How likely is this?

jpsartre

Dormant account
Joined
Aug 7, 2004
Location
Denmark
I need some brainy people to give some comments on this recent run I had at BJ. What I've done is to look at the dealers end hand in the cases where he got to draw. In other words, the hands where I bust or got BJ are not included. The results are (with dealers end hand on the left and the number of times it happened on the right):

Bust = 34
17 = 19
18 = 20
19 = 14
20 = 27
21 = 29 (including 11 BJ).
 

EasyRhino

Dormant account
Joined
Sep 17, 2006
Location
San Diego
First, if you think you're getting cheated, you're probably not. Assuming you're not playing at a shop that is rogued, or uses an unknown software provider, it's highly unlikely. I've played many thousands of online blackjack hands and experience some ridiculous streaks. Hell, just last night, I was playing a few hands of casino hold'em. I got two full houses in a row. The odds of that happening are four hundredths of 1 percent. (Not complaining).

That being said, it does look like you got an outsized proportion of dealer 21's during that stretch. And it looks like you may have gotten less 19's than expected.

I feel your pain, I've played through stretches like that, and I start to wonder if maybe my mouse is broken or something.
 

gerilege

Meister Member
PABrogue
PABnorogue
Joined
Mar 9, 2006
Location
Hungary
News flash

I have only made some quick calculations, and I haven't checked my results.

The below numbers should be only upper bounds, as you ignored the hands when you busted, and I think it means that the dealer's starting cards are usually worse in those cases when you do not bust. I mean you bust more often when the dealer shows a high card, and you will never bust when the dealer shows 3-6, in which cases the dealer will bust more often. Therefore your sample is a bit biased, and the real probabilites should be even smaller with your sample, but in general, the probability of:

34 or less busts in 143 hands: 0.150796126
29 or more 21s and BJs...: 0.003175056
18 or more non-BJ 21s...: 0.018978694
11 or more BJs...: 0.074982838
14 or less 19s...: 0.126549641

Nothing unusual I guess.
 

jpsartre

Dormant account
Joined
Aug 7, 2004
Location
Denmark
Just what I was looking for. Cheers :)

And for the record, I did not think I was being cheated. Just thought I had a really bad run and was curious just how bad it was, statistically speaking.
 

jpsartre

Dormant account
Joined
Aug 7, 2004
Location
Denmark
Statistically speaking, your sample size is too small to draw any conclusions :p

I agree that you need a much larger sample to establish that a software is unfair, however, there is nothing wrong with smaller samples if you just want to see how unlucky you were. Betting $100 on red 5 times in a row and getting black each time will obviously not prove that the software is cheating but seeing that this will only happen around 4 time sout of a 100 I'd still say it was pretty unlucky :)
 

EasyRhino

Dormant account
Joined
Sep 17, 2006
Location
San Diego
And for the record, I did not think I was being cheated. Just thought I had a really bad run and was curious just how bad it was, statistically speaking.

Okay, I hear ya. I had a ridiculously bad run autoplaying over 10,000 roulette spins a while back. It was like 4 standard deviations off. In a way it was nice to know that I really was getting hosed. :)
 

EasyRhino

Dormant account
Joined
Sep 17, 2006
Location
San Diego
Oh wait, looked up details: 9999 spins, 1 unit each, and a loss of 432 units. Some other dude estimate 3.265 SD? It was at Casinoshare (microgaming) grinding through a ginormous wagering requirement, in April.

That's how I got my personal conviction that running quickspin and autoplay together on MG roulette is even a bigger gamble than you'd think.
 

aka23

Dormant account
Joined
Jun 11, 2006
Location
Planet Earth
Oh wait, looked up details: 9999 spins, 1 unit each, and a loss of 432 units. Some other dude estimate 3.265 SD? It was at Casinoshare (microgaming) grinding through a ginormous wagering requirement, in April.

That's how I got my personal conviction that running quickspin and autoplay together on MG roulette is even a bigger gamble than you'd think.
It is easy to estimate SDs for 10,000 spins on roulette with 1 unit bets. The sqrt of 10,000 is 100. The SD per spin is roughly 1. So 1 SD is roughly 100*1. Assuming you were playing French Roulette, that would make it about 3SDs off from the expected loss of 134.

I've found FR autoplay to be streaky, including at Casinoshare (actually more so at Casinoshare than elsewhere from my personal experience). There are several Casinomeister threads on this subject, including similar abnormal results. I have not found the odds to be distorted in favor of winning or losing... just streaky.
 

Users who are viewing this thread

Top