Gambler's Fallacy

[Icewynd]

I was just reading something on the web picking apart lottery number analysis because it succumbs to the "Gambler's Fallacy", that is, the idea that since something has not happened recently that it must be 'due'. This idea is refuted by stating that the probability of each event (Lottery draw, coin toss) is exactly the same every time.

This argument is factual, but misses the point as far as I am concerned. It is clear from a quick glance at lotto stats that things move in cycles. For a while there will be more odds than evens, but then the cycle turns and there are more even numbers -- this is known as regression to the mean and is true for any distribution of numbers.

In lotto analysis I use a lot of moving averages and the numbers on all stats cycle from peak to valley and back again, within parameters that can be established over a sample of draws. When these averages are nearing an extreme one can be sure that the trend is about to change.

I don't think that this method of lotto analysis is part of the Gambler's Fallacy, but I am sure that there are other opinions out there (Colin?).

Lookin' forward to hearing your point of view.

 

[blitzed]

probability & other stats have many times led to me walkin out of the mini-mart with C-notes in hand

I'm still well out ahead of pick3/pick4 games...no matter what the pundits squawk about.

I'm ahead on the big lottery games because I pretty much quit after winnin enough chump change to break even heh! I'd rather play the dailies.

later,
blitzed

 

[CMF]

If it walks and quacks like a duck has a bill and webbed feet it probably is a duck.

I was just reading something on the web picking apart lottery number analysis because it succumbs to the "Gambler's Fallacy", that is, the idea that since something has not happened recently that it must be 'due'. This idea is refuted by stating that the probability of each event (Lottery draw, coin toss) is exactly the same every time.

This argument is factual, but misses the point as far as I am concerned. It is clear from a quick glance at lotto stats that things move in cycles. For a while there will be more odds than evens, but then the cycle turns and there are more even numbers -- this is known as regression to the mean and is true for any distribution of numbers.

In lotto analysis I use a lot of moving averages and the numbers on all stats cycle from peak to valley and back again, within parameters that can be established over a sample of draws. When these averages are nearing an extreme one can be sure that the trend is about to change.

I don't think that this method of lotto analysis is part of the Gambler's Fallacy, but I am sure that there are other opinions out there (Colin?).

Looking' forward to hearing your point of view.

You've made the usual assumption and then carried on as if it can be taken for granted. First, you must establish that the particular Lotto history for the game you are playing is more relevant than any other history for the same game. You will find that one history for the same game gives basically the same overall results as any other history for the same game but without synchronous highs and lows which are random - as is the history.

If the Lotto game pays on Odds or Evens or Sums then they are relevant. Have a look at the New York Lottery site where for Pick 3 they now pay on Sums. From my site, "The first question that should come to mind is, "How can they do that when the occurrence of the Sums is proportional to how many there are?" eg the Sums 0 or 27 only ocurr 1 in 1000 for 000 or 999. The answer is that they pay 28 different amounts depending on the odds varying from $500 to $6.50.".

You can see the full payouts here (Copy and paste the bit without the square brackets): -
http://www.nylottery.org/ny/nyStore/cgi-bin/ProdSubEV_Cat_400_SubCat_202130_NavRoot_300.htm

You can download a pdf article called A Mini Guide to Critical Thinking by Joe Lau here: -
http://philosophy.hku.hk/think/misc/miniguide.pdf

Colin Fairbrother

 

[Icewynd]

Colin,

Thanks for weighing in here, I knew you would have a point of view on this!

But, to boil the argument down to the most simple version, are you saying that a distribution of coin tosses would not establish reliable parameters (high and low values of a moving average) that could be counted on as turning points in the distribution within a high level of confidence? I guess I'm asking if you have seen evidence that the distribution would continue to push the high and low values of the cycle as the sample size increases?

If so, what sample size would you consider reliable? In the Canadian Lotto 6/49 we have almost 2800 draws, or almost 27 years -- a considerable number. I know you have worked with huge samples of lottery draws (6/49) combined from lotteries around the world, so I would consider yours to be an informed opinion on this matter.

Thank you for the link to the guide to critical thinking. I will be interested to have a read, but my initial impression is that you think that I (and everyone else on this board) need such a guide?

Lotto player to lotto player, best 'o luck to ya.

 

[CMF]

Jerome Cardan (1501 - 1576) Made sure his own death prediction came true!

Analysts do exist outside of Universities but those that would like their work to be taken seriously must abide by the rules of Critical Thinking (hence the article reference) and the work can then be judged on its merits.

I do not have the time or inclination to discuss something where the basic rules of rational argumentation are not adhered to. So, someone earlier boasting about being ahead in Pick 3 or Pick 4 with no evidence whatsoever and contrary to all my expertise in both those games, deserves to be ignored.

The realm you are delving in is basically the para-normal ie some psychic ability to ostensibly come up with something gainful from something that is quite separate to the process in producing results. An example is rubbing a crystal ball to tell someone's future.

Coin tossing has been extensively studied and there is plenty to find on the web. Generally, it is not difficult to have a run of 9 Heads or 9 Tails. After either one the odds of getting a Head or Tail are 50/50. This may mean that a run on one or the other is broken but this is an observation only not part of the tossing process.

The most famous 50/50 run occurred August 18, 1913 at the Monte Carlo casino when black came up successively 26 times. A movie has been made about it. After about 10 blacks interest was intense and after 15 blacks it was pandemonium. The casino made a fortune in just one night and of course punters lost a fortune.

Should the same situation occur in your life would you be certain of Red coming up on the 27th spin? Those that went with the flow without risking everything would have done alright on that occasion. Problem is you need the patience of Job to ever get the chance of participating and at every spin it is just as likely to go the other way.

Here is a link to an excellent article by John Walker: -
http://fourmilab.ch/rpkp/experiments/statistics.html

Colin Fairbrother

 

[blitzed]

hiya Colin, some people are adamant in their belief that lottery analysis is absolutely worthless...and I disagree, there is a lot to be said for using it to contain the draw numbers within say 50% of the number playfield or better.

unlike many pick3 players, I do not play each & every draw...nor do I dig Martingale holes...I mostly reserve play for what look like opportune times, sometimes get lucky, then bide my time waiting for more patterns to pounce on...if I have not checked recent draws in ages, I salivate upon seeing that for example the first column has been odd the last 6draws, 3rd column has been below 5 for last 6draws, 3 or more consecutive doubles and they are usually oversaturated, etc.

as far as the coin tossing, likewise I've seen either obverse or reverse landing for long consecutive stretches...however never very far into double digits.

cheers!
blitzed

 

[johnph77]

If the Lotto game pays on Odds or Evens or Sums then they are relevant. Have a look at the New York Lottery site where for Pick 3 they now pay on Sums. From my site, "The first question that should come to mind is, "How can they do that when the occurrence of the Sums is proportional to how many there are?" eg the Sums 0 or 27 only ocurr 1 in 1000 for 000 or 999. The answer is that they pay 28 different amounts depending on the odds varying from $500 to $6.50.".

Two things about this game came to my immediate attention - first, it is not a standalone game, but only an add-on - one has to purchase a "normal" ticket before this game can be played. Second - the payout percentages are consistent with other U.S. Pick 3 and Pick 4 games - 50% or slightly less. In the sums portion they range from exactly 50% (Sums 1, 4, 24 and 27) down to a low of 47.25% (Sums 11 and 17). Caveat emptor?

 

[rigged2lose]

The Gambler's Fallacy is correct and if we were to adhere by it's warning we would never gamble therefore we can never lose. However, we can't ever win either. Once we buy that ticket, we have become "a gambler" and as gamblers, we must throw the Gambler's Fallacy out the window. That way of thinking is for the non-gambler, the non-participant which we are not/no longer. We need to give ourselves the best possible chance of winning cause nobody wants to be a loser. This is assuming of course its a fair and honest game.

 

[blitzed]

Two things about this game came to my immediate attention - first, it is not a standalone game, but only an add-on - one has to purchase a "normal" ticket before this game can be played. Second - the payout percentages are consistent with other U.S. Pick 3 and Pick 4 games - 50% or slightly less. In the sums portion they range from exactly 50% (Sums 1, 4, 24 and 27) down to a low of 47.25% (Sums 11 and 17). Caveat emptor?

hiya John, yup most pick3 lotteries are locked down pretty tight, so I play frugally and at opportune times, otherwise I'd quickly dig a hole.

BTW, it is a shame that your site is down...cutting expenses?

cya,
bllitzed

 

[johnph77]

BTW, it is a shame that your site is down...cutting expenses?

cya,
bllitzed

No, the server upon which it was hosted caught fire - I'm having a little difficulty a the new host, but, thank goodness all the files may be intact. When I get some time it should be back up.

 

[blitzed]

No, the server upon which it was hosted caught fire - I'm having a little difficulty a the new host, but, thank goodness all the files may be intact. When I get some time it should be back up.

jeeez! hopeya can salvage it, great reference site...I kept intending to post here askin you about it, but forgot.

take it easy,
blitzed

 

[Icewynd]

Hmmmmm...33 hit last night just when it "had to" to avoid a new low. No, I'm not anthropomorphizing, just sayin' that these long-term high and low limits hold more often than not.

 

[CMF]

Will he win a fortune anticipating limits?

Ladbrokes offer the following odds for picking 1 to 5 integers in a 6/49 Lotto game: -
1 7: 1
2 61: 1
3 675: 1
4 7500: 1
5 150000: 1

So, if you bet $10 on an integer that hasn't appeared in 60 draws. Doesn't come up - bet again - lose - bet again - lose - bet again - lose - bet again - lose - bet again - lose - bet again - win. You break even!

Good luck - you're going to need it!

Colin

 

[rigged2lose]

Something doesn't seem right here. If Ladbrokes are truly giving those odds this may be a good chance for us gamblers to make some coin here. How are the winning numbers chosen? If they are chosen by Ladbrokes itself then we really have no chance but if they are taken from an independent source like say a National lottery it could profitable. I'm certainly interested, thanks.

 

[CMF]

Paying less than the odds means a bookie on average will win.

Here is the web address -
http://lottos.ladbrokes.com/en/how-to-play/odds.

You will be pleased to know the payouts are offered on national games.

Playing one line with 6 integers in a Pick 6, Pool 49 Lotto game your chances of getting 1 integer correct is 6 in 49 or 1 in 8.166. So, you bet $8.00 for 8 draws costing $64.00 and get 1 win which pays $56.00.

Au contraire, I think you are a really good opportunity for Ladbrokes to make some seriously good coin.

Colin Fairbrother