One more time
Okay already! Enough with the god particle! (For those of you who aren't movie fans, this photo is from a famous scene in the movie Pulp Fiction.)
The combined 4.9 sigma result reported for the Higgs boson is hailed as a stunning achievement that took trillions of recorded events, billions of dollars, and thousands of scientists.
By contrast, several classes of combined psi effects already provide empirical results that are much, much greater than 5 sigma, with hardly any funding and a few handfuls of scientists working the problem.
Some future day when physical theories tackle the mysterious boundary between objective and subjective realities, they'll start to predict psi effects (I believe that day is inevitable). When that happens psi data will suddenly make sense. Then I'll have to change the image caption to "Say psi particle one more goddamn time."
Comments
There is enough data from reliable sources to be sure that some future paradigm will indeed predict psi phemomena and you will be one of the people children read about in school. I just hope we all live long enough to understand the phenomena theoretically.
Best wishes
In its day, within the particle physics community it was almost as big a deal as the Higgs though it didn't get as much popular press (in fact, along similar lines to your cartoon, I remember a hand drawn cartoon on the corridor wall at the CMU physics department expressing similar sentiments to yours -- stick figures in rows of seat with a figure in the front gesturing with a pointer. The figure in the front says: "I have discovered a new particle", while the audience says in unison "ψ!").
The excitement was because the particle was the first proof of a fourth quark (the charm quark), and was the lowest energy form of charmonium (a charm quark bound to a charm anti-quark).
It was actually discovered, virtually simultaneously, by two groups, one dubbed it the psi meson, while the other dubbed it the J meson, so its official name is now the J/ψ meson -- the only particle with a digraph name (discounting subscripts and diacriticals).
Topher
is this 'God particle' also a wave?
...to which Dean replied
Yes, but only when you're not looking at it (like all particles).
True bu it's not looking at the Higgs that makes it significant.
Virtual particles are particles that the laws of physics allow to exist only because their existence is so brief that they cannot be observed and therefore their appearance out of "nothing" cannot break any laws. Such particles are constantly "boiling" out of the vacuum but disappearing again before they can be seen.
That they cannot be observed directly, however, does not mean that they cannot affect "real" particles (which are also waves of course) in ways that do not constitute a quantum "observation". The continuous ghostly haze of these virtual particles create fields. For example, the continuous generation of virtual photons creates a background electromagnetic field in "empty" space whose effects are observed in the "Lamb" and "Casimir" effects.
In order to explain why things have the mass that they do, it was theorized that there exists a directionless (scalar) field that interacts with observable particles creating a drag when the observable particles accelerate. But where does this field come from? From the generation in the vacuum of virtual particles of an unknown type -- the Higgs particle.
To confirm this theory, though, they needed to actually observe non-virtual instances of this particle. They have now apparently observed a particle within the right mass range and with some of the right characteristics to be the Higgs.
But the importance of observing the Higgs (if that has happened) is what it says about the sort'a existence of Higgs particles that do not have enough existence for us to observe them.
In physics the convention is that to be significant an experiment must be 5 sigmas from chance or about 3*10^-7. This probably comes from the characteristics of most physics experiments: identical trials with little variance is between them is to be expected (electrons are all the same, and if you put some work into it, so are spheres of, say, nickle), and once you've set things up its generally easy to crank out more trials. So why stop at 2 sigma?
There's even less justification for this then for the alpha of 0.05 in other fields, but it is the standard. Physicists are quite superstitious about it. They seem to tend to have an unexamined belief that this extreme somehow provides protection against experimental errors -- that it gives some assurance that the results are not only not a statistical fluke but "real" in terms of not being due to experimental "sloppiness" or poor analysis. Or to put it another way, they seem to think that unless the probability of a statistical false positive is less than .0000003 then the probability of a false positive is somewhere around .5, but if it is less than .0000003 than one doesn't have to worry too much about false positives.
I'm not exaggerating. When you challenge physicists on the rigidity of this standard they will tell you that "4 sigma" results are "common" but "frequently" disappear.
Think about that -- a "larger sigma" only protects you from statistically generated false positives. Only one properly conducted experiment with a false premise out of 33000 will produce a false positive at 4 sigmas. Its true that going to 5 sigmas will eliminate 99% of those, but how often do you need to worry about that? Well, 1 experiment in 33000 if almost all hypotheses tested are false -- much less often if physicists occasionally test true hypotheses.
Of course, this completely ignores the fact that cranking out trials until you get a positive result at 5 sigmas (or 4 or 1.6) is statistical nonsense. Random walk theory proves that if you keep at it long enough you will always reach whatever level of significance you want eventually. (Is it obvious yet that I'm not a physicist, however much I'm interested in the field?)
"Cern should really build a larger auditorium. The present one is nice and cosy, but it is embarrassing and sad to see many distinguished colleagues queueing up at five in the morning knowing that they have a slim chance to get a seat, after working for 20 years on finding the Higgs boson,said Dorigo (a scientist on the CMS experimental team at Cern)."
Considering that the thing cost 8 billion dollars (that they'd admitted to at that point) and who knows how much to operate and maintain, the lack of an auditorium with sufficient seating for the announcement of the "Higgs-like" particle, which no one I've asked can think of a practical use for, just breaks my heart.
I'm also finding the complaints about the "God particle" moniker funny as the physicist, Leon Lederman, seems to have first called it that. Not sure he's an atheist or not.
Makes me wonder what the next big toy the Lords of Creation are going to whine for us to buy them and what the next particle, or whatever, NASA style hype will surround. Also what percentage of 1% of those involved in the blog hoopla actually understand what it is they've discovered and what it means.
I don't want to rain on your parade -- we really need people to be open to be willing to look at parapsychological experiments seriously and to be willing to conduct them -- but you probably weren't seeing psi results.
Parapsychologists take great care with being rigorous with their experiments not because of unreasonable demands by self proclaimed skeptics but because there really are subtle effects that will produce the appearance of psi if you don't exclude them.
Roger Nelson has shown that for whatever reason, psi experiments generally have a roughly constant effect size per unit time. While your effect size (about .25) is in the range of typical psi experiments, those experiments are typically done over a period of months, so your effect-size/hour is way beyond what would be expected in a psi experiment.
Some people (including Dean, I think) have warned parapsychologists that a tendency to exclude data merely because it is "too good" could result in discarding the best data unreasonably. I agree with this, but at the same time we must pay attention to the characteristics of the phenomena we are studying that have been discovered in our own experiments.
Casual experiments don't make it harder to get high Z scores, they make it easier to get high Z scores that don't mean anything. Hence, routinely getting more than 5 sigmas in casual experiments doesn't really justify a "that's it" comment.
Lets say, for example, that you're casual experiments consisted of taking a standard playing card deck, shuffling it, then guessing what the next card is, then dealing it and seeing how you did, and repeating that through the whole deck. (I'm not saying that that is what you did, this is just an example).
The simple subconscious strategy of remembering if there have been more reds or blacks dealt previously and guessing the less frequently seen color will give you a 58% hit rate.
Moreover, shuffling a deck thoroughly is non-trivial. If you are good, fast riffle shuffler you would have needed to have spent at least an hour shuffling the deck to produce 1600 meaningful trials with this design. Cards tend to stay together so guessing the color of the card that followed the last card seen in the previous run would tend to give you quite good results -- with more sophisticated strategies (e.g., if the next card wasn't the card that followed the previous card in the previous run the one after that is likely to be that card) and combining it with the previous strategy would be expected to produce a very much higher hit rate in less the deck were skillfully shuffled at least eight times between each run.
I'm really not trying to be a dog in the manger here. Casual experiments are fun. But generally they don't say much about psi phenomena. That's the error that creates a Susan Blackmore.
Has such a "subconscious strategy" been demonstrated to happen or is it some imaginary faculty cooked up out of nothing more than speculation? Is the ability to do that effectively widespread or evenly distributed? And what in the "subconscious" would be motivated to make the effort to do that? Why would ones "subconscious" go to what would have to be a considerable amount of unconsidered trouble to make that kind of stab in the dark?
Wouldn't you have to establish that such a phenomenon existed before it could be reasonably introduced as a possible impediment to this kind of experiment? It would probably be possible to invent similar imaginary faculties or effects that would negate enormous realms of psychological research. Including that involved in "verifying" the existence of such a faculty.
"Moreover, shuffling a deck thoroughly is non-trivial. If you are good, fast riffle shuffler you would have needed to have spent at least an hour shuffling the deck to produce 1600 meaningful trials with this design. Cards tend to stay together so guessing the color of the card that followed the last card seen in the previous run would tend to give you quite good results"
How would your, I'm assuming "subconscious" know what part of a deck had been effectively shuffled to produce a different sequence and which hasn't? And how would it know that a sequence of cards was retained in subsequent shuffling? How much of this guessing of sequences, again, I'd imagine "subconsciously", would have to be accurate to have a serious impact in distorting some kind of mysteriously unconscious effect on card guessing?
No.
Has anyone run an experiment to see if the experimenter tells the subject what the card actually is before the subject guesses whether they get above "chance" results.
No again.
What has been thoroughly, repeatedly, and overwhelmingly demonstrated is that ones subconscious does this kind of thing all the times -- and is, in fact apparently better at it then ones conscious. The subconscious is extremely sensitive to picking up patterns and reacting to them. That's how you learned to speak as a child. Its how you judge how someone is feeling. Its how you drive a car (unless you are a very poor driver).
Since I can easily imagine doing this consciously, I have no doubt my subconscious is capable of it. In this case I can be stronger -- its been shown that people attempting to create a random sequence tends to try to balance things out -- precisely what I am talking about.
I believe without question that psi exists. I also believe, however, that a lot of what people do and think and feel happens in their minds at a level that they are not consciously aware of.
If I were to see a psi experiment with this flaw in it that produced typical results, I would say "Since you got results similar to the results typical in psi experiments, its very plausible that your results are due to psi, but unfortunately, we can't be sure."
But if I see an experiment with the flaw I mentioned, but with results as extraordinary as Raving Chemist reports, I would have to say "If it doesn't look like a duck and it doesn't quack like a duck, and it could easily be a plum tree, then it probably isn't a duck -- I'm afraid that this probably isn't psi giving you your results".
If someone gets from New York to LA in 8 hours, I'm going to doubt that they traveled by horse. Doesn't mean that I don't believe in horses, nor that people are incapable of riding them.
The subconscious wouldn't have to know. It would just have to follow the not unreasonable pattern that what happened previously is likely to happen again. When this seems to work, the behavior would be reinforced -- if something works people tend to stick with it.
How accurate would they have to be to produce the small effect that described here? -- not very.
Would "delaying" feedback eliminate the effectiveness of the first strategy I mentioned? Yes, absolutely. This would be a better, less "casual," more rigorous design. This was precisely why the Duke lab stopped giving trial by trial feedback in card experiments using "closed decks" back in the 1930s.
While the specific strategy I mentioned based on how poorly people tend to shuffle cards unless they work at it depends on trial by trial feedback, there are other ways -- fairly simple ways -- of taking advantage of a poor shuffle that do not.
(For the record, it has been shown that good bridge players take advantage of typically poor shuffling -- without being aware of it, i.e., subconsciously).
Of course, I do not know whether these particular flaws actually occurred in the incident that Raving Chemist reported -- he gave no details so I was just providing some examples of the kind of problems that one can run into.
Maybe there was 1600 well mixed cards and no checking was done until they had run through all of them. But this would hardly be described as a "casual" test, and there are lots of other potential problems that are hard to rule out in anything reasonably described as "casual".
I do not know whether these particular flaws actually occurred in the incident that Raving Chemist reported -- he gave no details so I was just providing some examples of the kind of problems that one can run into.
This reminds me so much of my recently looking into the accusations made against J. B. Rhine's experimental methods by Irving Langmuir, as repeated by one of the lesser of the CSICOPs. Langmuir, himself, misrepresented Rhine's methods so as to mock his work for the amusement of his audience. Then the, admittedly incomplete, tape of his talk, given decades after Langmuir's conversation with Rhine -with alleged quotes from Rhine - was allegedly transcribed, decades after Langmuir's death, to be used as "evidence" against Rhine, after Rhine's death, in a book by the pseudo-skeptic. Since then the flawed "evidence" has endlessly echoed in the pseudo-skeptical blogosphere. Only, if the CSICOP really was interested in the truth, he could have seen that Rhine had already said he'd been misrepresented and that any statistical review of the charges made would show that what Langumir said was impossible.
It's easy enough to see which side practices actual science instead of ideological conjecture and which side practices intellectual honesty instead of telling and repeating ideologically advantageous anecdotes instead of accurately reporting data. And that is not to mention outright lying.
I agree with that part of what was said. It would be good if conventional psychology had to practice a similar level of rigor, which it usually hasn't. If it did some of the extravagant claims made by psychologists over the past century might never have been published, only to be discarded later. I think the lax standards practiced in psychology are, actually, dangerous. But that's a different discussion.
The ad hoc creation of faculties and abilities, especially those alleged to operate on a subconscious level, has figured both in some of those claims and in popular pseudo-skeptical invective. I was merely pointing out that things like this ability to correctly remember sequences of cards in an insufficiently shuffled deck are also a proposed mental faculty, just as PSI abilities are mental faculties. Only, when a "skeptic" brings that into the criticism of experiments demonstrating a likely PSI ability, it will seldom be, itself, questioned and those alleging such an ability are never required to demonstrate that this new faculty actually exists or is present when they say it is.
Thinking more about it last night, a person who is trying to guess a card is doing so one at a time.
The scenario proposes that the deck of cards is insufficiently shuffled so as to retain sequences of order from the previous order in the pack that the "subconscious" remembers and can correctly guess in the subsequent deck order.
It would be monumentally unlikely that the entire or even much of the sequence would result from an honest shuffle. So which of the possible retained sequences that remained in the new deck order would not be determinable by reason.
Any sequences retained would almost certainly not appear in the new order in any reliable order they would have held in the complete, original deck order. If anything an unpredictable sequence of partially retained sequences would be expected.
I'd think that alone would make such a faculty impossible. It would be impossible to anticipate which orderings were retained and the sequence those would appear in the new deck order.
In order for a card guesser to identify the card they were guessing was part of a "remembered" sequence, they would have to be able to guess if the previous card or cards were parts of a partial sequence remembered from a previous run or that the next card up would be.
In any card guess, they would have to be able to pick out the correct possibility of one, out of any possible number of partially retained sequences of cards, being the right sequence to give them a greater chance at correctly guessing the next card.
I don't see any way for such correct subconscious identification and guessing of a sequence to work except through a PSI ability. If anything, remembering sequences that were part of the original deck order would tend to lead to incorrect guesses in the next deck order, resulting in result that would be no greater than chance, I'd guess.
In one game, Chinook [an extremely advanced checkers computer program], playing with white pieces, made a mistake on the tenth move. [Dr. Marion] Tinsley remarked, "You're going to regret that." Chinook resigned after move 36, fully 26 moves later. The lead programmer Schaeffer looked back into the database and discovered that Tinsley picked the only strategy that could have defeated Chinook from that point.
Marion Tinsley is widely regarded as the greatest checkers player in history, by far. I think his legendary feat in this game is probably less difficult than what was proposed in this instance.
Perhaps there is such an onus in your mind, but the scientific process (and I do not mean the versions of the scientific process created by scientismist pseudo-skeptics) places no such onus on my part. I am also under no onus to "scientifically demonstrate" that the sun rises in the east, nor (more apropos) that people can tie their shoes without thinking about it.
That people can do things like that has been overwhelmingly demonstrated. It is demonstrated every time anyone except the rankest beginner plays chess. It's demonstrated every time someone catches a ball. It's demonstrated every time someone drives a car. People don't consciously calculate every possibility in these activities nor do they plan the tensing and relaxing of each muscle necessary to accomplish them. I haven't proposed anything particularly skilled here -- nothing even 1% as difficult as counting cards in blackjack, for example. I'm proposing that someone might have gone with their natural inclinations and when that seemed to work, to have continued to do so.
It is up to the person making a claim to demonstrate that alternatives aren't more plausible. In this case, the claim is that a form of psi that has not been seen in over eighty years of parapsychological experiments took place in some casual tests. I say that in this case conventional explanations are more likely than this brand new phenomena. The onus is on anyone claiming that its unlikely that the results in this case could not have produced by people acting the way people generally do.
If you were to actually read those eighty years of experimental evidence you would find that one thing that has been found is that sloppy experiments frequently show strong affects, but those affects diminish when the experiment is tightened. Contrary to pseudo-skeptics, however, they do not disappear entirely, most of the experiments in the parapsychological literature are far from sloppy, and most of the weird supposed flaws claimed by pseudo-skeptics do not constitute sloppiness nor, if tested, lead to any change in results.
First of all PSI is a unit of pressure in the so called "English" system of measurements no longer used scientifically or in engineering (though still in use in the US in commerce). It is "ψ", generally lexicographically rendered "psi", that is the unknown explanatory principle or principles that underlie the phenomena that we study in parapsychology (and generally also used as an adjective indicating the phenomena themselves).
Sorry, this grates -- a pet peeve.
And second, the claim that it is "unknown to reality" (not even undemonstrated!) that people are capable of doing simple cognitive tasks without being consciously aware of it (i.e., subconsciously) is simply untrue -- countered by the fact that until I just thought of it as an example I was consciously unaware of how my fingers were hitting the keys on my keyboard, and I am still unable to access how thinking of a word gets translated into a spelling transcribed by my typing fingers.
Only PSI has been massively tested with great rigor that conventional psychology has almost never matched.
Hyperbole doesn't gain anything. Psychology routinely accepts results as established on the basis of astronomically less evidence than exists for psi phenomena. That doesn't mean that there aren't some things in psychology as well established as psi phenomena -- by dint of the immensely larger number of people conducting general psychology experiments than parapsychology experiments. I'm speaking of things -- characteristics of perception, cognition, memory, decision making, emotion, personality, states of consciousness, social and group dynamics etc. -- that have become foundational in further experiments and as a result get tested, over and over, again far beyond what anyone feels is necessary or even desirable to establish their truth. I'm also referring to some ground truths of psychology that can be established unambiguously by simple observation: people can see physical objects in their vicinity given the right conditions, people sometimes remember things accurately, people sometimes remember things inaccurately, people sleep, people generally like sex, people can be distracted by things, ... and people can effectively perform tasks without being consciously aware of what they are doing to make those performances effective.
Can you really say that these things have not been as well established as psi? Do you really believe that it is more thoroughly established that psi exist then that people can understand spoken language? Or to be more fair, that people are more likely to accept evidence that supports their existing beliefs than that which contradicts them?
The scientific study of PSI is entirely more rigorous in its testing of claimed phenomena.
The field of parapsychology, both because of the demonstrated characteristics of the phenomena itself and the social situation within science requires levels of rigor found in no other field (and I've looked). That doesn't mean that only parapsychology experiments are rigorous. Its about the need for some of that rigor to be able to say that something is due to psi that I'm supporting and that you are arguing against.
You keep arguing as if I were a psi skeptic rather than a parapsychologist. I can believe that magnetism exists without getting upset when someone says that a rock rolling down a mountain is probably not being drawn by a magnet -- especially if I note that the distance over which the force was exerted and the likely composition of the rock would have required a magnet more powerful than any known to exist.
The background here is that Irving Langmuir was a scientist who made a splash in multiple fields, winning the Noble Prize in Chemistry in the 30s. Mostly -- and this is telling -- his work was to go back and fill in the details of things that had been left incomplete when most of the research in the field had moved on to the exciting "frontier" research issues in the field.
He gave a talk at GE Labs (where he worked) in 1953 on "Pathological Science". The subject was a magic formula that could be used to discard research without having to actually bother with its facts or claims. (Somehow the magic formula failed to prevent him from throwing vehement support behind a rather transparently bogus weather modification scheme).
Some people in science were aware of the talks from word of mouth but its actual content was only known by the few who actually attended it until 1968. In that year, Hall did a transcription from recordings, filling in some incomprehensible bits on the basis of discussions with attendees of the talk. He also edited it into the form of a paper (transcripts contain ers and ahs and half finished sentences that the speaker retracts and corrects immediately afterward, etc). This was issued by GE as a technical report.
Although this still had limited circulation people who were interested could now request a copy and read it, so it became better (and more accurately) known.
In 1989 the paper was published in Physics Today, which is the house journal of the American Institute of Physics. It put forward a list of six characteristics of "pathological science". Anything that met a few of these characteristics could safely be dismissed without thought. In support of this he presented examples of what he considered "pathological science," including his memories of a few hours discussion with Rhine 20 years before.
I left it to others to comment on how bizarrely out of line his memories were with Rhine's actual experiments, knowledge and opinions (meeting criterion 4: "fantastic theories contrary to experience"), and how unlikely he would be able to recreate 20 years later, purely from memory, word for word dialog with Rhine (criterion 3: "claims of great accuracy").
Instead I wrote a long letter essentially pointing out that attempting to create a classification based only on positive exemplars is logically invalid. I then presented many examples of breakthrough, foundational experiments that fit his criteria -- that essentially his criteria essentially defined innovative, potentially revolutionary science (what his career had generally avoided). I hoped that they would publish enough of it to make my point, and that I had included enough examples that they could choose some that they liked.
To my surprise, they accepted the whole thing, requesting only minor revisions (mostly things like referring to "Albert Einstein" rather than "Einstein"). It, with other comments on the article, was published in early 1990.
Was the skeptic you were referring to Robert Parks who in 2000 published a book called "Voodoo Science" which apparently (I haven't read it) referred to "Pathological Science" extensively (I doubt with any reference to my rebuttal or any of the others)? If that is the case, then I should point out that characterizing Parks as a minor CSICOPS member is hardly a fair representation. Parks is a major skeptic, able to use his office as the Director of Information for The American Physics Society as a platform, as well as his long time "What's New" column (a news column about physics and science in general) for the APS and now on the University of Maryland website.
I did leave out the majority of the details and the context, I just wanted to mention an interesting highlight from one of the many experiments I've conducted, offhand in a short blog comment, without getting too verbose.
No offense taken. My friend and I have been studying parapsychology for years. Every single alternative you’ve mentioned, we’ve thought of long ago. I think we’re aware of all of the major sources of potential statistical bias and counter-theories to psi. Selection bias, subconscious bias, multiple-test issues, etc. We have collected mountains of data, under a wide variety of conditions, and while it’s nothing journal-worthy due to the informal nature of the data collection, and the small number of participants, (it’s me, my friend, and a couple of our close friends for the most part that have done the experiments), it’s been extremely valuable in *our* development and learning about psi phenomena. We’ve done everything from decks of cards, the Russel Targ ESP Trainer for iPod, other similar ESP games for iPod, dice, custom written programs, pseudo-RNG algorithms, and RNG run from a Geiger counter and a chunk of mild uranium ore.
The card-guessing game I mentioned was conducted in basically two contiguous blocks (We try to collect in contiguous blocks to avoid selection bias). The first block of 18 decks was conducted face-up, with feedback, so the possibility of detecting whether there are more reds or blacks in the deck is plausible, though I am not sure where you obtained the 58% figure. This resulted in 604 hits in 936 trials, 64.5%, z > 8, p < 1E-18, ES = 0.29. Even assuming an expected hit rate of 58% due to autistic-savant level, which we certainly considered because this friend is extremely intelligent like that, it’s still significant at 2.5E-05. It was exciting, but we couldn’t totally exclude the possibility that having feedback of the cards was affecting the hit rate in some way.
The second block of experiments was conducted face-down – the cards were sorted, face down the entire time, on a non-glossy surface, into two piles. After the deck was exhausted, the results were tallied. Shuffling in all instances is very though shuffle and riffle multiple times, at least 4-5. This resulted in a small decrease in effect size, 577 hits out of 908, 63.5%, z > 8, p < 1E-16, ES = 0.27. The exact p for greater than or equal to 577 hits of 908 is 6.08868E-17. I had to write a custom program in Python to solve for really small p values using the discreet equation, since most online programs don’t handle huge factorials or exponentials. Most of these are ‘close enough’ estimates, using excels NORMSDIST function.
(Part 1, continued...)
I definitely see your complaint about “looks like a duck,” and I agree, it does look exceptional, but that might be partially because I just picked one of our experiments with a really high z, without much context. My experimental statistics teacher would say that p value is only useful in telling you whether a set of data is statistically significant, or not, and you stop there. However, believe the other useful thing about p values (at least for Bernoulli binomial trials, and experiments with very discreet probabilities) is that if your experiment is valid, when you hit a certain p value such that you would have to run an experiment (a block of trials) continuously for multiple lengths of the age of the universe to get that as a spurious result, you don’t have to worry about the “sampling to a foregone conclusion” issue. If an experiment took someone a week to do, and hit twenty- or hundred-to-one odds, it’s possible that maybe they conducted a bunch of experiments, and put all the non-sig ones in a file drawer. But if you have an experiment that doesn’t have any gaps in the data collection from start to finish, and that hits odds ratios in the billions or trillions, that particular error becomes less of an issue. Obviously, no test in the world can really tell you if your experiment is valid or well-designed, this is just a way of addressing the complaint of, “Oh, you just sampled until you got the results you liked.”
Physical decks of cards have lots of issues, which is why we generally prefer to use something digital, such as the Russel Targ ESP Trainer app for ipod. It’s pretty much four colored buttons, and you have to press the correct one. It then gives you feedback if you’re correct or not. We’ve looked for patterns in the sequence of colors, ways to mess with the machine, etc. There is no way we’ve found to spoof it, and it avoids most of the issues with cards. Just flipping through my excel file of all the data, there is a single experiment that has 239 hits out of 744, 32.1%, p = 3.6E-06;
63 out of 192, p = 6.21E-03;
109 out of 432, p = 0.45, NS - that was an interesting failure (the subject said ‘their heart wasn’t in it’, and they were “conflicted’);
85 out of 432, p = 0.99 (that was a weird one, ‘significant’ in the wrong direction, but probably just chance)
181 out of 696, p = 0.46, NS
382 out of 1200, p = 2.3E-08;
1024 out of 3696, p = 7.3E-05
166 out of 624, p = 0.43, NS (the notes on that one are “frustrated” and “pissed off”)
And a bunch more but I think you get the picture. The hit rate varies depending on the type of experiment, the subject, how the subject is feeling (by far the biggest factor). But we tend to either see: 1) it appears that ESP is present, the hit rate is roughly stable at some value above chance, and so the odds go up the more trials that are conducted
2) ESP is not present during that experiment, odds are jack bupkis. It often correlates with the subject’s mood or disposition.
Which is exactly what one would expect if ESP were a real, mental phenomenon. We’ve taken good enough notes that we’ve actually constructed a rubric to determine whether or not the subject is in a mental state that is conducive to psi. That way, we can take a break from data collection so as not to dilute potentially ESP-demonstrating data with noise. Things like boredom, frustration, apathy, “not feelin’ it,” lack of focus, physical discomfort, and hunger tend to correlate pretty well with chance experiments.
Are you seriously saying that "memory" is an unproven ability that is only offered on an ad hoc basis?
Here is what I am saying:
1) It has been demonstrated that people can do subconsciously pretty much anything that they can do consciously. In fact very little of our decision making, perception and thinking is conscious at all.
2) Many magicians and others have learned to consciously memorize a deck of cards with a few seconds viewing. Mostly learning this is a matter of learning to pay attention and finding ways of tapping into the abilities of the subconscious. The trick seems to be more making the memories conscious than memorizing per se.
3) We are talking about a single incident with a single individual who might have learned to do this subconsciously. Whether or not this ability is commonplace is an entirely different issue. This is something that people can learn to do, and have learned to do, and that has been demonstrated to have been done subconsciously by some individuals (high level bridge players, whose choices have been shown to be biased in just the way expected from recognizing the inadequacy of most shuffling, but who had no conscious awareness that they were taking this into account in their decisions).
4) It doesn't require anywhere near perfect memorization -- just occasionally accurately remembering (presumably subconsciously) that the 4 of diamonds followed the jack of hearts last time and guessing red to generate an excess of hits.
5) The rapid production high scores is extremely unusual in rigorous experiments and commonplace in "casual" experiments that allow this kind of thing.
There are lots of possible conventional ways that these results could have come about, I threw out a couple for illustration. The reasonable conclusion is that while psi effects unquestionably exist the effects found in this test probably are not due to them.
And that pseudo-skeptics make ad hoc unjustified assumptions ("he could have stood on a stool in a public hallway peering in a transom that does not appear in any building plans without anyone noticing") to try to force a conventional explanation on a careful, rigorous experiment, does not mean that conventional explanations never produce the appearance of psi, nor that any proposal of such an explanation is ad hoc.
though I am not sure where you obtained the 58% figure
I probably could have calculated it with some difficulty, but I took the lazy path and ran a few thousand simulations.
rate of 58% due to autistic-savant level
I'm sorry, counting easily distinguished items up and down is not autistic-savant level: its normal seven year old level of skill.
Shuffling in all instances is very though shuffle and riffle multiple times, at least 4-5
A deck of conventional playing cards is not well mixed with fewer than seven or eight riffle shuffles. At 4 or 5 shuffles adjacent cards tend to stay adjacent or at least close.
----
I'm guessing that by "selection bias" you mean optional stopping (which can be seen as a particular type of selection bias). Avoiding this doesn't require contiguity. You can leave gaps of any length and interleave different serious however you want. The only requirement is that you decide in advance how many trials you are going to do (actually there are other valid stopping criteria, but that is the simplest and most common), and decide in advance of each trial what series it is part of.
-----
Inadequate shuffling can also produce an opportunity for conventional high scoring when feedback only occurs at the end of each run. I'll have to run some more complicated simulations (if I get the time -- quite a bit of work has piled up, unfortunately) to see if the ways that have occurred to me could produce those levels of scoring.
Please note that my comment was only about the single casual experiment you mentioned, not about the other more careful experiments you did. You cited astronomically high p-values in very short time period in a test that you described as "casual" to note the triviality of merely 5 sigmas. I questioned the relevance of an experiment that seemed unlikely to have been due to psi.
---
Custom Python program:
I developed a simple formula for extreme Z-scores that can be easily used with any scientific calculator. I'll write it up on my blog and, with Dean's permission put a heads up here.
That's a rather old-fashioned view of things. The whole idea of "significance" was developed to make it possible to create tables of reasonable size back before computers made it easy to compute the actual p-value. Significance still has some value as a rule of thumb to help decide whether something interesting is taking place, but the focus in statistics is now on the p-value itself (along with other quantities, of course, like the effect size).
The p-value is a measure of the amount of evidence you have that something other than chance deviations within the null hypothesis are operating. One use for that is, as you say, to judge how plausible it is that a result is due to the file drawer effect, but their are lots of other reasons to pay attention to it.
I haven't proposed anything particularly skilled here
I think you overlook the problems I pointed out.
- It would be quite a feat, in itself, to correctly remember an entire deck order on one exposure. Has that ability in the "subconscious" been demonstrated? Has that ability been shown to be widespread in the population or is it an incredibly rare ability? Has that ability been shown to include the I think something like that would be necessary for your proposed problem to be significant. And you leave out the possibility of the "subconscious" incorrectly remembering the deck order and the effects that would have on correctly guessing cards.
- There would be no way for a card guesser to know what parts of the original order, which they'd been exposed to exactly once, would survive a card shuffle. There would be no way for them to know the lengths of the retained fragments of order.
- Those retained fragments of the original order would be of unknown length, they would start at two cards but could not be counted on, reliably to be very long in any kind of adequate shuffle. The would soon end, perhaps as early as the third card up.
- Those fragments of that original order could not be relied upon to appear in the new deck order in the same relationship they appeared in the original deck order, especially if one or more cuts were involved in making the new deck order.
- I'll remind you that the card guesser would have been exposed to the original deck order once. In order for your proposed ability to influence the number of correct guesses a guess would have to correctly identify each card as being part of such a fragment of retained order from the previous run of the deck. Incorrectly guessing that the card being guessed at was part of a retained fragment of that original order would give no such advantage, eventually the retained order in the new deck order would break. I am not certain that the disadvantage that might give towards is knowable or even consistent but I'm fairly confident that your proposed mental faculty would tend to drive down the correct hit rate instead of increasing it.
The more I think about it the less credible it seems. I doubt that such an ability could a. exist as an ability, b. or that, even theoretically, it could be a significant factor in driving up positive hits in card guessing experiments c. that it would be at all reliable. As I said, I suspect it would tend to lead to incorrect guesses if it existed as an ability.
Presumably that ability wouldn't disappear during the second run and the "subconscious" would be just as liable to retain a memory of that second deck order. What would the allegedly retained memory of two different deck orders mean for using those differing memories in a third trial?
I'd say that, certainly by the third run, such an ability to remember card orders could considerably diminish the likelihood of making correct guesses. If anything it would be a hurdle that a proposed PSI ability would have to overcome, making a statistically significant positive result even more remarkable. The mathematical analysis of the problem taxes my limited knowledge of probability. Has anyone has ever analyzed this kind of proposal for those possible problems?
I wonder if, as Topher Cooper suggests, that such a (dis)ability exists, it could have some effect on the phenomenon of diminishing abilities during extended periods of card guessing.
debunk an experiment on the basis of a proposed mental ability that is not proven to exist, perhaps one never before proposed to exist
We aren't talking about an experiment, we're talking about an informal test.
No one has ever proven, and probably has never before proposed, that people have the ability of multiplying the value of pi to 15 places times the value of e to 15 places. But this unheard of feat is within the range of the abilities of most minimally educated people (if they are presented with those two values), many, many people can do much more complicated feats of arithmetic, some people can do it in their head, and a few people routinely do this kind of thing (and much more difficult) entirely in their head.
This is neither ad hoc nor untested. If you were unable to things of this kind and difficulty you would be unable perform the most elementary actions. You would certainly be unable to compose and type the notes you have been blessing us with (what! you claim to have memorized the spellings of thousands of words! Don't be ridiculous. Surly you must be consciously figuring out the spelling of each word, but using Pounds per Square Inch to determine when the infamous irregularities of English spelling apply).
Or, to say it simply (once again, but you haven't noticed the previous times): we are talking about a fairly simple application of a thoroughly demonstrated ability. Taking this into account is part of what makes experimental parapsychology so rigorous.
Certainly pseudo-skeptics are prone to inventing implausible explanatory abilities out of whole cloth, e.g., Richard Wiseman "debunking" the auto-ganzfeld by assuming that subliminal perception works with much greater accuracy than ever found in the laboratory working on sound orders of magnitude smaller not only than has ever been demonstrated by below thresholds physically possible according to QM.
This isn't such a proposal though. Its an ability well established by psychologists, and both accepted and proven by parapsychologists. If a skeptic were to point out the possibility of this flaw any decent parapsychology would say "oops" and either find a way of proving that it could not have influenced the results or withdraw their publication.
I think you overlook the problems I pointed out.
Nope, just treating them as already addressed.
However,...
The general kind of ability we are talking about, the ability to subconsciously recognize (the only way people recognize anything) some cards, and to subconsciously remember (the only way people remember anything) the color of the next card, without that recognition and association being conscious has been demonstrated with astronomical probability to be widespread among humans (actually universal except for the severely brain damage).
The ability to learn over repeated trials to apply an ability that has desirable consequences is universal throughout all multicellular life and for at least some unicellular life (if not all).
Whether circumstances necessary for someone to have learned to use this ability in this way is a bit hard to say. It certainly seems to be a specific instance of a much more general, widely useful ability (assume things haven't changed completely and act accordingly) so I would guess it wouldn't take much exposure to the specifics to bring it to the fore. It certainly has been shown, pretty explicitly, to be widespread in the population of expert bridge players.
But the issue is irrelevant -- we aren't talking about a test being given to the general population. We are talking about a single, apparently highly gifted individual.
But if we are talking about how widespread abilities are in the general population. It has been thoroughly demonstrated (the outstandingly rigorous and thorough experimental base of parapsychology that you thrust forward as if I were ignorant of it and as if it somehow proved that this particular test had to be valid, despite violating the rigor represented there and being inconsistent with the characteristics of the phenomena studied there) that the ability to produce hit rates via psi that were shown in this test are very clearly not widespread in the population -- at the very least it is very rare if not unique.
It would be quite a feat, in itself, to correctly remember an entire deck order on one exposure.
You're changing the subject: my comment that you are addressing was clearly and explicitly about the counting strategy not the memorization.
Addressing your comments anyway: yes it is quite a feat. It is done routinely by magicians, though, and the subconscious generally seems to be better at this then the conscious. In fact, memory is entirely a subconscious process -- the best the conscious can do is to find ways to encourage (by repetition, focus and reframing) the subconscious to memorize things.
More importantly, one doesn't need to memorize the whole deck, just a few cards. We're talking about explaining one extra hit out of 10 not a perfect score.
And you leave out the possibility of the "subconscious" incorrectly remembering the deck order and the effects that would have on correctly guessing cards.
Nope, I quite definitely took that into account (ignoring the scare quotes which seems to imply that you believe that people are fully conscious of all levels of their own cognition). It would be damaging to my case (and remember, this is only an example of how a casual test like this admits to non-paranormal explanation), only if scoring were quite close to perfect -- which it isn't.
There would be no way for a card guesser to know what parts of the original order ... would survive a card shuffle. There would be no way for them to know the lengths of the retained fragments of order.
Nope, but in the specific scenario I gave as an example, with trial by trial feedback, and a very simple way of taking advantage of it, they don't have to. They only need to observe the previous card.
(I elided a parenthetical that was a repetition of the previous point -- difficulty of memorization -- and had absolutely nothing to do with knowing the effects that the shuffle would have on the card order.)
Those retained fragments of the original order would be of unknown length, they would start at two cards but could not be counted on, reliably to be very long in any kind of adequate shuffle. The would soon end, perhaps as early as the third card up
Quite true, but:
1) We are not talking about an "adequate" shuffle, rather I was explicitly presupposing in my example (correctly as it turns out) a typical shuffle -- with too few riffle passes.
2) The rather simple method I spoke of only makes use of single card distance. The higher separation is irrelevant.
3) A more mathematically sophisticated method (which, however, would make use of a more primitive cognitive process -- general association rather than specific sequence memorization) could make use of longer range ordering -- but it would not be destroyed by the fact that the higher order effects are statistical rather than deterministic.
4) Card shuffling has been carefully studied, mathematically and experimentally, by the leading parapsychologist, Dr. John Palmer, and by the leading statistician, magician and pseudo-skeptic Persei Diaconis (the last role means we must be careful that bias hasn't interfered with his conclusions -- but they are, in fact, quite objective).
If a single, ideal riffle pass is made, the probability that any card will still be immediately followed by the card it was previously followed by is 50%. There is a 25% chance that both the following cards will be retained, and so on. The probability, however, that a card that was close to a given card will still be fairly close is close to 100% (its actually 100% except for cards that are close to the break). It takes quite a few passes to eliminate this.
But that is the ideal. Actual shuffling tends to drop much larger packs together than the ideal. Moreover, unless the deck has never been handled, there is a strong tendency for the same cards to stay together with each repeated shuffle. Packets near the top especially tend to stick together (one half-pack gets used up first, generally the one in the non-dominant hand which typically holds the bottom half-deck(which tends to let slip larger packets) leaving the remaining cards in the dominant hand to be dropped on the top in one piece).
If you haven't shuffled at least eight times -- more if you don't make a properly sloppy cut between each riffle pass -- you haven't done an adequate shuffle.
1) A single riffle shuffle perfectly maintains the order of all the cards in each of the two half-decks. It takes more passes than you would expect to eliminate this.
2) Cuts do not change the (cyclic) order of a deck at all. Multiple cuts are exactly equivalent to a single cut. The advantage of cutting between riffle passes is to compensate for some of the severe limitations in human riffle shuffling.
3) Since we are talking about a trial by trial feedback design, neither the order nor the position of the pairs (or packets in some other scenarios) is a problem. The identity of the last card guessed provides the information about what packet is relevant.
...[a long paragraph discussing the problems of attempting to gain an advantage by blindly calling the previous order of the deck]...
Since this addresses the problems with using a worthless strategy that bears no resemblance to what I was talking about (even if I wasn't as clear as I intended about, this bears so little resemblance to it that I don't think there could be any confusion), I can only agree entirely with the statement and ask how this is in any way relevant.
but I'm fairly confident that your proposed mental faculty would tend to drive down the correct hit rate instead of increasing it
If you mean it would interfere with the effectiveness of psi, then that could well be the case, though the experimental evidence for the effects of response bias on this kind of experiment is mixed to say the least.
If you mean that it would cause the hit rate to be worse than random then you are very definitely wrong.
Actually, you are really claiming that the test is invalid. If any non-psi strategy can drive down the hit rate than the non-psi strategy of always guessing the contrary of what the other strategy would recommend will reliably drive the hit rate above chance levels. Therefore, there would then be a conventional, non-psi reason for the observed hit rate and the test is debunked.
I prefer better to more thought myself. In any case that you doubt things that have been thoroughly proven (you appear to be sure that they haven't been proven because you doubt the consequences -- a clear example of complex subconscious cognition -- unless you are consciously engaged in begging the question) does not constitute a valid argument.
Presumably that ability wouldn't disappear during the second run and the "subconscious" would be just as liable to retain a memory of that second deck order. What would the allegedly retained memory of two different deck orders mean for using those differing memories in a third trial?
So now the ability to remember things that you doubted could possibly exist is now known by you to be obviously so strong that it is of indefinite extent?
OK, setting aside the implicit contradiction:
You seem to be claiming that any game or other activity, the depends on remembering recent history cannot be done reliably a second time immediately. Is it then your opinion, that no waiter in the history of the planet is capable of getting any but the first customer of the day's order right. That bridge and poker players would be fools to play more than one hand? That ER doctors when they go from patient to patient inevitably treat each patient for another patient's malady? That children playing the card game "concentration" must be world class geniuses (at the very least) to be able to play a second round?
No, I don't think you really think that, but why would this problem only exist for this single task.
Temporary memory (a.k.a., medium term memory) is an important part of the cognitive tools humans have at hand to deal with the world. Not all memory is long term (permanent) memory.
That isn't to say that there isn't a small problem. I must say that I get annoyed at myself when I play Ma Jong solitaire when I think that I have already removed a pair of tiles that I had actually removed in the previous game. However, I would not call myself a skilled player, and even so it is a minor annoyance rather than a severe problem in winning.
I wonder if, as Topher Cooper suggests, that such a (dis)ability exists, it could have some effect on the phenomenon of diminishing abilities during extended periods of card guessing.
Nope. Such a disability clearly exists, but psi decline would require long term memory would be involved, its one thing to have a somewhat reliable sense that the 2 of clubs was followed by previously by a red card. Its quite another to have every such impression be inevitably remembered years later.
Something similar (that does relate to long term, though not "biographical" memory) was proposed as a reason for psi decline by Dr. Charles Tart several decades ago. He thought that it was possible, or even likely, that learning worthless strategies on the basis of random hits, might drive out whatever internal strategies that successfully elicit psi ability.
------------
Anyway, I'm done. Parapsychology is often as badly damaged by some of its supporters as by any of the pseudo-skeptics. I have every hope that The Thought Criminal isn't one of them.
I'd be very interested in that program.
My program is very verbose, but it works well, it's accurate, and it's fairly fast. But I'd love something that works for scientific calculators.
I'd especially like you to address the inability to know which parts of the original order survived an honest shuffle - it's possible that not a single one did, the length of the retained sequences, the order those would appear in the new deck order.
You could include a point I didn't state very well, that, not knowing when a retained sequence ended, the next card or cards out of sequence might be more likely to be incorrectly guessed because of the mistaken expectation that it would continue.
Also, how many cards into one of these remembered sequences would a card guesser need to correctly identify the correct sequence. Any sequence would have to begin with two cards in a specific order, no sequence could even be suggested until the second, possibly the third card into a sequence. Presumably a number of retained sequences could begin the same way but not continue in the same way, not to mention new sequences that could begin the same way as remembered sequences but would end in different ways.
Your scenario seems to generate an unusual number of problems for the attempt to correctly identify retained sequences. I wouldn't think I'd identified all of the possible ways it could go wrong yet.
It is done routinely by magicians
If someone else is handling the cards? I'd like to know where this ability is documented when the shuffle is honestly done by someone else, with no control of the magician and the magician doesn't handle the cards. Of course I'm assuming an unmarked deck or, better, that the magician never sees the cards or has any other means of using deception or subconscious cues. Those would negate the presence - though not necessarily the existence - of your proposed faculty in the experiment. Of course, that's an assumption that magician "skeptics" have made to negate the psi faculties (you like that case better?).
No, I'm fairly confident that if a card guesser believed consciously or subconsciously that they remembered sequences from a previous deck order after a reasonably honest shuffle they would frequently be mistaken and it would be a hindrance, instead of a help in correctly guessing cards.
I still wonder if it's possible that actually remembering card sequences from previous runs or believing that those were remembered might help account for decreasing success as card guessing goes on. Though that's complete speculation on my part and I'd never assert it was likely without very careful and extensive research into it. Though I'm doubting such an ability really exists, though it's possible for someone to falsely believe they remember a sequence of cards at any point as well as "remembering" sequences that were never in the first deck order.
I really can't do anything about your lack of perception. I've addressed them at great length in multiple different ways. I don't know how explaining again can make any difference.
With all due respect, I suggest that you read what I wrote again. You seem to have got it into your head something entirely different than what I'm suggesting.
I said I was done, but against my better judgement, I'll address your sort of new points.
re: having an ineffective strategy would lead to hitting at a hit rate below chance.
This isn't new and I did address it but unlike most points I only addressed it once, and you were more specific, so...
If you mean that an ineffective strategy would interfere with psi functioning...
This would seem to be logical, but psi just doesn't work according to our expectations. Ineffective unconscious strategies (things like avoiding repeating a call) are known as "response bias". There were some studies that seemed to show that response bias interferes with psi hitting. Some other studies, however showed otherwise. In an "independent study" for college back during the reign of the Pharaohs, for example, I analyzed one of the classic card calling experiments (i.e., one that was not done specifically to study response bias) and found that the hitting perfectly conformed to the percipiant's response bias. I suspect the successes were due to an experimenter effect.
if you mean that an ineffective strategy will produce below chance hit rates statistically...
Nope. This is a classic statistical error. Its a version of what is known as the "Inverse Gambler's Fallacy" (the belief that since the Gambler's Fallacy is a fallacy, betting the opposite must be effective; the Gambler's Fallacy is the belief that, for example, if red has occurred on a roulette wheel frequently recently that black is more likely to occur next to "balance it out").
Ineffective strategies in predicting random sequences are ineffective, not "anti-effective". Otherwise, there would be an effective strategy of simply guessing the opposite of the ineffective strategy and you would have an effective strategy, and the sequence would not be random. In other words, if you are right here, then I am wrong about the specific strategy but right that there is an effective strategy and the procedure is flawed.
re: magicians able to memorize a deck.
You'll find it "documented" all over the magical literature. Harry Lorrayne, "The Memory Man" as well as a magician, demonstrated and taught it to the public. I personally know two magicians who can do it well enough to incorporate it (with "outs") into their acts -- and I generally don't run in the highest magical circles.
I'd be very interested in that program.
OK, done. You will find it at A Simple Formula For Extreme Z Scores (not my most imaginative title).
If you have a question, or need some help, post a comment there (identify yourself as The Raving Chemist -- I'll respond to anyone but I would like to know its you).
I wasn't talking about abstract strategies in the context of a card game, which is quite a different thing with a different goal.
I was talking about an ability to do what you asserted, to successfully increase correct guesses through an accurately remember the order of a deck of cards in the context of a card guessing experiment and being able to use that memory to successfully guess cards based on retained fragments of the original deck order after an honest shuffle done by another person producing a new deck order.
Guessing one card at a time in an adequately designed card guessing experiment is not the same thing as trying to remember cards dealt, many of those known, and trying to guess what opponents in a card game hold based on their decisions in the context of a game. What similarities you might be able to propose are more than made up for in differences.
I've thought of other problems for your claims since my last comment as well as quite obvious means for overcoming the possibility of those in an experment. Simply not telling the card guesser what cards were in the original deck order and not allowing them to see the cards would take care of it. Whatever "order" the guesser could remember would have nothing to do with the actual deck order and could only be based on a memory of what they'd guessed, a memory of their guesses which might, itself, be inaccurate. I'm assuming adequate intelligence and attention in the experimental design to insure information leakage couldn't be a plausible source of distortion to the results.
The requirement of rigorous control against all possibly eliminated sources of distortion in a experiment should not be allowed to, itself, become a distortion of the results. Any proposed skills or mental faculties shouldn't enter into that criticism unless those can be successfully demonstrated in the context of the experiment.
Criticism of an experiment should also serve reality instead of inserting irrational, impossible proposed flaws in order to deny what the data means. That allows what should be legitimate criticism open to being ideological debunkery. The invention of mental faculties whose actual existence or presence isn't confirmed is one of the ways that kind of distortion has entered into scientific and, even more so, popular scientific discourse.
I think popular debunkery of controlled science is as dangerous as what it is supposed to oppose, especially in the context of global warming and other environmental destruction, has become a major problem. It's way past time to take a very hard look at its practices and effects. You can look at both James Randi and Penn Jillette in that regard. I'm more than a little skeptical of the value of allowing people in their line of work, dependent on an ability to deceive, to become important figures in the popular understanding of what is popularly passed off as science on cable TV and the internet, not to mention some major newspapers and magazines.
I could also point out that what magicians and card-players do, they do consciously and deliberately, not "subconsciously".
Ah, where is the experimental evidence demonstrating the effectiveness of his teaching this "to the public" and that his method is retained for long?
We are talking about science dealing with mental abilities. I would doubt that one person in ten thousand could do that without training, I would doubt that most of his students could consistently do that even with practice. If I'm wrong, show me where the study is published. And, I'll point out, again, in this case it wouldn't be the result of a "subconscious" act, it would be the product of a rigorously, intentionally, consciously applied skill. So it really doesn't fulfill the scenario you propose.
And then show me where it is similarly documented that such a skill is of any practical advantage in guessing unmarked or unseen cards after an honest shuffle, with at least one cut, by another person.
I've always been annoyed with the idea that experimental subjects must be presumed to uniformly have the skills that professional magicians have to practice over and over again. And that the extremely remote possibility of someone in the general public, who likely never had the training to do what is asserted, has that ability must always be presumed. Where else in science is that level of control demanded?
Considering how much of the evidence in even the physical sciences could possibly be faked, it's rather important that the person making such an accusation that all possible avenues of fraud or error prevented, be required to back up their charge. The insistence that psi rigorously control for avoided the remotest possibility of a very skilled magician sneaking in and corrupting the data is an insurmountable and moving hurdle that isn't imposed on other science, certainly not on conventional psychology. I'd doubt that most of physics could stand up to that as a ground rule. You might look at the, I think far fetched, speculation that Eddington fudged his photo confirming Einstein as an example.
You say that in a properly designed experiment this is not a possibility, a statement I heartily agree with. But you apparently object to what is considered proper design.
You apparently believe that the only hidden power of the mind is psi, I believe in psi but believe, on the basis of both personal experience and familiarity with the experimental literature that people have amazing subconscious capabilities that they are unable to access consciously.
You believe that expert card game skill is not comparable in difficulty to what I am describing here. Again I agree -- skill at games like poker and bridge are much more difficult. An expert card player must be simultaneously taking many other things into account -- strategy, tells, bidding or betting strategy, managing ones pools, interpreting the play of the other players.
You believe that if most people don't have a skill then we do not have to be concerned that this particular person does not. I believe that even if most people don't have the ability to remember a few cards unconsciously, or to count up and down by ones, that a skill that some people provably have might be had by this one particular person. Especially, since we are talking about what is called in parapsychology a "special subject", i.e., someone who scores exceptionally highly. People take pride in what they do exceptionally well, which means that special subjects are very well motivated to learn to do well consistently. Even if we set aside the possibility of conscious fraud, then we must understand that an exceptional subject is almost certainly well motivated to unconsciously learn techniques that produce strong results. They certainly would not have any way of consciously knowing that that is what they are doing at least some of the time.
You believe, speaking metaphorically, that since there are grand masters, that those who appreciate art and want to understand it better, should not worry about whether a painting is a cheap forgery or a naive mis-attribution. If it appears to be "great art" to an untrained eye than it should be considered to be "great art" and one can learn more about great art (or about a particular artist or style or technique) by studying it. You further seem to think that any art expert who says that there are problems with any work, is clearly someone who dislikes art, and this silly talk about how easily work gets mis-attributed is just an excuse to avoid appreciating great art.
If you want to believe that, that painting you bought at a flea market is an undiscovered masterpiece by Leonardo Da Vinci -- after all it has the signature "Leo D" on it -- you're welcome to it. But an expert who says that this is unlikely really isn't being stupid when he says that that is unlikely.
Psi is, by definition, what is left over after conventional explanations have been eliminated. There are lots of things that look like psi, which are not -- that's not just in the heads of the pseudo-skeptics. Just as there is a difference between a properly signed painting found in a back room of the Vatican archives, and an atypical work found at a flea market in Sandusky, Ohio. There is a difference between a well done experiment and a "casual test". In general, a casual test means that one has to say that one can't be sure that psi is operating. If it furthermore shows results that are unusual for what has been shown to be psi, then one is justified in saying that there is no strong reason to believe that it is psi and a good reason to believe that it isn't.
Reliably testing psi is difficult. Reliably testing psi using cards is more difficult. Reliably testing special subjects is very, very difficult, especially with cards.
Now, I've just discussed things that we know can be done. If we don't eliminate not only explanations that we know will happen, but those that are physically possible, then we may be spending our time studying some unknown capability of human cognition, memory or perception rather than psi. Worse, we may confuse some actual psi with some conventional known phenomena, and with some conventional unknown phenomena. This will certainly lead to an incoherent mess (I'm not just speaking about scientists -- if you find psi interesting then you should care whether or not you are observing psi).
But that's me. You enjoy your genuine Renaissance Master.
I think the never ending, never to be satisfied, demands placed on PSI researchers who take reasonable precautions against information leakage, etc. and then many extraordinary ones in response to those never to be satisfied objections should not be used to automatically reject earlier research that was never reasonably taken as lacking validity through invalid objections.
I don't think psi researchers should remain in the business of re-proving the reality of phenomena they've massively confirmed. That is a waste of time and the extremely limited resources they have. Most people accept the reality of those phenomena, anyway and many people within science do, even if some of those won't admit it in public. When it comes to the ideological skeptics, a number of whom have a financial interest as well as their entire stock of public credibility dependent on any evidence being buried, there will never be enough evidence of the most massive persuasiveness to get them to admit they are wrong. I've said before that they are the modern day Cardinal Bellarmines, only he was actually far more reasonable.
They can do what they think best but I think it's time they declare victory and move on, faster and to present their work to the people who can take it, forgetting the pseudo-skeptics. Ignoring them will be necessary, anyway. Producing good work won't win them over. They've got to be marginalized.
I overwhelmingly agree that it is a waste of time to pure general proof oriented research on psi. It has been proven. Demonstrating that different manifestations of psi occur is still important. Also important is to study the characteristics of psi, and to study how psi may be produced more reliably or more strongly.
An even greater waste of time then continued attempts to prove psi (which at least has the benefit of providing the opportunity to observe psi phenomena) is to expend resources studying non-psi phenomena as if they were psi. This is not only a waste of resources but actually damaging to the enterprise of understanding psi (this is true both for formal scientific study and for the personal study). A chemist studying a particular chemist makes sure that she is actually experimenting with that chemical, and an ornithologist interested in a particular species of bird makes doesn't just say, "well its got feathers, and we are in the swamp so this must be an Ivory-Billed Woodpecker".
There are lots of things that can look like psi, many of them producing a much stronger effect and so, if present, guaranteed to wash out psi. That psi exists does not mean that frauds don't exist, that subtle sensory leakage has no effect, that statistical errors cannot occur in analysis, that subconscious cognitive processing is rendered powerless, that memory and reporting becomes perfectly reliable, etc.
If you are actually interested in psi, you have to make sure that what you are studying is psi. Nothing I have said in any way implies that psi needs to be proven to exist. But before you can study it, you have to prove that what you are studying can only be psi.
I said I was done, but this is an important point.
I wonder if your ability has ever been tested as being significant when an entire deck order is retained. I'd propose that a larger deck order be used, a small group of cards at the beginning of the test which would intentionally be rearranged to destroy any retained order, the number of which would be unknown to the guesser and changed in different runs, to be followed by the entire, retained deck order. I'd be curious to know if the guesser could have more success in "remembering" the deck order that they'd been exposed to once and if their success rate would go up after several such exposures to the same deck order.
Do you know if that has ever been done to test your proposed ability?
Are you attempting to prove that sometimes people don't have any awareness of things repeated to them over and over? That really didn't need to be proven?
In any case, proof is not of any consequence. I was giving an example of the type of thing that needs to be kept in mind in having any assurance that what looks like psi phenomena is. It happens to be something that has been proven to occur. But all that matters is that it is something that might take place. When there is something within the bounds of conventional science (especially something so thoroughly within those bounds, however sure you are that its impossible) that might be taking place than there cannot be any confidence that what is happening is outside the current understanding of conventional science.
If you care about psi this matters, if you don't care whether something is psi or not -- if it could be psi, it must be psi than you get upset when someone points out that there are (or perhaps just might be) non-psi phenomena that has the appearance of psi.
But I've said that several times too.
I really shouldn't keep getting sucked in. I suspect you are the only person who doesn't get it. If there are others they are no more likely to accept facts and logic than you. You've pointed out the (pseudo-)skeptics are not likely to be swayed by facts and reason (I do read what you say). This is true -- even if they reject conventional beliefs rather than alternatives to conventional beliefs.
I promise I will not respond further to The Thought Criminal.
I gave a suggestion for how that might be tested to see if there is any evidence, whatsoever, that such a faculty exists. If what you proposed is a "subconscious" ability is a significant factor in card guessing, it should show up in a test in which the exact card order is retained, with a small amount of distraction at the beginning of the repetition of the sequence.
I don't believe it is likely that more than a few people in a hundred thousand would have the ability to do that, though I'd certainly be open to being shown to be wrong.
Agree to disagree! Especially on the internet! :)
1. Would it help if you got multiple decks of cards, all properly shuffled, and used each deck only once?
2. If it's easy to get high significance with casual tests, then is it also just as easy to get minimal significance?
3. Topher wrote, "The simple subconscious strategy of remembering if there have been more reds or blacks dealt previously and guessing the less frequently seen color will give you a 58% hit rate." Can that be eliminated by having a combined deck and only using one quarter of that deck before changing to a new one of similar construction? Surely, if the subject is not aware of the results before the experiment is finished, that would eliminate the problem.
Though Criminal, you mentioned a friend who scored 64%. Has he tried guessing lottery numbers? That's a serious question!
It depends on the argument. If it is off topic, or immature, or repetitive, or ignorant, and if it goes on and on and on, it can make it hard to find and follow more interesting discussions by other posters.
Yes, that would help. There are, however, lots of possible experimental flaws that can result in spurious results -- especially when using a single "special subject". I didn't really know anything about the test except that it was described as "casual" and it apparently used an ordinary deck with a binary (color) choice. So I gave some examples of flaws that could have been present.
Rigorous tests can be done, though, even in card tests -- it's how Rhine first proved psi exists (and he did do so, according to any reasonable scientific standards) eighty years ago. One of the the strengths of Rhine's (and his group's and independent replicators') approach though, was not depending on special subjects. That inevitably a potential battle of wits that few researchers are qualified to compete in and none can ever be sure that they've won unless they have proven fraud and therefore lost. For example, I would be very surprised if the mentalist community hasn't developed a hacked version of the Targ Trainer. I don't think Russel ever intended it as a reliable tool for testing special subjects.
On the other hand...
Using special subjects, with all the safeguards you can bring to bear, given the potential for a strong effect, is a very useful way to look for patterns from which hypotheses can be derived to test with a more general population.
Its always easier to get "minimal significance" (by which I assume that you mean p=0.05 -- 1/20 probability of the given hit rate if only chance is operating -- rather than p=0.0) than high significance -- just throw out trials until your level drops to that point.
But if I said that it was easy to get such levels I only meant so in a relative sense. You either need someone who is good at picking up (consciously or not) such information (a special subject) or the right psychological environment.
And yes, we've made several attempts at the lottery. There are several problems we have been trying to engineer through.
1. Since accuracy is not 100%, you need to use error correction and redundancy in order to generate the correct numbers. You need to run X trials for every bit of data, where X is derived from the statistical power equation. The lower your native hit rate, the more trials you have to run.
2. Experiments have shown that precognitive effect sizes drop the farther you go into the future.
3. The more time you spent on a given ESP trial, the larger (on average, to a point) the effect size.
Put that all together, and there is basically a limit to how much data you can generate about the future. If you want to generate more data with any confidence, you have to do more trials. But when you do more trials (with a limited number of psi operators), you're going further into the future, which means your effect size drops, which means you need to do even MORE trials. You can try to limit the amount of time spent on each trial, but again, effect size drops, and you need more trials.
It works out for even a small lottery, like Win-4 (4 ten-digit numbers), you need thousands of trials in a 24 hour period. If you only have a few people generating data, you just can't achieve any predictive power.
But research suggests that associative remote viewing can be used to discern future events.
As I said, the problems I mentioned can be eliminated and might not have been present in any case (The Raving Chemist did let us know that insufficient shuffling -- less than eight riffle passes -- was used. If nothing else this invalidates the statistics, since they are computed on the basis of random trials. To what extent this is a real problem given all the circumstances of the test would need some work to investigate.) These were just a couple of possible flaws out of many that could have been present.
Your suggestion would weaken but not eliminate the problem. I'm not sure what you mean by a "combined deck" but lets take it to mean that four decks are shuffled together. In that case, the deck will contain 208 cards -- 104 red and 104 black. So the chance for either a red card or a black card on the first trial is 104/208 = 0.5. Say that a red card is the actual target for the first trial. Now there are 103 red cards and 104 black cards and a total of 207 cards in all. The probability of a black card, then is 104/207 = .5024 which makes a black card very slightly more likely than red on the second trial.
One alternative is to use an "open deck" which has been randomly selected so that each trial will have an exactly 50% chance of being red or black. Creating an open deck is straight forward but a real pain. For example, you can separately thoroughly shuffle 52 decks individually then assemble a single deck by taking a single card from each deck. You can't just shuffle that deck again for the next run -- you have to do the whole thing again using 52 complete decks.
The electronic equivalent of an open deck, however, is actually easier than a closed one which is one of the many reasons that electronic tests have almost entirely replaced card or other mechanical tests.
The alternative to an open deck is to not give trial by trial feedback, you only tote up the results at the end of the deck. This is what The Raving Chemist did with the second block of trials.
I think you are referring to The Raving Chemist -- it was the test where his friend produced this level that we have been talking about.
In any case psi has shown itself to be very sensitive to attitudes, feelings and expectations. People raised at least in part in social groups which sees money as to some extent "unclean" are likely to have some level of discomfort, whether they are conscious of it or not, in using their ability (or the ability they believe that have) to gain "filthy lucre". This attitude is prevalent in western Europe in general and England quite strongly. That attitude has carried over (rather paradoxically and ambivalently, given the exactly contrary Calvinist strain of seeing wealth as a sign of God's favor) into Anglo-American and from there general American culture.
It is very, very common to hear people involved with psychics, mediums, etc. that psychic abilities cannot be used for personal gain. That supposed use almost always involves at least psychological personal gain is not seen as a contradiction -- "personal gain" is mostly unconscious code for "financial gain".
All of which is to say, that there is a very good probability that a legitimate ability would be psychologically throttled from operating effectively on the lottery.
There are also some classic experiments that increasing motivation produces an "inverted U" in psi scores. Up to a point increasing the rewards for correct hits (financially or otherwise) increases psi scores under tight conditions, but beyond that point the rewards decrease -- perhaps because of stress ("choking"). (This is one of many characteristics of psi that none of the pseudo-skeptics theories come close to explaining and there are virtually no attempts by the pseudo-skeptics to address them).
There are other big differences between the lottery and most successful psi tests. Most notably and probably at least psychologically relevant is a rather long time between the effort and the feedback/reward. Non-conscious processes rarely deal well with delayed gratification.
Furthermore, although that 64% hit rate seems large it only represents less than 1 bit of information in per 17 trials. That means you would need 226 trials for any real chance to get a 4 digit "daily numbers" game and over 465 trials for a typical weekly lottery. That's at a minimum, communication theory says that with a high noise channel you need additional "error correction" bits to be able to reliably decode the message (i.e., you need information about what part of the "received signal" is the message and what part is just noise in addition to the information contained in the message). Given the high noise levels, I'm guestimating you would need to triple the number of trials to have a decent chance of hitting the lottery on the nose (it depends on the nature of how the noise mixes with the signal to determine what an effective way to encode this is and how many bits it would take, not to mention how you define "a decent chance", so guestimate is the best I can do). That's a lot of trials with no encouraging feedback over which to maintain interest and enthusiasm.
So the skeptics (not just pseudo-skeptics, this is a reasonable question) who wonder why a real "psychic" wouldn't just quietly amass a fortune and not waste time with "party tricks" and seemingly marginal uses of their supposed abilities have a reasonable answer if they are honestly interested in one.
Thank You.
I said:
Up to a point increasing the rewards for correct hits (financially or otherwise) increases psi scores under tight conditions, but beyond that point the rewards decrease...
I meant to say:
...but beyond that point the scores decrease...
I think that was my friend that I mentioned...
Obviously I agree with this post, since a response by saying basically the same thing. The primary difference is that for simplicity, I took as a starting point the hit rate you mentioned as a reliable indicator of the actual possible hit rate given similar conditions. In fact at best this is not the average hit rate but an exceptional one even for this subject, not to mention (oops just did) my obvious doubt that this represents psi hitting at all.
As a point of interest, though, I have a nit to pick on one statement, without that having any real impact on the point you were making:
2. Experiments have shown that precognitive effect sizes drop the farther you go into the future.
I think that that overstates the situation.
The major demonstration of this was a meta-analysis that combined the results of experiments that had not been done to gather information on this issue. That is, it looked at for example, precognition experiments whose targets were generated at varying time periods from the "calls" and found overall that shorter time gaps resulted, overall in higher scores.
This kind of meta-analysis cannot be used to reach conclusions -- they are "exploratory" and any results are suggestive rather than conclusive. The reason is that the experiments in which a long time gap may differ from those with short time gap in many ways that might be the actual cause of the difference. Even within a single experiment or series which was not designed to test this (and which was therefore designed to minimize other differences) there could be differences other than the time gap between those trials that were near future and those that were more distant future.
...
Pseudo-skeptics like Ray Hyman have cited discussions by statisticians on the inappropriateness of using this type of meta-analysis for inference as meaning that all kinds of meta-analysis is at best "only exploratory". I would say that the fault is not entirely with the pseudo-skeptics, since statisticians frequently just call either one "meta-analysis" and depend on their expert readers to understand what they are talking about from the context. But the pseudo-skeptics really shouldn't publicly publish absolute, unqualified statements about things they don't really understand.
In any case, I think that a major confounding factor in this meta-analysis is confounding the time over which precognition is operating and the time between call and feedback. Long term precognition trials cannot, by definition, provide short term feedback, and there is little reason for a short term precognition experiment to delay feedback.
On the other hand, there is the classic interpretation problem of "what is the target." The Observational Theories and their ilk, imply that the feedback is the only target that means anything. So perhaps feedback lag should be equated with precognition time, except that this introduces a clear non-psi psychological reason having nothing to do with underlying psi mechanism.
Much of the experiments that actually were designed to address this issue suffer from the same problem. That's in addition to the inconsistency of their findings (to be expected given comparisons of small effect sizes) as well as artifacts introduced by experimenter effects (psi mediated or otherwise), differential effects, expectations, and the possibility that the result might be mediated by an unknown external time-dependent factor (e.g., to pick a completely arbitrary possibility, what if precognition works best when the call and the target times are within a few hours of the same phase of the moon. Then short term would trials would generally meet that condition while longer term trials would only due so if the time difference were almost precisely an integer number of lunar months apart. Time might appear to be the critical factor, but it actually is only typically correlated with what is important: if the prediction was for something occurring exactly 2628 days 4 hours later (about 10 years but an even number of lunar periods) would get the same score as a prediction 10 minutes in the future).