dominated hands winning 35% of the times

**kavelot** · 06-07-2008

hello

I'm playing tournments $3.25 (45 people) on pokerstars for a while, and started to notice that I was taking too much bad-beats when I had a dominating hand

Since "felling" that is not enough (we usually remember the bad-beats we take, but forget the ones we give), I requested my complete hands history and found my assumption is correct.

Here's what I did
From the sample I got, I extract JUST the hands played heads-up (ie, me vs someone else... no one else in the hand) AND when at least one of the players was all-in BEFORE flop (in case anyone is interested, I can share my script that do this job).

I found I had 158 situations in which it happened.
In 57 of them, I was the dominated. In 101 of them, my adversary was dominated.
I won 8 of the hands I was dominated (14%). My adversaries won 35 (34.65%) of them.

In fact I'm happy because it means I'm making better decisions then my adversaries, but isn't that out of order?

You can argue the data sample isn't big enough. I agree, so just trying to do an statistical test, I checked t-student
I believe in this situation, I would get (supposing a dominated hand will win 26% of the time):

me:
(0.26-0.14)/(2.77366/7.54983) = 0.32663

my adversaries:
(0.26-0.3465)/(4.86082/10.049875) = 0.17884

I just don't know what it means now,

I probably should compare this values to t(56) and t(100), but how? (I made this stat course long time ago)

can anyone help?

PS: I'm not thinking pokerstars has something against me, but I'm kind curious to at least convince myself I'm not that bad luck

**kavelot** · 06-07-2008

ok, I guess that's what I was looking for:

t-experimental = 2.889356
P(type 1 error) = 0.004414

so the probability of the mean being the same (in other words, the probability that I'm having the same odds when I'm dominated then when I'm dominating), is 0,44%. So I'm real unlucky!

**piv** · 06-07-2008

sample size buddy i'd want thousands of hands for any kinda significant data

**kavelot** · 06-07-2008

Originally Posted by piv

sample size buddy i'd want thousands of hands for any kinda significant data

I covered that. T-test says that the probability of this difference happening just by randomness is 0,44%

it's something similar than getting coin A and coin B
then you want to determine if they have the same probability of flipping head and tail

in order to do that, you flip coin A and coin B 50 times and gets:
coin A: head = 25 / tail = 25
coin B: head = 11 / tail = 39

the chances that they're the same are close to the ones that I'm getting the same odds when I'm dominated and when I'm dominating

if you wish more samples, you could flip 100 times and get:
coin A: head = 50 / tail = 50
coin B: head = 19 (instead of 22) / tail = 81 (instead of 78)

as you can see, for a difference as big as that, this sample size is enough

**piv** · 06-07-2008

if your calculations are correct (....) then it's entirely possible, its like a ~1/200 chance so it's not exactly out of the ordinary, you're prob just running bad, but you cant blame coolers for losing at poker.

**jharding666** · 06-07-2008

Can you just verify that by dominated hands you are talking specifically about scenarios such as AK v AQ, JT v T8, 99 v 97?

**kavelot** · 06-07-2008

Originally Posted by jharding666

Can you just verify that by dominated hands you are talking specifically about scenarios such as AK v AQ, JT v T8, 99 v 97?

by dominated hands I'm talking specifically about one player having cards AB and the other having cards BC, where A is different from C

I'm not counting AA vs AB with B < A because the probability of this winning is much lower

**MJPerry** · 06-07-2008

If dominated hands are winning 35% of the time (and we've established that dominating hands are AK v AQ, T9 v 98 etc) then that's only a 10% increase over a very small sample.

PokerForums.org

Welcome to PokerForums.org

Thread: dominated hands winning 35% of the times

LinkBack

Thread Tools

Display