS1sFHTt [5722]

The z in "Fisher z-transformation" is not the same as the z in "z-test for means/proportions".
https://x.com/i/grok/share/6XDeMGtpPBXLh4lJJjxmqR5Zr [5788]

"P Value and the Theory of Hypothesis Testing: An Explanation for New Researchers" by Biau et al (2009) is a history which delves into the Fisherian, Neyman-Pearsonian, and Bayesian approaches to p-values:


https://pmc.ncbi.nlm.nih.gov/articles/PMC2816758/ [4962]

The problem with p, sample size, power, and other biases
https://stillinthestorm.substack.com/p/most-published-research-findings [2363]

Does an alpha of 0.5 mean you are prepared to get it wrong 50% of the time?

                                                   Reality

       Decision                     No effect                There is effect

       Claim effect                 False positive           Correct (Power)
       Claim no effect              Correct                  False negative

False positive rate = only the top-left cell — (top row)
Overall error rate = (top-left + bottom-right) ÷ all cells

"No — alpha = 0.5 doesn't mean 'wrong half the time in general'.

It means 'I'm okay with screaming "significant!" about half the time even when nothing is actually going on.'"
https://x.com/i/grok/share/fab52fcb9e1847c896bf6c3e69ef083b [5875]

The other alpha: why is there an alpha (α) in E ∝ P / d²?

http://www.12v.si/gn [5960]

Variance and power analysis
https://x.com/i/grok/share/xOrfLrWvgQ9P7OB9DFqrMsj85 [5782]

Daylight finder, amplitude of Ptuj and sun positions

Daylight hours ≈ 12 + A ⋅ sin(2π (d − d₀)/365.2422)
https://x.com/i/grok/share/ptrZcBUpu6L7hQySLNF7LbhpD [5855]
https://www.gaisma.com/en/location/ptuj.html [5856]

Sun times for Ptuj without DST (2020).

Full day-by-day sheet.
https://docs.google.com/spreadsheets/d/1ldyxgehiLeOzDsHGnw4p-tEVqSl6kxgW5LfsiZWaK2g/edit?usp=sharing [5904]

Grok's mistake about night length as a proportion of 24 hours and why it doesn't matter.
https://x.com/i/grok/share/f0f49b9df99149e8b54bb68b456d3e7c [5711]

Does the coupon collector problem assume an infinite stock of equiprobable coupons and doesn't this fail to represent a real situation? Differences with consecutive k and reliability testing are examined.

We learn of a

"...widely cited 'rule of thumb': To demonstrate 90% reliability (R = 0.90) with 90% confidence (C = 0.90), you need n ≈ 22 consecutive successes with zero failures.
For 95%/95%, it's around 59 trials.
For 99%/90%, it's about 230 trials."
https://x.com/i/grok?conversation=2011745699675681272 [5870]

"How do you establish 'I just happened to notice' 16+ times without it looking like stalking?

Rival AI bot Claude checks Grok's answers, makes some mistakes, but ultimately draws similar conclusions, offers insights into the practical value of different sampling densities, designates how excessive sampling density leads to autocorrelation, figures out without prompting that the surveillance warrant is a supervening issue and, with prompting, that Ptuj is a terrible town.

Claude's initial results http://www.12v.si/cp
Full chat at http.12v.si/cc [5582]

Claude concluded at least 15 random or 13 consecutive observations were needed.

If you "just happened to notice" you could equally "happen to not notice" when something isn't there.

Strictly, null pollyannaism and surveillance warrants obviate further investigation of Claude's low k compared to Grok's 37. But asked about this, Claude criticised himself again and supported Grok, with a comparison. However his charts frankly didn't really match his own data. It subsequently emerged that the chart is correct, the 144 n for 75% was in error, and the actual result is 15, as shown here
http://www.12v.si/cg [5853]

With further egging on Claude tried four different ways Grok could have calculated an n of 37. Two came up well short, one was too complicated, but Wald's Sequential Test with overshoot correction yielded 37.1. This is explained in detail in the full chat [5852] and amended artifact:
http://www.12v.si/ca [5854]

"See why" you could lose the case - Police calculator for beelzebulbs

3D histogram for p1 = 0.98, with raw effect on the x-axis, statistical power along the z-axis; on the y-axis the height of the columns represents sample size for each combination. Hover to see Cohen's h while experiencing the Gelb Effect.
http://www.12v.si/cy [5859]

$\text{Square Degrees}=\text{Steradians}\times {\left(\frac{180}{\pi }\right)}^2$

$\text{Square Degrees}\approx \text{Steradians}\times 3282.806$

Sumljivometer

The Sumljivometer is a nomographic tool for justice-related superstitions.

Calculate Defendants' suspiciousness on a lighting basis.

Instantly evaluate guilt, based on night lighting activity.

Find hours and minutes of darktime (night + astronomical twilight + nautical twilight) as a proportion of any date (blue number).

Compare any lit percentage of darktime (brown) for Court evaluation of shadiness on a quantitative lighting basis. Use in your own G*Power calculations to make sure you stay on the right side of Ptuj's draconian lighting rules.

Covers 98.78% of the Earth's surface between the polar circles.

Substitute lighting statistics for hard-to-quantify lifestyle- and tidiness-based evidence.

Catch your neighbours red-handed, with concrete proof of atypical lighting.

Experience darkness as vibrations with the Nightsound feature (under development).

With AMSA Advanced Mood Swing Alcorithm and profiling tool for persistent lighting offenders.

Sumljivometer has non-prosecutorial uses too: calculate sensor-based lighting hours, predict costs, scale batteries for any latitude, etc. etc.

Fully functional tactical blind. Quickly hide your results from other suspicious folk.
http://www.12v.si/sumljivometer [5900]


26 hour days

Time-based calculations about indoor farming like those in the reverse engineered study assume a 24-hour cycle, but it turns out this is only one possibility among many.



While the Police have been hunting down fans of the Benedictions, cannabis growers have been solving world hunger by discovering new information about plant development cycles, and that plants can do better on supercycles (over 24 hours):

"Some plants needed longer nights. Others exploded under extended days. They tested strawberries, calendulas, cherry tomatoes, and flowers, and all showed signs of hyperproduction."
https://hightimes.com/grow/fuck-12-12-supercycle-breaking-the-cannabis-flowering-rule/ [5289]





Quick guide to logical fallacies
https://x.com/SahilBloom/status/2009252565293125901 [5857]

More statistics

Sample size - page 270
null hypothesis - page 459
p values - page 462
Stats - Modeling the World
by David E Bock, Paul F Velleman and Richard D de Vaux
https://www.chino.k12.ca.us/site/handlers/filedownload.ashx?moduleinstanceid=26467&dataid=45087&FileName=Stats%20Modeling%20the%20World.pdf [2362]


Student v Fisher

"Prior to Gosset the relation between sample size and the level of statistical significance was rarely explored." [4555]

Writing in the Journal of Wine Economics (2011) about the history of statistical significance - some of which grew up around the alcohol trade - Stephen Ziliak of Roosevelt University Chicago says

"Student [William Sealy Gosset (1876—1937) aka 'Student' — he of Student's t-table and test of statistical significance] opposed Fisher's randomized field experiments on grounds that, as Student proved as early as 1911, decisively so in 1923, and again in 1938, balanced designs are more precise, powerful, and efficient compared to random.

"Brewers and economists alike have not noticed as much as they might that Student's exacting theory of errors, both random and real, marked a significant advance over ambiguous reports of plant life and fermentation asserted by Priestley and Lavoisier down to Pasteur, Fisher, and Johannsen, working at the Carlsberg Laboratory in Denmark."

In field experiments on barley:

"In an admittedly crude way, Student proved in 1911 with Mercer's and Hall's mangolds data that randomization is inferior to deliberate balancing. Student spent three decades refining his theory and further proving the statistical and economic advantages of balanced designs."

The debate on randomization (Fisher) vs. balancing (Student/Gosset) was still going strong in 1937:

"The point that Student made and Fisher refused to acknowledge is that deliberate randomization gives less assurance that treatments and controls will have equal access to good and bad soil, to good and bad growing conditions. Randomization-based 'validity' is lovely to contemplate in theory. Yet compared with balanced designs, random designs have been found to be imprecise, expensive and comparatively powerless.

"Balanced designs are deliberate or systematic arrangements of treatments and controls made by the analyst conscious of real ground and/or other confounding sources of fluctuation in the output of interest. 'Balancing' means creation of symmetry in all of the important sources of error, random and systematic, good soil and bad."
https://cpb-us-e1.wpmucdn.com/blogs.roosevelt.edu/dist/b/153/files/2012/02/William-S-Gosset-and-Experimental-Statistics-Ziliak-JWE-2011.pdf [4555]

Insofar as they were concerned with observational studies at all, the early bans which drive the present case could therefore be based only upon the anecdotal evidence and unreliable perceptions of statistical power available at the time.



A major step in the evolution of the scientific method was R A Fisher's tea tasting lady, and Chapter II on her, on testing significance, and the birth of the null hypothesis, is worth reading if you want to understand what was still groundbreaking in the 1930s.

The null hypothesis was occasionally mentioned before "The Design of Experiments" was published, but didn't really get its legs until after a decade after the 1925 Opium Convention, also the year of the last international cannabis review.

Its publication coincided with Logik der Forschung (1934, 1935 imprint) introducing Karl Popper's criterion of demarcation between scientific and non-scientific investigations, and theory of falsification.

It follows that none of the medical evidence submitted at the treaty negotiations or in all probability the 1935 review could have employed Fisher's Exact Test or Popper's post-Vienna philosophy of science.
https://archive.org/details/in.ernet.dli.2015.199572/page/n23/mode/2up?view=theater [2755]

The same applies to Lindley's Paradox:
https://en.wikipedia.org/wiki/Lindley%27s_paradox [3624]

Fun facts: proofs in CaPs prohibition vs. incombustible salamanders

Aristotle probably died believing salamanders are incombustible - and even extinguish fires. This fitted a belief system about four elements somehow.

Has this been scientifically verified? Even top geezers like these get it wrong sometimes, and for the usual reasons - see page 471 et seq.
https://ia800505.us.archive.org/17/items/mythslegendsgold00bulf/mythslegendsgold00bulf.pdf#page=25.08 [5866]

Pliny the Elder later expressed skepticism, and was later (probably) misrepresented as having tested the assertion by throwing a salamander on the fire, where it died. Nevertheless...

"The myth persisted for centuries, influencing medieval bestiaries, alchemy (e.g., Paracelsus viewing salamanders as fire elementals), and even heraldry, but it was gradually debunked through direct observation."

Grok was unable to unearth any statistically significant experiments in salamander ignition, with anecdotal case histories and stories of their flight from wet logs on the fire rather tending to negate the hypothesis.

Yet it's "common knowledge", not our own, that make these notions about asbestos amphibians feel like ancient history; they linger in New Age circles. Paracelsus is frequently cited in support of monotonic dose-response and made to sound more anti-hormesis than he actually was. Aristotle's logic is reputable enough for a name-check in the present case.

But importantly, the continuing absence of real-world, standardized flammability tests on an adequate salamander sample does not allow us to conclude anything whatsoever about the real facts one way or the other.
https://x.com/i/grok/share/6gOCiXsRBTgA68dhngdRxzIA8 [5867]

Though it makes us uncomfortable and is hard to confront, the certainties upon which the international prohibitions on CaPs were partly grounded (the rest being bluff and hearsay) were no more reliable than the superstitious blunders about salamanders expressed in European heraldry - such as in the crest of the City of London's Worshipful Company of Fire Fighters...

https://wcoff.org/modarms.php [5868]

...and this one from 1963 belonging to the UK's Institution of Chemical Engineers, whose motto "Findendo Fingere Disco" sounds great, but merely means "I learn to make by separating." And as you can see, the party-goers are separated.

https://www.heraldry-wiki.com/wiki/Institution_of_Chemical_Engineers [5869]

For me, this ruins all of Aristotle. What an idiot! These symbols of earlier beliefs, and the lack of amphibian fire-extinguishing interventions in infernos from chip pans to emergency landings and reports thereof in chemical engineering journals, all suggest myths should not be banned, but not taken seriously either. 

Slovenia should not be creating offences and offenders out of myths.

SUMMARY OF INTERROGATIONS USED IN THE REVERSE ENGINEERED STUDY:

What are the null and alternative hypotheses?

What effect are the Police attempting to measure?

What metrics should be chosen?

Does mechanistic or ecologic research provide any reason to logically connect the thing under observation with the inferred effect?

Are the metrics and effect reliably relevant to the hypothesis?

Does the selection of metrics affect interpretation of the results?

Is there a temporal element? Are rates involved? Means or proportions?

Will the observations be scheduled or random?

Can opportunities for observations to be made be missed or avoided?

If observations are non-standardised, what adjustments (e.g. cohort matching) need to be made to prevent sampling error?

With what equipment will the values be recorded?

Will null or zero values be recorded with the same diligence as other values?

Is the nature of the enquiry pre-loaded with suspicions or superstitions about the outcome?

What unmeasured factors, excluded from the experimental model, could interfere with a fair comparison?

Can causation be inferred from association by the chosen experimental design?

How did they ask the question?

Did they formulate the hypothesis in advance of the actual testing?

How did they test the hypothesis/answer the question?

Were the observations counted and all the relevant data recorded?

Did the outcome show the value of a priori power analysis and scientific testing?

Did the outcome show the danger of confirmation bias, post hoc and informal hypotheses, and wishful thinking?

Could the theory of Officers L and T have been tested?

Was it tested?

Did they get a high power, statistically significant answer to an illogical question?

THE MOTION SONG: DEDICATED TO THIS MOTION
https://www.youtube.com/watch?v=k-HdGnzYdFQ

----------------------------------------------------------------------------

The Englishman stands for the rights of everyone disadvantaged, discriminated against, persecuted, and prosecuted on the false or absent bases of prohibition, and also believes the victims of these officially-sanctioned prejudices have been appallingly treated and should be pardoned and compensated.

The Englishman requests the return of his CaPs and other rightful property, for whose distraint Slovenia has proffered no credible excuse or cause.

The Benedictions represent both empirical entities as well as beliefs. Beliefs which the Defence evidence shows may be reasonably and earnestly held about the positive benefits of CaPs at the population level, in which the good overwhelmingly outweighs the bad. Below, the latest version of this dynamic list.


THE BENEDICTIONS                            REFERENCES                        TIMELINE OF DRUG LAW v. SCIENCE