Some Offsuit hand Winrates
 

I want to gather some stats on a few hands: K9o, Q9o, J9o, T9o. After I get sufficient responses I will post a reply with a weighted average of everyone's results. I want just the overall winrate with the hand, and only from 6 max statistics (you can combine all your 6 max levels). I also want the number of times you've had each hand. Thanks to everyone who takes the time to contribute.

K9o: 1677 times, -0.08 bb/hand
Q9o: 1708 times, -0.07 bb/hand
J9o: 1721 times, -0.06 bb/hand
T9o: 1678 times, -0.05 bb/hand

Re: Some Offsuit hand Winrates
 

What about KTo,QTo

Re: Some Offsuit hand Winrates
 

Hi Soko,

Ideally I'd like composite statistics on every hand. I am just starting off with 4 marginal ones to make it manageable for people to post a reply.

Re: Some Offsuit hand Winrates
 

K9o: 1426 times, -0.06 bb/hand
Q9o: 1408 times, -0.03 bb/hand
J9o: 1408 times, -0.09 bb/hand
T9o: 1402 times, -0.07 bb/hand

Re: Some Offsuit hand Winrates
 

10/20 6max: 117k hands

K9o : 1044 times, -.06 bb/hand
Q9o: 1064 times, -.06 bb/hand
J9o : 1083 times, -.05 bb/hand
T9o : 1041 times, -.01 bb/hand

5/10 6max: 144k hands

K9o: 1271 times, -.01 bb/hand
Q9o: 1310 times, -.01 bb/hand
J9o : 1295 times, -.10 bb/hand
t9o : 1316 times, -.06 bb/hand

I don't have time to combine those. So you can just use these as two seperate data sets if you want.

-Dave

Re: Some Offsuit hand Winrates
 

Hmm, it seems like you might want to have hands played outside of the blinds. If you include the times you have to pay .5 BB while on the BB, it makes the results harder to interpret IMO. If you only have the times people voluntarily play those hands, you will be able to see if people are overrating these hands, and playing them in EV- situations.

Re: Some Offsuit hand Winrates
 

you need to filter so that the blinds are not included. otherwise, all the hands you defend with have a -.5bb or -.25bb winrate to begin with. this is very important.

here are my stats filtered "not a blind"
k10 .21
k9 (.03)
k8 .21

qj .22
q10 .18
q9 (.01)
q8 .01

j10 .20
j9 (.05)
j8 0.0

109 .01
108 .03

Re: Some Offsuit hand Winrates
 

Hi Spicymoose,

You are right. I would have gained more info by asking people to post their winrates with each of those hands a.) out of the blinds and b.) in the blinds. I decided to keep it at overall winrate however, because

1. It's less work for the people replying. I didn't want to turn people off and would rather have tons of overall data then a few responses that are detailed. Having said that, any poster who wants to provide just their winrate on the button with those hands is more than welcome to. Go to the 'Position Stats' tab in PT to find this info (General Info won't combine your selected limits for position stats if you have more than one selected).

2. Overall winrate is correlated to playability outside the blinds. A hand that is playable for a steal on the button will do better in the big blind than a hand that is not playable for a steal.

Stats on the button:

K9o: 335 times, 0.03 bb/hand
Q9o: 329 times, 0.06 bb/hand
J9o: 310 times, -0.01 bb/hand
T9o: 312 times, -0.02 bb/hand

Re: Some Offsuit hand Winrates
 

Hi Victor,

Can you post the number of times you've had each hand? Just K9o, Q9o, J9o, and T9o is enough for now, and your overall winrate including the blinds for each of those. To combine everyone's samples I need everyone to post the same stat. I am ok with overall winrate for now since this gives me some good information. I can make a future post asking for non-blind or positional stats.

-v

Temp Results
 

I have samples from 3 people so far. We need many more for these numbers to be meaningful.

The results so far are:

***WINRATES FOR VARIOUS HANDS FOR 2+2ERS IN 6 MAX GAMES***

Hand #Times WINRATE(bb/hand)

K9o 5418 -0.05
Q9o 5490 -0.04
J9o 5507 -0.08
T9o 5437 -0.05

Re: Temp Results
 

[ QUOTE ]
I have samples from 3 people so far. We need many more for these numbers to be meaningful.

The results so far are:

***WINRATES FOR VARIOUS HANDS FOR 2+2ERS IN 6 MAX GAMES***

Hand #Times WINRATE(bb/hand)

K9o 5418 -0.05
Q9o 5490 -0.04
J9o 5507 -0.08
T9o 5437 -0.05

[/ QUOTE ]

I really don't get what you are trying to show here. I doubt any of these will end up being >0 for their winrate, and I imagine we could expect to see K9 more valueble than Q9 and so on.

Re: Temp Results
 

Hi Spicymoose,

What I really want to know is the strength of some of these marginal hands, and how playable they are out of the blinds. I picked offsuit hands because it's easier to get a statistically sufficient sample of them since they occur so frequently. The specific hands I picked, K9, Q9, J9, and T9, are one notch south of hands I KNOW are profitable overall in 6 max games.
Are these hands only playable on the button?
How much do they make on the button?
Can some of them open in the cutoff as well in typical games?
I suspect that any hand that you can open with from the cutoff as a default play should be at least breakeven EV overall (in 6 max games). My stats with all 4 of these hands are miserable, so I am interested in learning more about them.
Lastly, I enjoy looking at and pondering the statistics of hand values. I learned a lot about relative hand values from the EV charts at http://www.pokerroom.com/main/page/g...stats/expValue . The problem with that site is that you have the combined stats of everyone, including poor players, which skews the results of marginal hands.
If I had access to a hundred people's DBs I could look at all sorts of things, but I don't; so I have to come up with a simple question that I can get a high volume of responses on.

Re: Temp Results
 

[ QUOTE ]
What I really want to know is the strength of some of these marginal hands, and how playable they are out of the blinds

[/ QUOTE ]

then why the [censored] are you including the blind occurences?

Re: Some Offsuit hand Winrates
 

vkh,

K9o: 643, -0.02
Q9o: 630, +0.01
J9o: 601, -0.01
T9o: 601, -0.08 (I still like a button limp behind those 2 though [img]/images/graemlins/wink.gif[/img])

Jay

Re: Some Offsuit hand Winrates
 

small sample. but here it is anyway.

K9o: 501 times, -0.00 bb/hand
Q9o: 550 times, -0.04 bb/hand
J9o: 542 times, +0.06 bb/hand
T9o: 490 times, +0.02 bb/hand

Re: Some Offsuit hand Winrates
 

non blind hands

K9o 294 .15 bb/h
Q9o 286 .07 bb/h
J9o 310 .04 bb/h
T9o 299 .14 bb/h


including the blind
K9o 508 .03 bb/h
Q9o 485 -.02 bb/h
J9o 519 -.13 bb/h
T9o 458 .04 bb/h

Re: Temp Results
 

[ QUOTE ]
then why the [censored] are you including the blind occurences?

[/ QUOTE ]

I'll illustrate why with an example.

Let's say I want to know what my winrate could bewith K9o in the cutoff if I opened with it. Let's say that I've been dealt it 300 times in the cutoff and I've folded it 300 times. I'm making 0.00 bb/hand with K9o in the cutoff. What does that tell me about whether it's playable in the cutoff or not?

Absolutely nothing.

As I stated in another post in this thread, the strength of a hand outside the blinds is correlated with its strength in the blinds. A hand that is playable UTG for instance, usually breaks even or better in the blinds, whereas a hand that you muck first in on the button usually does pretty close to it's "loss floor" (roughly (-0.12)bb/hand in the big blind in 6 max) in the blinds. Since many people don't play the four hands I listed in the cutoff, and some don't even play them on the button (I have 130k hands of 5-10 as a significant winner without ever playing any one of those four hands on the button), I believe I will get a BETTER idea of their strength OUTSIDE the blinds by incorporating data INSIDE the blinds then by not incorporating this data.

Re: Temp Results
 

I disagree on several points.
1. There are enough people who do play K9-J9 in the c/o and button that you should be able to tell the potential that these hands have.
2. Blind play verses button play are two very differnet animals, and its not difficult for a non expert to be above average in one area and below in the other. Extrapolating that since sit. A is profitable that sit B with the same hand should also be profitable can lead people down a dangerous path.

Re: Temp Results
 

[ QUOTE ]
1. There are enough people who do play K9-J9 in the c/o and button that you should be able to tell the potential that these hands have.


[/ QUOTE ]

The people who play K9-J9 in the cutoff are a small subset of the people who play them on the button. And those people are a subset of the total population of players. On top of that, you have less than half the number of hands in the cutoff and button as in the blinds. So we are wheedling down the number of samples a lot by leaving out the blinds. You need tens of thousands of samples for winrate to mask the standard deviation of these hands, so I don't agree that there are 'enough people,' at least that will respond to this thread.

I agree with you on number 2. But I disagree that this invalidates the statistics if I incorporate hands from a large number of winning players. Maybe a few will play great in the blinds and horrible out of the blinds or vice versa but most winning players won't exhibit this bipolarity on an extreme level and with enough players this should be smoothed out.

Re: Temp Results
 

do you realize this?

if your winrate for a certain hand from the bb is -.45, then you are making money with it.

Re: Temp Results
 

[ QUOTE ]
do you realize this?

if your winrate for a certain hand from the bb is -.45, then you are making money with it.

[/ QUOTE ]

Of course I realize that. For you to ask me that question tells me that you did not understand my post. All I can suggest is that you re-read it.

-v

Re: Some Offsuit hand Winrates
 

K9o: 17 times, .25 bb/hand
Q9o: 23 times, .15 bb/hand
J9o: 43 times, -.14 bb/hand
T9o: 23 times, .08 bb/hand

Re: Some Offsuit hand Winrates
 

K9o: 198 times, -0.02 bb/hand
Q9o: 216 times, -0.06 bb/hand
J9o: 220 times, -0.03 bb/hand
T9o: 201 times, 0.00 bb/hand

So you are using these just as a figure of merit against other neighboring hands right?

Re: Some Offsuit hand Winrates
 

you need to take into account vpip as well.

Re: Some Offsuit hand Winrates
 

[ QUOTE ]
So you are using these just as a figure of merit against other neighboring hands right?

[/ QUOTE ]
No, he isn't.

I spoke to him about his intentions for the hands listed in this thread. Vkh is looking to combine all samples that are contributed into one giant sample size (sort of like pokerroom.com has). The biggest and most useful difference is that this sample will be made up of stats from only winning players as opposed to all players.

These four hands are just a start. If all goes well with these, he intends to continue developing a daily referrence that all of us can use.

Re: Some Offsuit hand Winrates
 

ok heres my totals:
hand..times..vpip..winrate
k9....1345...34....(.06)
q9....1362...22....(.07)
j9....1366...15....(.06)
109...1374...28....(.03)

Re: Some Offsuit hand Winrates
 

[ QUOTE ]
you need to take into account vpip as well.

[/ QUOTE ]
That will all come out as a wash. With hundreds of thousands of samples from winning players, everything that varies (VPIP, postflop play, etc.) will average out.

Re: Some Offsuit hand Winrates
 

K9: 833, +.09
Q9: 812, -.05
J9: 762, -.04
T9: 781, -.04

Obviously, the last 3 just haven't converged yet.

/mc

Re: Temp Results
 

[ QUOTE ]
[ QUOTE ]
then why the [censored] are you including the blind occurences?

[/ QUOTE ]

I'll illustrate why with an example.

Let's say I want to know what my winrate could bewith K9o in the cutoff if I opened with it. Let's say that I've been dealt it 300 times in the cutoff and I've folded it 300 times. I'm making 0.00 bb/hand with K9o in the cutoff. What does that tell me about whether it's playable in the cutoff or not?

Absolutely nothing.

As I stated in another post in this thread, the strength of a hand outside the blinds is correlated with its strength in the blinds. A hand that is playable UTG for instance, usually breaks even or better in the blinds, whereas a hand that you muck first in on the button usually does pretty close to it's "loss floor" (roughly (-0.12)bb/hand in the big blind in 6 max) in the blinds. Since many people don't play the four hands I listed in the cutoff, and some don't even play them on the button (I have 130k hands of 5-10 as a significant winner without ever playing any one of those four hands on the button), I believe I will get a BETTER idea of their strength OUTSIDE the blinds by incorporating data INSIDE the blinds then by not incorporating this data.

[/ QUOTE ]

I still don't see how you can possibly get any meaning out of these numbers. Blind play is way different from non blind play, and some people may only be playing T9 in the blind, and never out of the blinds. This will skew results one way or the other, and will not give you a "overall" picture. I really think you need to have seperate winrates for in blind, and out of blind play with the different hands.

Please, just give me an example of potential long run results, and tell me what those results would mean. For example, lets say K9o, Q9o, J9o, T9 lose -.04, -.07, -.08, -.09 respectively. What meaning do you draw out of that?

Re: Temp Results
 

[ QUOTE ]
Please, just give me an example of potential long run results, and tell me what those results would mean. For example, lets say K9o, Q9o, J9o, T9 lose -.04, -.07, -.08, -.09 respectively. What meaning do you draw out of that?

[/ QUOTE ]
I think I can answer that.

If we folded every one of those hands all of the time, our loss rate would be around -0.12 bb/hand. This is derived by taking the amount of both blinds (-0.75 bb) and dividing it by average players at the table (5.3), which is -0.1415 bb/hand. That number needs to be adjusted slightly in our favor in order to compensate for the times we get a walk in the big blind.

By playing these hands in profitable spots (which according to HUSH would be button, blinds and sometimes the CO), we are able, not so much get those hands in the green, but to reduce our losses.

By taking a collection of overall winrates for these hands from only winning players (breakeven to 3bb/100 winners), we can determine a relatively accurate hand value for each one. That information is extremely useful in determining where you or I should be with all hands in terms of winrate, and will also provide a more accurate guide for opening certain hands in certain situations.

Just entertain vkh for 2 minutes of your time and provide your personal stats. He's willing to invest much more of his time for our benefit.

Re: Temp Results
 

[ QUOTE ]
[ QUOTE ]
Please, just give me an example of potential long run results, and tell me what those results would mean. For example, lets say K9o, Q9o, J9o, T9 lose -.04, -.07, -.08, -.09 respectively. What meaning do you draw out of that?

[/ QUOTE ]
I think I can answer that.

If we folded every one of those hands all of the time, our loss rate would be around -0.12 bb/hand. This is derived by taking the amount of both blinds (-0.75 bb) and dividing it by average players at the table (5.3), which is -0.1415 bb/hand. That number needs to be adjusted slightly in our favor in order to compensate for the times we get a walk in the big blind.

By playing these hands in profitable spots (which according to HUSH would be button, blinds and sometimes the CO), we are able, not so much get those hands in the green, but to reduce our losses.

By taking a collection of overall winrates for these hands from only winning players (breakeven to 3bb/100 winners), we can determine a relatively accurate hand value for each one. That information is extremely useful in determining where you or I should be with all hands in terms of winrate, and will also provide a more accurate guide for opening certain hands in certain situations.

Just entertain vkh for 2 minutes of your time and provide your personal stats. He's willing to invest much more of his time for our benefit.

[/ QUOTE ]

I still don't get how any of this is usefull. You mention how you can get a loss rate per hand, but this is meaningless unless you know where it is comming from. A hand like T9o is rarely played from outside the blinds, so the majority of the stats you are getting are solely from blind play. On the other hand, K9o is a hand that is often played from the button, CO, and maybe even MP. Without having a weighting of how often these hands are getting played from what position, and in what circumstances they are being played (a blind defense EV is different from a blind steal EV is different from an isolation EV), I really just don't see what the use of knowing that K9o has an EV of -.07 is.

Re: Temp Results
 

[ QUOTE ]
still don't get how any of this is usefull.

[/ QUOTE ]

It's not useful. It's completely useless, irrelevant information. Knowing the overal expectation of a hand serves no purpose whatsoever, academic, practical, or otherwise. Please, don't waste any more of your time with this useless thread.

Re: Some Offsuit hand Winrates
 

K9o: 377 times, -0.04 bb/hand
Q9o: 448 times, -0.03 bb/hand
J9o: 402 times, -0.04 bb/hand
T9o: 469 times, -0.07 bb/hand

Re: Some Offsuit hand Winrates
 

K9o: 597 -0.04
Q9o: 580 +0.06
J9o: 567 +0.02
T9o: 542 -0.07

Re: Some Offsuit hand Winrates
 

Inclusive of blinds

K9: 3352 (.05)
Q9: 3480 (.09)
J9: 3459 (.05)
T9: 3403 (.07)

Temp Results II
 

We're starting to make headway -- about 1/8 of the way to getting 100,000 samples of each hand. Keep submitting data!!

***OVERALL WINRATES***

Hand Samples Winrate
K9o 13789 -0.04bb/hand
Q9o 14076 -0.05bb/hand
J9o 13988 -0.05bb/hand
T9o 13779 -0.05bb/hand

Re: Temp Results
 

[ QUOTE ]
[ QUOTE ]
still don't get how any of this is usefull.

[/ QUOTE ]

It's not useful. It's completely useless, irrelevant information. Knowing the overal expectation of a hand serves no purpose whatsoever, academic, practical, or otherwise. Please, don't waste any more of your time with this useless thread.

[/ QUOTE ]

Meh. Although I disagree with you, there could be something that you could draw out of your results. I just don't see what that could be, and I am asking you how you intend to use these numbers. I guess I will see when you finally complete everything. My main reason for putting so much pressure in this thread is because I honestly think that you will gain so much more if you seperate things out, rather than clumping everything together.

I may sound harsh, but I am honestly posting so that I can learn more, along with everyone else who may read this thread or results. If I am wrong about the clumping, then I am wrong, but I would like to know why I am wrong.

Re: Temp Results
 

Actually I apologize for getting sarcastic before.

I am just cataloging overall winrates with some hands. The end result will be a post that will contain the average winrate of these hands across a wide range of good 6 max players; a result that exists nowhere that I know of on the internet. It will be available as a reference for anyone who wants to see it by searching for this post. This clearly does not cause any harm. There is absolutely no reason for anyone to provide resistance to this thread. If you have an idea for some other statistical project then I more than welcome you to spearhead it.

Good luck,
-v

Re: Temp Results
 

you still arent taking into account vpip. why?

Re: Some Offsuit hand Winrates
 

blinds included:

K9o 1126 0.11BB/hand
Q9o 1100 0.02BB/hand
J9o 1114 (0.01BB/hand)
T9o 1095 (0.06BB/hand)