Two Plus Two Older Archives  

Go Back   Two Plus Two Older Archives > General Poker Discussion > Poker Theory

Reply
 
Thread Tools Display Modes
  #1  
Old 10-09-2004, 04:31 PM
srt19170 srt19170 is offline
Senior Member
 
Join Date: Sep 2004
Posts: 272
Default Using Past Performance to Make Showdown Decision

Pokertracker says that when I've played a Pair to the showdown I've won 26% of the time. Suppose I now turn around and use that number to guide whether or not I call a bet on the river. I expect to win 26% of the time, so I call a bet when I'm getting 3:1 pot odds.

If that turns out to be too liberal, my Won % will start going down. That, in turn, will force me to higher pot odds. E.g., if my showdown % drops to 10%, then I'm going to want 9:1 pot odds to call a bet on the river. Similarly, if it is too conservative, my pot odds will fall.

All other things being equal, this should eventually settle out to the break-even point.

Any thoughts on this idea?

-- Scott
Reply With Quote
  #2  
Old 10-10-2004, 04:16 PM
Position Position is offline
Junior Member
 
Join Date: Nov 2003
Location: San Diego, CA
Posts: 20
Default the general idea


I like your basic idea. But, that two-pair-type data sounds useless for two big reasons: 1) that win-percentage # is MOSTLY times when you weren't even calling the river (because you were the aggressor or you just checked it down), 2) even if you did filter the data to just where you called the river bet, the texture of the board & the specific opponents' personalities are much more important than the general quality of your own hand (pair, two pair, flush) taken alone out-of-context.

I've long been using your basic idea, though, in NL play: I keep an eye on my W$_at_showdown_when_I_call_on_river %. If that number is less than about 50%, I suspect I'm calling too much. If it's more than about 30%, I suspect I'm calling too little. (Yes, I should often suspect both!)

Right now, I'm at 63%, so I almost KNOW I'm calling too little -- just hard to say exactly where [img]/images/graemlins/wink.gif[/img] Illustrating: Any systematic approach like this requires so much good individual-cases judgment to apply at all -- much less to fine tune.
Reply With Quote
  #3  
Old 10-10-2004, 08:56 PM
srt19170 srt19170 is offline
Senior Member
 
Join Date: Sep 2004
Posts: 272
Default Re: the general idea

If I wasn't clear, I was looking at the "Won $ at SD" percentage (thoughtfully provided by Pokertracker on the "Misc. Stats" tab). So in my case, I've only carried high card to showdown 3% of the times I've ended with high card; of those times, I've won 11%.

However, I think your second point is very interesting. I hadn't thought about it in exactly these terms, but I think the value of what I'm suggesting is that it provides an empirical way to calculate the EV of your judgement.

Ed Miller posted something (in micro-limits, I think) which said "don't fold a big pot for one bet on the river." (Or maybe I read it in SSH.) At any rate, part of his point is to play the pot odds *instead* of your intuition. Your intuition is that you're beat, so you want to fold your hand to save a bet. But with big pot odds, your intuition only has to be wrong occasionally for you to profit.

What I'm suggesting is that if you track your performance you can empirically calculate what "big pot odds" means -- the point where the pot odds break even for you. One of the nice things about this is that it is a personalized number for you. You might play many more high card hands to showdown than I do, in which case your pot odds number is likely to be much higher.

Obviously, if you have other information that you think is reliable you have to use it. But -- speaking for myself -- in many cases where I'm holding high card or a pair on the river I really haven't a clue where I stand. This method gives you an analytic tool to narrow in on the +EV strategy.

-- Scott
Reply With Quote
  #4  
Old 10-10-2004, 09:40 PM
Position Position is offline
Junior Member
 
Join Date: Nov 2003
Location: San Diego, CA
Posts: 20
Default specific PT stats

hey,

Yes, it was clear to me which number you meant! For my last few days of NL, that number is 58% (at the bottom of the column, for all high cards, pairs, everything). BUT my W$_when_called_river is only 50%. (My W$_when_I_bet_river_and_get_called is 68%, and also many times the river gets checked through in my games. This is how they net average to the Misc Stats' 58%.)

I claim that the when_I_called_on_river qualifier is absolutely necessary -- because that's the only situation you're trying to quantify in your post! Things get really muddled if you look at a statistic that includes your own value bets & bluffs on the river...

You can find the W$_when_I_call_river stat your General Info's More Detail pop-up window, but it doesn't break it down my hand type. You could use various filters to break it down though -- some of which are surely more directly relevant than hand type in your original sense!

(I'm using too many exclams. Must be 68%+ caffeinated!)
Reply With Quote
  #5  
Old 10-10-2004, 09:50 PM
Position Position is offline
Junior Member
 
Join Date: Nov 2003
Location: San Diego, CA
Posts: 20
Default the EV of judgment


I really like that phrasing: "the EV of your judgment." Looking for new analytic methods like the one you introduced is surely great for that.

We MOST care about improving our judgment of EV in specific instances -- and I totally admit that improving our judgment of the EV of our judgment (!) is a valid & sometimes useful way to improve that.

It may well be that this method you've introduced is fairly useful for your limit games; I do know that it can't possibly be even one of the 1,000 most useful such analytic methods for my games. (And, no, I don't think I've actually reached 1,000 such methods yet!)

The ultimate best long-term measure of the EV of your judgment, of course, is your BB/hand in each specific seat position or filterable situation. When you don't have a ton of hands yet at your current style&level, though, it's certainly useful to look at a lot of ultimately correlated more specific indices.

On intuition, your Ed Miller quote is terrific. I still find I need to keep it (or the equivalent) in mind when I play limit.
Reply With Quote
  #6  
Old 10-10-2004, 10:34 PM
srt19170 srt19170 is offline
Senior Member
 
Join Date: Sep 2004
Posts: 272
Default Re: the EV of judgment

You're right that this really applies only to limit games. I tried to think about what the equivalent technique would be in NL games and my head started hurting so I quit! :-)

-- Scott
Reply With Quote
  #7  
Old 10-10-2004, 11:32 PM
LetsRock LetsRock is offline
Senior Member
 
Join Date: Nov 2003
Location: California
Posts: 1,495
Default Re: Using Past Performance to Make Showdown Decision

Aside from the fact that you don't want to break even, nothing. Even is NOT an edge.
Reply With Quote
  #8  
Old 10-11-2004, 12:33 AM
Saint_D Saint_D is offline
Member
 
Join Date: Aug 2004
Posts: 96
Default Re: Using Past Performance to Make Showdown Decision

The other slight problem with this approach is the randomness in poker. You would have to play at least 10K hands or so each time you made an ajdustment to see if you were on track. With smaller sample sizes, the changes in these %'s aren't all that meaningful.

If your numbers are WAY off where you expect them to be, then you can make adjustments quicker. Like when I first started using PT and realized my VP$IP was almost 50% when it _felt_ much lower.

Just changing my pre-flop standards made a big difference in my game. I am still a tiny bit high at 22%, but I need a few thousand more hands before I can tell if I am still too lose (probably) or if it's just a run of cards where I am getting slightly above average hole cards.
Reply With Quote
  #9  
Old 10-11-2004, 11:11 AM
srt19170 srt19170 is offline
Senior Member
 
Join Date: Sep 2004
Posts: 272
Default Re: Using Past Performance to Make Showdown Decision

[ QUOTE ]

The other slight problem with this approach is the randomness in poker. You would have to play at least 10K hands or so each time you made an ajdustment to see if you were on track. With smaller sample sizes, the changes in these %'s aren't all that meaningful.


[/ QUOTE ]

Good point. I'm assuming you start with some substantial number of hands already played. At that point, your numbers are going to have some "inertia" -- that won't be substantially changed by a few hands. You're not going to go from playing 10% of high cards at showdown to 50% overnight.

On the other hand, now that I think about it, you likely play so few hands in that situation that maybe you would see radically swings. Hmm, I'll have to think about it.

I've started doing this just as an experiment, and so far it has affected exactly 1 decision :-).

-- Scott
Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -4. The time now is 11:41 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.