Winning Hand Count Distribution |
|
Posted Wed Mar 08, 2006 12:16 pm GMT by lwestatbus
Had a horrible time coming up with a Subject to describe this post.
I'm looking for a report or study on the distribution of the number of times a particular hand won out of a large number of hands (1,000 or more) at a 9- or 10-handed table. That is, if the report covers 1,000 hands it should read something like
High Card won - 18 times
Pair won - 76 times
...
Full House won - 82 times
...
Straight Flush won - 3 times
Ideally the report should cover actual play so that it will only account for hands that actually went to showdown and will not include those hands that were folded that would have won had they stayed in. So a simulation wouldn't work very well. It would be interesting to see a report that only reported on hands that went to showdown and a different one that reports on the winning player's hand even if there was no showdown.
The reason for the post is that I have been having what I think is an unusual occurrence of my 80%-win hands losing to better hands. That is, my very unscientific observations have generated a list of hands that I think will win at least 80% of the time or better. But I think I need to rely on some evidence now instead of opinion.
Thanks.
Did you know that participating in a poker forum can help you improve your own game? Be it by sharing experiences or simply asking for help, participation in a forum helps you focus and keep 'on topic' which will help you improve your game. You can learn from other players feedback and from their experiences. Why the THP poker forums? We offer one of the best managed texas holdem poker forums available, and the community within is far more friendly than those typicaly found on other sites. We've made a 'lurkers edition' of the poker forum available here on Holdem Poker Online, but we encourage all visitors to register and join in on the conversations on TexasHoldem-Poker.com
Posted Wed Mar 08, 2006 12:26 pm GMT by Jernej Zorec
ok my stats from 18k played hands at 1/2
Final Hand, went to showdown, won at showdown
High Card, 91, 14.29%
One Pair, 402, 41.54%
Two Pair, 392, 53.57%
Three of a Kind, 88, 78.41%
Straight, 67, 80.6%
Flush, 72, 86.11%
Full House, 69, 82.61%
4 of a Kind, 8, 100%
Royal Flush, 1, 100%
Posted Wed Mar 08, 2006 3:22 pm GMT by lwestatbus
Wow. What great information. This isn't quite what I was looking for but very, very useful and I appreciate you sharing this.
My understanding of your data is these were YOUR cards that you took to showdown and that of the 91 times you went to showdown with just high card you won 14% of the time. Correct?
The specific data that I am looking for is the rank of the winning hand regardless of who won and either the count or percentage of won hands for each rank. That is, if there were 1,000 hands played the count of all winning hand ranks should total 1,000. If by percent then the total should be 100 percent.
Many, many thanks, though, for this great response.
It is really, really interesting to see that 17.5% of the time a full house wasn't good enough. Thanks again.
Posted Wed Mar 08, 2006 3:59 pm GMT by MJJ
you can get what yuo want from his stats-
Just add the total # of hands reported then use it to convert to the format you want...
ie- he reports on 1204 hands, 402 times 1 pr won- 402/1204=33.4% 1 pair won
it would be more acurate w/ larger sample and/or if he had all hands won at showdown at the table (he is not exactly 'average', he is either looser or tighter) But, it is a start
Posted Wed Mar 08, 2006 4:28 pm GMT by Geno
1204/18000 = 6.7% hands played to the showdown. Who wants to comment on that?
Posted Wed Mar 08, 2006 4:36 pm GMT by Sean_in_NJ
| lwestatbus wrote: | | It is really, really interesting to see that 17.5% of the time a full house wasn't good enough. |
17% of the time that it went to showdown, it lost. You're missing the numerous times that the FH bets the flop, turn and river and wins without a showdown.
Not to mention the various ways that a FH in hold'em is vulnerable: two pair on board (top vs. bottom), trips on board (PP against smaller split pair), ad infinitum.
I'd be very surprised if any results you were able to gather where showdown is a requirement would be telling at all.
| Geno wrote: | | 1204/18000 = 6.7% hands played to the showdown. Who wants to comment on that? |
I don't know if my post covers that or not...
Posted Wed Mar 08, 2006 4:41 pm GMT by MJJ
quote="Sean_in_NJ"
I'd be very surprised if any results you were able to gather where showdown is a requirement would be telling at all.
quote
That's the only kind of results you could use...
There is nowhere to find out what the best hand was (or which handwas ahead) when 'player A' took down the pot on the turn. The only way to say that '1 pair wins x% of the time' is to use showdown results. The problem w/ the ones above is that they only show 1 players hands instead of hands for the whole table.
Posted Thu Mar 09, 2006 2:09 am GMT by Sean_in_NJ
| MJJ wrote: | That's the only kind of results you could use...
There is nowhere to find out what the best hand was (or which hand was ahead) when 'player A' took down the pot on the turn. |
If this is the case, then why are we even bothering with a sample at all? Pick two cards, pick a board of 5, and then find the number of two-card hands that beat it. There's your probability that a certain hand will win against the universe of possibilities. Compare that number to your results, and see how much of a statistical outlier you are.
If you just want to use showdown numbers, then you damn well better have an allowance for the times that a hand wins without a showdown. Are you NOT going to count Player A's royal flush as a winner when he's holding the A  , the board reads K Q J T and Player B folds on the turn? I'd say we can be pretty sure that Player A was ahead, and that Player B knew it. The only time this gets to showdown is when B has the 9 , and now your straight flush winning % at showdown is 50%.
Yeah, that doesn't sound right to me...
| MJJ wrote: | | The only way to say that '1 pair wins x% of the time' is to use showdown results. The problem w/ the ones above is that they only show 1 players hands instead of hands for the whole table. |
Well, what happens when both players at showdown have a pair and one loses on kicker? Doesn't 1 pair now only win 50% of the time? Or even 10% of the time if the other 8 players stayed in with other 1 pair hands?
Posted Thu Mar 09, 2006 2:23 am GMT by Sean_in_NJ
| lwestatbus wrote: | | I'm looking for a report or study on the distribution of the number of times a particular hand won out of a large number of hands (1,000 or more) at a 9- or 10-handed table. |
The probability of making a straight flush is 1 in 64,974. You'll need a LOT more hands than 1K to get a reasonable sample size.
Posted Thu Mar 09, 2006 9:25 am GMT by MJJ
| Sean_in_NJ wrote: | | MJJ wrote: | That's the only kind of results you could use...
There is nowhere to find out what the best hand was (or which hand was ahead) when 'player A' took down the pot on the turn. |
If this is the case, then why are we even bothering with a sample at all? Pick two cards, pick a board of 5, and then find the number of two-card hands that beat it. There's your probability that a certain hand will win against the universe of possibilities. Compare that number to your results, and see how much of a statistical outlier you are.
If you just want to use showdown numbers, then you damn well better have an allowance for the times that a hand wins without a showdown. Are you NOT going to count Player A's royal flush as a winner when he's holding the A , the board reads K Q J T and Player B folds on the turn? I'd say we can be pretty sure that Player A was ahead, and that Player B knew it. The only time this gets to showdown is when B has the 9 , and now your straight flush winning % at showdown is 50%.
Yeah, that doesn't sound right to me...
| MJJ wrote: | | The only way to say that '1 pair wins x% of the time' is to use showdown results. The problem w/ the ones above is that they only show 1 players hands instead of hands for the whole table. |
Well, what happens when both players at showdown have a pair and one loses on kicker? Doesn't 1 pair now only win 50% of the time? Or even 10% of the time if the other 8 players stayed in with other 1 pair hands? |
I understand your point sean; but if we use data from hands that don't go to showdown.... You should have a pretty high percentage where you win w/ nothing- (you bet on an unmade hand and everyone folds, etc). I think he is more worried about the strength of the hands. Also, there is no way to get info on hands not showed down (unless you can get one of the sites to release that info to you)
Your second point about 1 pr vs 1 pr/k- that doesn't mean 1 pr wins 50%... He wants to know in what percentage of total hands played that 1 pair would win- so if there were 100 hands played and 1pr won 12 of them it would be 12% (regardless of how many times 1pr lost to a better hand)
He wants to know what % of hands will be won by a certain hand
Not, what % of the time that certain hand will win...
Posted Fri Mar 10, 2006 3:13 pm GMT by lwestatbus
| MJJ wrote: | Your second point about 1 pr vs 1 pr/k- that doesn't mean 1 pr wins 50%... He wants to know in what percentage of total hands played that 1 pair would win- so if there were 100 hands played and 1pr won 12 of them it would be 12% (regardless of how many times 1pr lost to a better hand)
He wants to know what % of hands will be won by a certain hand
Not, what % of the time that certain hand will win... |
MJJ is correct. Maybe I should have posted the reason for my inquiry with the original post. I think that my made hands are getting beaten more often than they should be OR I need to adjust my perception of how good a hand is. As I play I get a feel for the hands that are winning in showdowns, but I don't write it down--maybe I should be. My sense is, though, that my winning percentage with two pair, for example, is less than the percentage of hands won by two pair. Hopefully you'll believe me when I say that I am aware of threats on the board, sudden activity from passive players, etc. I realize that two pair (again, as an example--this post isn't about two pair) loses value based on several observable factors and I adjust my play for them.
I saw in an interview once that in his early poker playing days Doyle Brunson dealt himself thousands of poker hands to get a feel for the distribution of hands. I'm looking for any reference to something slightly different but similar in intent.
Thanks to all for this lively discussion. Thanks also to Jernej for his great contribution.
Posted Tue Apr 11, 2006 1:52 am GMT by snoogins47
Well, I'm lazy and didn't read the entire thread, but...
...from about 28k played hands at at 5/10 and 10/20
Final Hand, went to showdown, won at showdown
High Card, 342, 23.39%
One Pair, 1362, 40.46%
Two Pair, 1005, 55.72%
Three of a Kind, 195, 75.38%
Straight, 200, 81.5%
Flush, 146, 84.25%
Full House, 150, 88.67%
4 of a Kind, 10, 100%
From about 3k hands of NL, between nl200 and nl600
High Card, 31, 9.68%
One Pair, 81, 28.40%
Two Pair, 63, 44.44%
Three of a Kind, 22, 72.73%
Straight, 10, 80%
Flush, 12, 83.33%
Full House, 11, 90.91%
I figure the high card/one pair stuff is so low mostly because of how often heads up pots get checked down after the flop, or after a flop bet here and there. In general, there seems to be a lot more of the "accidental showdowns" in NL than in FL. Bet the flop, get called, give up, and lose, or whatnot.
I wish that I had a more extensive database, but my computer is silly, hard drive wipes are the devil, and I go months at a time without paying attention to pokertracker.
|
|
