Worst losing streak

GrandMaster

Ueber Meister
CAG
I was playing Vegas Strip BJ yesterday on autoplay. I managed to lose 25 units, which was my stop loss level, in 36 hands. I had a look in my playcheck log, and I found this losing streak including some hands before and after those 36.

52 hands, won 7 (including 1 BJ), tied 4, lost 41, total loss 35.5 units, 4.22SD below expectation. :( The probability of losing this much or more in 52 hands is about 1 in 120000 (including some correction for the small sample size). The chi squared value is about 26.9, the probability of the value of chi squared with 2 degrees of freedom being greater than or equal to 26.9 is about 1 in 700000. Can anyone top this?

My luck turned later and I made a small profit in the end.
 

Clayman

Dormant account
lanidar said:
And why Vegas Strip BJ?
I would guess because it has the lowest HA of all the BJ games offered by Viper.

GM - Nice going that you made a profit in the end. That you didn't bust out down 35.5 units, shows the value of a small starting unit that is less than 2% of starting bankroll.
 

cashroom

Dormant account
huh how did you learn all this?

i just heard loosing 30 something hands in a row equals 1 in 4 trillion or something like that. I've been messing up I think placing too high bets at the wrong times watching BJ tournaments on TV. But I played the Brandy vote promo and got it up from $33 to $800 something and wagered $38,000 but placed some $100 bets and lost a little sick of playing it after wagering that much. Hard to get that thing to $2003 played it tournament style
 

caruso

Banned User - repetitive violations of 1.6 - troll
Can anyone top this?
Small potatoes.

My last two Gambling Federation sessions combined came in at around five SDs down, approx. 200 units over approx. 900 hands.
 

GrandMaster

Ueber Meister
CAG
caruso said:
Small potatoes.

My last two Gambling Federation sessions combined came in at around five SDs down, approx. 200 units over approx. 900 hands.
Yes, but mine happened on software I believe to be fair.
 

Clayman

Dormant account
GrandMaster said:
The chi squared value is about 26.9, the probability of the value of chi squared with 2 degrees of freedom being greater than or equal to 26.9 is about 1 in 700000.
Not that it matters much, and mostly because I have little faith in most of my calculations, I get a chi-square of only 21.2 (~1 in 40,000). I used a W% of 43.31%, L%=47.89%, T%=8.8% to derive the theoretical number of wins, losses, ties versus observed of 7,41,4.

Cashroom - tournament BJ and regular BJ have almost nothing in common so I wouldn't bet that big in regular BJ if I were you. Of course, as always, the risk-of ruin you are willing to play with is entirely your own choice.
 

marcolino

Dormant account
caruso said:
Small potatoes.

My last two Gambling Federation sessions combined came in at around five SDs down, approx. 200 units over approx. 900 hands.
Why Gambling Federation can be considered?
 

GrandMaster

Ueber Meister
CAG
Clayman said:
Not that it matters much, and mostly because I have little faith in most of my calculations, I get a chi-square of only 21.2 (~1 in 40,000). I used a W% of 43.31%, L%=47.89%, T%=8.8% to derive the theoretical number of wins, losses, ties versus observed of 7,41,4.
Clayman is correct. I should stop trying to do chi square calculations late at night. :)
 

Clayman

Dormant account
GrandMaster said:
I should stop trying to do chi square calculations late at night. :)
Hey, if I throw in the 1 BJ and do it over 3 degrees, it's down to 1 in 10,000.

Now aren't you glad you can look forward to this happening again some time real soon? :D
 

jpm

Dormant account
caruso said:
Small potatoes.

My last two Gambling Federation sessions combined came in at around five SDs down, approx. 200 units over approx. 900 hands.
I hope that was the last time you played there too!
 

GrandMaster

Ueber Meister
CAG
Clayman said:
Hey, if I throw in the 1 BJ and do it over 3 degrees, it's down to 1 in 10,000.

Now aren't you glad you can look forward to this happening again some time real soon? :D
You need the expected value in each group to be about 5 or 6 for the chi squared test to be valid. It is already marginal with the expected value of the number of pushed hands, but the big discrepancy was not there, so it should be OK. The expected number of blackjacks is only about 2.5, that's definitely too low for the chi squared test.
 

Clayman

Dormant account
GrandMaster said:
You need the expected value in each group to be about 5 or 6 for the chi squared test to be valid. It is already marginal with the expected value of the number of pushed hands, but the big discrepancy was not there, so it should be OK. The expected number of blackjacks is only about 2.5, that's definitely too low for the chi squared test.
Right. I figured you'd say something like that. I was just being facetious.

At the risk of boring everyone to tears, what does the Yates correction do and would it be better in a situation like this?

Heck, while we're on the subject, what was the "small sample" correction you made in the first place?

If all this is too complicated to describe here, you know where to reach me!
 

cashroom

Dormant account
Gaming Federation paid me quick and hassle free

Ther are based in Costa Rica but the check came from Montreal. I received a free scratch card in the mail for $20 and cashed out $100. I was surprised I didn't have to go through some complicated speal. received check in a week.
Commodore is where I played
 

caruso

Banned User - repetitive violations of 1.6 - troll
jpm said:
I hope that was the last time you played there too!
LOL, it was.

Apparently they've now issued ANOTHER version. How many is that now? Three at least. Obviously this'll follow the same pattern as before: clean game, suck in the mugs, stick in the fix, clean out the mugs.

Be smart - stay away from Gambling Federation cheating scum.

EDIT: Ooops, thought this was the GF thread. Excuse my off topic remarks. Still, they bear repeating.
 

GrandMaster

Ueber Meister
CAG
Clayman said:
Right. I figured you'd say something like that. I was just being facetious.

At the risk of boring everyone to tears, what does the Yates correction do and would it be better in a situation like this?

Heck, while we're on the subject, what was the "small sample" correction you made in the first place?

If all this is too complicated to describe here, you know where to reach me!
The Yates correction is normally used with only one degree of freedom, typically when there are only two groups or a 2x2 table, or if the expected frequencies are between 5 and 10. You take the differences between the observed and expected frequencies and substract 0.5 from the modulus before squaring it. The chi square distribution with n degrees of freedom is just the distribution of the sum of the squares of n independent normally distributed random variables with mean 0 and variance 1. It is a continuous distribution used to approximate a discrete distribution. The Yates correction is supposed to provide a better fit and makes the chi squared test more conservative, reducing the likelihood of rejecting the null hypothesis.

My correction for the small sample size is nothing sophisticated, I merely included the 0.5 in the formula that I normally omit. :) My typical sample sizes are in the hundreds or thousands, the biggest error comes from not knowing the exact variance of the game.
 

jpm

Dormant account
caruso said:
LOL, it was.

Apparently they've now issued ANOTHER version. How many is that now? Three at least. Obviously this'll follow the same pattern as before: clean game, suck in the mugs, stick in the fix, clean out the mugs.

Be smart - stay away from Gambling Federation cheating scum.

EDIT: Ooops, thought this was the GF thread. Excuse my off topic remarks. Still, they bear repeating.
lol, good advice. I hope everyone takes it to heart.

I too got one of those wonderful 'everyone's a winner' scratch card in the mail the other day. What a shock, I won a free $20 (that I had to wager 20x or something ridiculous like that). Well, I played some multihand vp on their new & improved software and, as expected, lost it fairly quickly while hitting nothing but low end winning hands. Blew the rest of it out on nickel slots which still play the same as the last revision of the s/w. Aside from maybe a couple of new games, it looks the same, plays a little faster and has a better interface. Other than that, its a pure money pit. Abandon hope all who enter!
 
Top