Do you believe in Luck?

[GlennB]

His skill isn't so terrible. In fact, his particular poker group has been running a year long competition and he's currently in the lead. That's probably why they give him such a hard time about it when he complains about his bad luck.

If he's playing face-to-face in a smallish group then book stats on the probability of given hands winning go out the window, as people learn quickly about the other players' styles. People can be 'read' and aren't even aware of the 'tells' they give off.

I'd say that recalling hand histories from many online games involving countless unknown players would be a much better test of his so-called 'luck'.

 

[Beth]

On the other hand, what do you tell someone when they collect data on a supposedly random outcome, using a number of different approaches, improving them over time and the results consistently show their outcomes being worse than expected by random chance?

That would be expected in about half the cases, wouldn't it?

Half would be expected to be below the expected value, half above. The p-value is the probability of getting the actual results assuming the null hypothesis (random chance) is true.

Let's say we have twenty coin flippers who claim to be unlucky. We'll define heads as lucky. They offer as evidence their last ten coin flips, when they got mostly tails.
What are the odds that any randomly selected member of the twenty will continue to get more unlucky tails in their next flips? I think it would be about one in two, not rare at all.

Yes, assuming a fair coin and unbiased flipping technique.

In other words, a larger sample size is needed. If one claims that Joe Blow is unlucky and correctly predicts his bad luck will continue, there will be about a 1:2 chance of being right. If one claims that these twenty people are unlucky and correctly predicts they'll all do worse than chance, and they do, the improbability of being right is starting to indicate something.

It would certainly be more convincing to have multiple individuals collecting data. However, this particular effort is for a single individual. His is the only conclusion that matters. He has done multiple data collection efforts spanning several years at this point. Sort of like a single coin flipper who repeats the experiment of flipping a coin multiple times, occasionally changing the coin and flipping technique to better guard against bias. Every single coin flip experiment has had fewer than expected heads, which was what he originally predicted and set up the experiment to test.

 

[carlitos]

[FONTImpressive. Although I might argue that Luck isn't necessarily a supernatural concept. For example, if someone wins a random lottery more often than others, there may be nothing supernatural about it. Nevertheless few would deny that they are lucky. In other words, it can be seen a mere description of circumstances rather than a supernatural event.[/SIZE][/FONT]

I don't think that we could get enough of a sample size to determine whether someone "wins a random lottery more often than others." I did see a 60 Minutes bit on a guy who thinks that his lottery winnings make him an investor, rather than a gambler. Maybe he's an outlier.

http://www.cbsnews.com/news/seven-time-lottery-winner-shares-secret-to-winning-powerball/

 

[Beth]

Beth, I have a question: is 'luck' strategic in that it only applies to live poker games as opposed to anything else? In other words, is there any logical reason to suppose one's 'luck' or lack thereof only manifests itself in a live poker setting?

He claims his luck at poker is similar to his perception of his luck in a more general sense. Poker has the advantage of being able to compute random chance expectations exactly. Other life events do not lend themselves to such an analysis.

If not (and I see no logical reason why it would), then why waste all that time collecting stats in live games?

Because it isn’t a real life situation, it’s only a simulation. Results would not be accepted as being applicable.

 

[Beth]

You should make it clear to yourself that random chance _does not expect_ that one hand-picked individual will get the middle score of the gaussian curve.

We are both aware of that. The problem is he’s consistently landing in the left tail; he’s never landed anywhere else.

 

[Beth]

It's never been clear to me, Beth, whether you understand that the hypothesis that past luck can predict future luck is a paranormal hypothesis, or equivalently, nonsense. Do you agree?

I don’t agree that it is “a paranormal hypothesis, or equivalently, nonsense.” If it were to be supported by the evidence, I would consider it perinormal, not paranormal and not nonsense. At any rate, I am someone who keeps an open mind about unexplained phenomena. I do not reject the alternative hypothesis a priori .

And if so, then why are you doing this "study"?

We are doing the study because I could not convince my husband that his perception of consistently poor luck was observational bias. I have had to eat those words. His data collection has consistently supported his informal observations that his luck is worse than would be expected due to random chance. Whatever might be the cause, observational bias on his part does not suffice as an explanation.

The null hypothesis is true by the laws of physics, and so any rejection of the null must be due to random or systematic error.

If you believe this, then experiments are not necessary. You have rejected the alternative hypothesis prior to testing. This is not to say that random or systematic error is not occurring, only that such errors need to be identified and established as the cause rather than presuming they must exist because the alternative hypothesis is ‘nonsense’.

On the other hand, if you disagree, can you explain why it is not a paranormal/nonsensical hypothesis?

No. I can only say that the empirical evidence he has collected to date strongly supports his hypothesis of his outcomes being worse than those expected due to random chance. The cause of those results is currently unknown.

 

[Beth]

If he's playing face-to-face in a smallish group then book stats on the probability of given hands winning go out the window, as people learn quickly about the other players' styles. People can be 'read' and aren't even aware of the 'tells' they give off.

That goes back to the data he is collecting, which was designed to eliminate skill to the greatest extent possible. Certainly the data on how frequently he is dealt the hands of {A, K}, {Q,8} and {5,2} would be unaffected by any level of skill he or the other players have.

I'd say that recalling hand histories from many online games involving countless unknown players would be a much better test of his so-called 'luck'.

The 2013 dataset is so heavily weighted toward online games that the live games have virtually no impact on the results. I did pull out and do an analysis on just the on-line games and the results were very nearly identical to the 2013 results.

 

[stanfr]

We are both aware of that. The problem is he’s consistently landing in the left tail; he’s never landed anywhere else.

Excuse me?? Then how did he get a winning percentage of .56?? Your bias is showing, Beth.

 

[stanfr]

He claims his luck at poker is similar to his perception of his luck in a more general sense. Poker has the advantage of being able to compute random chance expectations exactly. Other life events do not lend themselves to such an analysis.
Because it isn’t a real life situation, it’s only a simulation. Results would not be accepted as being applicable.

He is lucky to have you, no?
I don't see how your second statement follows...accepted by whom? It begs the question--so luck only applies in 'real' situations and not 'simulated' ones....howz that?
If you want a quick and robust test--just record EVERY 2 cards he is dealt.
Very easy to do (I know, I play online)
That will give you a large sample size pretty quickly. Since the cards can be ranked (albeit not perfectly in the mathematical sense) that will give you a reasonable idea of where his 'luck' stands. Surely you've thought of this approach--so why have you rejected it?

 

[Ambrosia]

I am a former online poker player, I spent a long time watching videos, reading books, practising, collecting and analysing data etc I think I'd played ~500k hands by the time I stopped.

I was an overall winner, but not a great one and while I can (could) comfortably beat 50NL for a reasonable income over time the playing standard improved and I wanted to play higher stakes and got crushed every time I tried.

So now I play once in a while for some fun and leave it at that.

As far as luck goes, here's my take on it.

There are 2 extreme samplesizes. i) an infinite number of hands. ii) one single hand.

We know 100% for sure that given an infinite number of hands that you will be 'lucky' exactly as often as probability says you will.

We have no idea what is ever going to happen in any single hand.

Any samplesize that a player plays for is a subset of i) the closer this subset gets to being i) the closer to average the results get, but as you can't ever get there the results won't either.

Different players have different subsets, and some will run hotter and some will run colder than expectation.

People call this luck. Maybe your fated to only ever experience the colder side of the curve. Maybe it's all completely random and all of us suck at understanding random, what with having brains that have pattern finding fetishes.

If you played long enough it would all balance out, but you only get so many years and can only play so many hands.

 

[stanfr]

I am a former online poker player, I spent a long time watching videos, reading books, practising, collecting and analysing data etc I think I'd played ~500k hands by the time I stopped.

I was an overall winner, but not a great one and while I can (could) comfortably beat 50NL for a reasonable income over time the playing standard improved and I wanted to play higher stakes and got crushed every time I tried.

So now I play once in a while for some fun and leave it at that.

As far as luck goes, here's my take on it.

There are 2 extreme samplesizes. i) an infinite number of hands. ii) one single hand.

We know 100% for sure that given an infinite number of hands that you will be 'lucky' exactly as often as probability says you will.

We have no idea what is ever going to happen in any single hand.

Any samplesize that a player plays for is a subset of i) the closer this subset gets to being i) the closer to average the results get, but as you can't ever get there the results won't either.

Different players have different subsets, and some will run hotter and some will run colder than expectation.

People call this luck. Maybe your fated to only ever experience the colder side of the curve. Maybe it's all completely random and all of us suck at understanding random, what with having brains that have pattern finding fetishes.

If you played long enough it would all balance out, but you only get so many years and can only play so many hands.

Interesting points! I have played online for a couple years, not a huge number of hands but a fair sample size. Like you, I'm better than even but not good enough to make a living at it (and it's frustrating keeping up with the software along with the possibility of cheaters...) One thing that has interested me about my 'luck' is that it does seem to come in 'streaks'--It seems like for a few games up to a few weeks worth I'll hit everything, full houses, quads right and left--then just as quickly ill go into a prolonged period of being dealt 26 and having my AK and JJ hands get crushed on the river. You'd intuitively think that 'luck would somehow be more 'even'--but that's not been my experience, it's stochastic and weird!

 

[Beth]

We are both aware of that. The problem is he’s consistently landing in the left tail; he’s never landed anywhere else.

Excuse me?? Then how did he get a winning percentage of .56?? Your bias is showing, Beth.

No, that's not evidence of bias. We aren't assuming a 50/50 chance of winning or losing after all. He's in the left tail because his computed expected probability of winning was above his actual percentage wins. For the number of hands involved, this difference (0.0191) had a p-value of .0672 which places it in the lower left tail of what was expected according to the null hypothesis of random chance.

He is lucky to have you, no?

Yes. And I'm lucky to have him too. We will be celebrating 35 years of marriage next month.

I don't see how your second statement follows...accepted by whom?

Accepted by my husband - the only person for whom these results actually matter.

Surely you've thought of this approach--so why have you rejected it?

Same answer as before - because it is not a situation that the person involved feels will replicate the results.

 

[MuDPhuD]

...whether you admit it or not, the hypothesis that people have an attribute called "luck" is a supernatural hypothesis, and no matter how statistically significant your results may be, the probability that those results are due to errors in the experiment are overwhelmingly greater than the probability that they are due to the supernatural hypothesis. If not due to random error, then it is almost certain that significant results are due to systematic error, and the smaller the p-values, the more that systematic error is the favored explanation. Convincingly small p-values of supernatural hypotheses give us an opportunity to learn how a well-intentioned experiment can go wrong.

^The above continues to be absolutely true.

He claims his luck at poker is similar to his perception of his luck in a more general sense. Poker has the advantage of being able to compute random chance expectations exactly. Other life events do not lend themselves to such an analysis.
Because it isn’t a real life situation, it’s only a simulation. Results would not be accepted as being applicable.

We are doing the study because I could not convince my husband that his perception of consistently poor luck was observational bias. I have had to eat those words. His data collection has consistently supported his informal observations that his luck is worse than would be expected due to random chance. Whatever might be the cause, observational bias on his part does not suffice as an explanation.
No. I can only say that the empirical evidence he has collected to date strongly supports his hypothesis of his outcomes being worse than those expected due to random chance. The cause of those results is currently unknown.

So the obviously biased observer is collecting the data? Are you really suprised by your results? I'm only surprised that the results do not deviate farther from expected!

His skill isn't so terrible. In fact, his particular poker group has been running a year long competition and he's currently in the lead. That's probably why they give him such a hard time about it when he complains about his bad luck.

So his perception of bad luck is in fact demonstrably wrong!

 

[Beth]

So the obviously biased observer is collecting the data? Are you really suprised by your results? I'm only surprised that the results do not deviate farther from expected!

Yes, actually. The idea is to validate or refute the bias with actual data. I have no doubts of the honesty of his data collection. While some mistakes are inevitable in writing them down by hand, the data from the on-line games does not suffer that flaw and shows the same results.

So his perception of bad luck is in fact demonstrably wrong!

No, that doesn't follow. Winning is a result of both skill and luck, so therefore cannot be considered a measure of his luck, bad or good. The idea of the data collection he's doing is to measure outcomes based solely on luck, without being influenced by the skill component.

 

[Strange famous]

I absolutely do believe in luck.

I think it is not some inexplicable thing, but maybe some mechanism a bit like the placebo effect.

 

[jt512]

I don’t agree that it is “a paranormal hypothesis, or equivalently, nonsense.” If it were to be supported by the evidence, I would consider it perinormal, not paranormal and not nonsense. At any rate, I am someone who keeps an open mind about unexplained phenomena. I do not reject the alternative hypothesis a priori .

We are doing the study because I could not convince my husband that his perception of consistently poor luck was observational bias. I have had to eat those words. His data collection has consistently supported his informal observations that his luck is worse than would be expected due to random chance. Whatever might be the cause, observational bias on his part does not suffice as an explanation.
If you believe this, then experiments are not necessary. You have rejected the alternative hypothesis prior to testing. This is not to say that random or systematic error is not occurring, only that such errors need to be identified and established as the cause rather than presuming they must exist because the alternative hypothesis is ‘nonsense’.
No. I can only say that the empirical evidence he has collected to date strongly supports his hypothesis of his outcomes being worse than those expected due to random chance. The cause of those results is currently unknown.

Yes, the laws of physics make your alternative hypothesis impossible a priori. You do not need to do experiments. Believing that people have a "luck" attribute that they carry around with them is physically impossible. I now clearly see that you fail to understand this and fail to see that you are testing a paranormal hypothesis.

Once again, if your experiment rejects the statistical null hypothesis, then all you have demonstrated is that your experiment is flawed. I've repeatedly listed several possibilities as to what those flaws might be. They are individually and collectively a billion times more probable than your paranormal "luck" hypothesis.

You think you're being open minded, but you're actually being closed minded. If you were really open minded, you'd have a chance to learn something about what makes seemingly well-designed experiments go wrong. Instead, you'll go on chasing your paranormal hypothesis and learn nothing. Maybe these people can help you.

 

[Beth]

Yes, the laws of physics make your alternative hypothesis impossible a priori. You do not need to do experiments.

How convenient. You know the truth and are able to do away with experiments. I am not so certain. I hope that you are not offended that rather than simply trusting in your expertise about that, we continue to collect data and determine whether the empirical data are as we expect according to the laws of physics and random chance.

Believing that people have a "luck" attribute that they carry around with them is physically impossible. I now clearly see that you fail to understand this and fail to see that you are testing a paranormal hypothesis.

But I don't believe that people have a 'luck' attribute. I am merely willing to test that hypothesis, which only means I do not reject it a priori.

Once again, if your experiment rejects the statistical null hypothesis, then all you have demonstrated is that your experiment is flawed. I've repeatedly listed several possibilities as to what those flaws might be. They are individually and collectively a billion times more probable than your paranormal "luck" hypothesis.

I have appreciated your contributions to the discussion in that regard. I have no problem with identifying and eliminating flaws as we identify them and are able to do so.

You think you're being open minded, but you're actually being closed minded. If you were really open minded, you'd have a chance to learn something about what makes seemingly well-designed experiments go wrong. Instead, you'll go on chasing your paranormal hypothesis and learn nothing. Maybe these people can help you.

I disagree. I am open-minded enough to try and learn something about why can make experiments go wrong. In fact, if you read this thread from the beginning, you might notice that we have incorporated suggestions to improve the quality of the data collection and analysis. I'm simply unwilling to presume that when experiments show results it cannot be due to the effect we were testing for. It's possible that it is experimental or data collection error, but such an assumption is not falsifiable. In fact, my dh says if he weren't collecting the data himself, he would assume it was such an error.

 

[dlorde]

You'd intuitively think that 'luck would somehow be more 'even'--but that's not been my experience, it's stochastic and weird!

I expect that's because we are generally intuitively awful at probabilities, and we expect random sequences and distributions to be smooth and even, not structured and lumpy as the actually are (as shown here).

 

[GlennB]

Out of interest, how do his stats look if you compare the specific 'race' conditions for the exact hands? 48/52 is a somewhat crude guide and you've only counted 54 races.

Suited overcards have an extra edge with flush chances, connected overcards ditto with straights, and very small pairs have extra ways to lose vs overcards such as when you hold 22 and the board comes 77553 and you lose to your oppo's kicker.

eta: using the holdem race analyser here I see that :

KJs is 46.16 to 53.46 vs. TT

KJo is 42.98 to 57.82 vs. TT

KQs is 50.99 to 48.11 vs. 22

JTs is a whopping 53.28 to 45.32 vs. 22

(they don't total 100 because of rare ties, and it varies slightly if the pair contains a card the same suit as one or both overcards)

I suspect your 50-50 assumption is having a significant effect on the results for such a small sample. Apologies if it's been extended in later posts but it's a longish thread.

 

[stanfr]

Same answer as before - because it is not a situation that the person involved feels will replicate the results.

Waah?
Why wouldn't it? Do you think it would be a fair test?
I'm beginning to see a pattern emerging here from your last few responses...originally I thought your statistics might be a bit off, now I think maybe you're just being fooled. Allowing your husband to collect the stats for example...he may be very well earnestly trying to mark down the results correctly, but his bias could influence how they actually are being written down. It's like the folks at the check-out counter--they're not consciously trying to cheat you--it just works out that the mistakes are almost always in their favor!