Yep, more gamblers fallacy for you. GOOD TIMES!

[kelly]

Im sure you do.

[MrJ]

This might be a first for me or maybe its not? Back to back exact days. Same number of hours played and exactly +$800. Hottest number in the last 22 spins. Wait, wait......it can't be so. Perhaps I should of left an $800 tip?     I dont deserve that money.

Ken

[amk]

[Skakus]

Yeah, that was it. My bad, sorry guys.

I should have been suspicious - if the simulation doesn't agree with the math, the simulation must be wrong.   

This explains why the shorter simulations didn't show any difference, and why taking more spins increased it. If the code was correct, it should have been the other way around. The problem with simulations is that they can only tell you what happens in the long run, which we know anyway because of the math. The math, by definition, can't say anything meaningful about the the typical session of a couple of hundred spins, or less. You can simulate 100 spins over and over, but then you end up back in the "long run".


Typical run gives Hot = 37137,   Cold = 36853   Not a significant difference.

Hey Bayes, thanks for further testing this with my twist.

I was actually expecting the hot numbers to still outperform the cold, and they did but not by much, eh.

Given a choice, I also would prefer to bet the hot numbers, but many players swear by the cold numbers so who are we to judge.

If hot numbers stay hot for statistically significant samples then the possibility of a bias would merit consideration. As Kelly said, in this case the cold numbers aren't going to heat up.

But the reality is that on a fair wheel the cold numbers will eventually heat up, and then to the casual observer those will actually be the hot numbers. Both the hot and cold numbers are aberrations from the norm and the norm only exists in light of the aberrations.

You can simply observe this regular behavior, and past behavior is a good indicator of future behavior, or you can try and do something with it. Mr.J has decided to encapsulate this behavior within a 22 spin cycle, and limit his focal point to 2 numbers. Just one way of many.

The thing is, if there were a bias present then from time to time whether by giving or taking it will contribute to these natural aberrations by way of hot & cold numbers. Therefore the chance of a player such as Ken sporadically melding his bets with the bias is strong.

With this scenario being a distinct possibility, why would you play cold numbers?

Besides, it makes no sense because you want them to hit, and hopefully a few times, so to win they need to get hot, don't they?

[MrJ]

"But the reality is that on a fair wheel the cold numbers will eventually heat up" >>> Absolutely they will......but not during my 22 spins. 

Ken

[MrJ]

Yeah, that was it. My bad, sorry guys. 

I should have been suspicious - if the simulation doesn't agree with the math, the simulation must be wrong.     

This explains why the shorter simulations didn't show any difference, and why taking more spins increased it. If the code was correct, it should have been the other way around. The problem with simulations is that they can only tell you what happens in the long run, which we know anyway because of the math. The math, by definition, can't say anything meaningful about the the typical session of a couple of hundred spins, or less. You can simulate 100 spins over and over, but then you end up back in the "long run".

Anyway, here's the corrected code:

Code Select Expand

include misc.e
include roulette.e
sequence wheel, spins, s, cold, hot, sc
wheel = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,
     21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36}
sc = {0,0}
cold = {}
hot = {}
spins = {}
constant SAMPLE = 600

procedure bet()
    sequence score
    integer x
    score = {0,0} -- initialise scores (hot is 1st element, cold is 2nd)
    while 1 do
    x = wheel[rand(37)] -- get next number
    if find(x, hot) then
        score[1] += 1
    end if
    if find(x, cold) then -- outcomes NOT mutually exclusive!
        score[2] += 1
    end if
    if score[1] = 2 and score[2] = 2 then -- a tie
        exit
    elsif score[1] = 2 and score[2] < 2 then  -- hot numbers won
        sc[1] += 1
        return
    elsif score[2] = 2 and score[1] < 2 then  -- cold numbers won
        sc[2] += 1
        return
    end if
    end while
end procedure

for i = 1 to 100000 do     -- 100,000 sessions of 600 spins
    for j = 1 to SAMPLE do
        spins &= wheel[rand(37)]  -- get a spin
    end for
    s = roul_freq(spins, 37)  -- get list of frequencies (sorted from hot to cold)
    s = reverse(s)
    for k = 1 to 3 do
        cold &= s[k][2]  -- get the 3 coldest numbers
    end for
    s = roul_freq(spins[$-46..$], 37)  -- get list of frequencies for last 47 spins
    for l = 1 to length(s) do -- find the hot numbers
        if s[l][1] = 3 then   -- number must have hit exactly 3 times
           hot &= s[l][2]    -- add it to the list
           if length(hot) = 3 then  -- exit loop if 3 hot numbers are found
              exit
           end if
        end if
    end for
    if length(hot) = 3 then -- only start betting if 3 hot numbers have been found
        bet()  -- goto procedure
    end if
    spins = {}  -- reset arrays for next session
    cold = {}
    hot = {}
end for
? sc -- print output


Typical run gives Hot = 37137,   Cold = 36853   Not a significant difference.

Hi Bayes. I was wondering in regards to my reply #8. Can you code/test this EXACTLY how I asked the question?

Ken

[Bayes]

Hi Bayes. I was wondering in regards to my reply #8. Can you code/test this EXACTLY how I asked the question?

Ken

Ken, I thought I'd done that - what did I miss?

[ego]

Fact and Fiction.

1. One thing roulette system player can not change is 37 or 38 degree of freedom.

I assume that pepole don't understand what that means.
As there is so littel understanding about the game and what happens when one dealer spin the ball.

If pepole would understand the physics they would see why there is no probability that dictate that just there number would show up in the future.
The common phrase regarding that is just that nothing is due to happen, because you have 37 or 38 degree of freedom.

2. The explanation is very simple.

a) When a dealer release the ball, he or she will do so with a different snap/force or could try to do it with similar behavior, it does not matter.
Same apply to when the dealer push the rotor to move with certain speed.

First the ball traveling time will be different every time from beginning to end.
Rotor speed will be pretty much the same from beginning to end of the ball spin or during that time - but rotor speed after rotor speed will not be the same.
So we have two moving parts witch is uniq every time or session.

b) Then at the end of the spin the ball will strike different deflectors and jump or scatter random on the rotor before it lose its total force to rest in one pocket or one winning number.
Each time rotor position or numbers will be at different places deepening on speed and witch deflector that strike.
During this time you have four random factors that will occur each time the dealer spin the ball - so there is no way to tell what the next number will be, ever.

That is what your "Gambler Fallacy" is about or why your past result does not matter as its only about if you can "F u c k Lady Luck" from behind or not.

Cheers

Here is a question I asked at the Wiz site. I can always tell when its a good question based on if I get few responses. People are so inconsistent when it comes to GF.

A) All 38 numbers are in a hat. I blindly pick one number, its the #17. I go to the casino and make a small bet on the 17, for ONE spin, never bet again and then went home. Did I just use gamblers fallacy?

B) One day later I change my mind, went back for ONE MORE SPIN of the 17. Did I NOW use gamblers fallacy?

Ken

[MrJ]

There are VERY FEW guarantees in roulette, something that is 99% a SURE THING, even the Advantage-play (cough) goofs can't deny this >>> In a short period of spins.......25?....37/38?.......there will be number(s) that will get 3 hits on it. Why not take this 99% fact and *TRY* to make something of it? Now don't even tell me, a person witnessed every number hitting in 37 spins. (LoL)

I know the anti-method AGENDA is strong but come on.  Look how many methods there are out there (including my own) that are NOT 99% guaranteed something will happen. If a person really has their s**t together, they can come quite close to having a Hol. Gra. with repeaters (it must be flat betting, no progressions).

Ken

[MrJ]

@Bayes >> I guess I don't understand these 2 comments.


"I wasn't taking due account of tie situations, which may be biasing the results"

"This explains why the shorter simulations didn't show any difference, and why taking more spins increased it"

(Also, why cant anybody PM you?)

Ken

[ego]

Fact and Fiction.

1. One thing roulette system player can not change is 37 or 38 degree of freedom.

I assume that pepole don't understand what that means.
As there is so littel understanding about the game and what happens when one dealer spin the ball.

If pepole would understand the physics they would see why there is no probability that dictate that just there number would show up in the future.
The common phrase regarding that is just that nothing is due to happen, because you have 37 or 38 degree of freedom.

2. The explanation is very simple.

a) When a dealer release the ball, he or she will do so with a different snap/force or could try to do it with similar behavior, it does not matter.
Same apply to when the dealer push the rotor to move with certain speed.

First the ball traveling time will be different every time from beginning to end.
Rotor speed will be pretty much the same from beginning to end of the ball spin or during that time - but rotor speed after rotor speed will not be the same.
So we have two moving parts witch is uniq every time or session.

b) Then at the end of the spin the ball will strike different deflectors and jump or scatter random on the rotor before it lose its total force to rest in one pocket or one winning number.
Each time rotor position or numbers will be at different places deepening on speed and witch deflector that strike.
During this time you have four random factors that will occur each time the dealer spin the ball - so there is no way to tell what the next number will be, ever.

That is what your "Gambler Fallacy" is about or why your past result does not matter as its only about if you can "F u c k Lady Luck" from behind or not.

Cheers

vs.

(Advantage-play is for suckers) Gambler's Fallacy is a term coined by unsuccessful gamblers to validate their reasons for losing.

What a joke.

Cheers

[Bayes]

"I wasn't taking due account of tie situations, which may be biasing the results"

Ok, suppose after collecting the hot and cold numbers, they were:

hot = 13, 21, 7
cold = 34, 13, 17

The first program (the one which gave the hot numbers an apparent advantage) would finish the session immediately when a group (hot or cold) got 2 wins, but what happened when number 13 hit TWICE?

The first time it hit, the program recorded a win for cold and hot (so the score was 1 - 1), but when the number 13 hit again, HOT was recorded first, and having reached the "winning" score of 2, the session would end. But of course, this wasn't really a win for the hot numbers - it should have been a TIE.

If you were recording results manually, you wouldn't think twice about this, but computers are VERY dumb, so if you don't think about what's happening and tell the computer to do it, that's when mistakes get made.

"This explains why the shorter simulations didn't show any difference, and why taking more spins increased it"

This relates to the first point about the tie situations. Because it's a relatively rare event that you get 2 (or more) numbers the same in each group, then a tie won't happen that often, so in a smaller test it's unlikely to show up in the stats (the randomness will swamp out the effect), but in a much longer test the outcomes will tend to stabilise to the true result. Nothing strange about that; it's why you need to test systems over many spins before you can be confident it's not just luck.

(Also, why can't anybody PM you?)

Sorry, I'll add you to my "buddy" list. Nothing personal, there was time a while ago when I was getting a lot of pms and I just got tired of reading them, so I cancelled all pms. 

[MrJ]

Anyone who disagrees with this, please chime in >> In 37/38 spins (most likely less than that) we will have number(s) with 3 hits on it. Anyone disagree?

Ken

(thanks Bayes for your response)

[macduff]

Yes i disagree,its more like 36 numbers you will get 3 hits. This is a fact not a faint hearted notion. 

[MrJ]

Yes i disagree,its more like 36 numbers you will get 3 hits. This is a fact not a faint hearted notion. 

?? Thats what I said. So *WITHIN* 37/38 spins (or less) we will get number(s) with 3 or more hits. So this is a FACT, correct?

Ken