Friday, February 3rd: Table Selection
I think that I've started writing a post on this topic at least 3 or 4 times, but stopped for various reasons: sometimes I've thought that it was too mundane a topic, other times I thought I might just be re-hashing information that was common knowledge to most, and still other times I wasn't quite sure I was ready to delineate, in detail, one of the tools that has helped me achieve pretty phenomenal winrates at the medium-to-high stakes limit games. But I think that it's about time I wrote a little bit about table selection, what I consider to be the most underrated aspect of playing winning poker.
I have managed to run at 3.5BB / 100 over my past 150,000 or so hands at 20/40 and 30/60, a pretty lofty winrate (and one that prompted one 2+2'er to christen me "maybe the best limit poker player around"). While I humbly accept the nomination, that 3.5BB figure has a huge asterisk next to it, one that I'm not quite sure most people fully appreciate: I'm a 3.5BB player when I sit in very juicy games. The importance of that last bit really shouldn't be underestimated -- how good a player would I be if I just sat down at the first table I opened? I have no idea -- 2BB? 1BB? Would I even be a winner at all?? I don't have the answer to any of these questions, but the cool thing about it is that I really don't care. The biggest strategical advantage of online poker over live play is not the speed the hands are dealt, nor the slightly lower rake or lack of tipping -- those are each dwarfed by the incredible advantage of being able to pick and choose from dozens of different subsets of potential opponents.
Poker is a game of relative strengths: it really doesn't matter how technically skilled you are -- the primary determinant of your eventual success is the discrepancy between your abilities, and those of the opponents you sit with. Accordingly, there are two primary avenues to improve your winrate: 1) improve your own play, or 2) do your utmost to ensure a poor ensemble of opponents. There are hundreds of paper and electronic resources dedicated to the former, and while some mention the importance of table selection, I have yet to see any author give the topic the due it deserves.
So let's get down to the nuts and bolts of just how you find yourself a table full of donators. First things first. If you don't have Pokertracker and PokerAce, you're doing yourself a disservice. Spend $75 and get them. It's embarassing just how valuable they are; you won't find a better bargain as long as you live. I would honestly estimate that the seventy-five bucks I dropped on those two applications has earned me upwards of $50,000. Pokertracker (PT) is software that lets you collect information on all the hands you play, and PokerAce (PAHUD for short) is add-on software that will display your opponents' playing styles and statistics right on top of the table as you play. Some poker sites even let you collect data on hands you just observe -- e.g. you don't even have to be playing in a hand to collect data on the players. Party Poker, however, appears to be putting an end to the ability to collect stats on observed hands...it's still possible for now, but most people believe that they will soon release a new version of their software which does not enable datamining observed tables.
So first step is building up a robust PT database full of statistics on your opponents. Even when I'm busy with schoolwork and other responsibilities that don't let me play poker, I make sure that my computer is actively datamining hands from many different sites so that when I finally do get around to playing, my database is as complete as I can make it. So what to do once you've got your PT database chalk full of statistics? Maybe I can best illustrate by giving an example of my typical table selection process. I login to the poker software, and open up as many tables as I can -- let's say I'm feeling like playing some full-ring limit games: I'll open up as many 20/40 and 30/60 tables as I can find. I've got PAHUD configured to show me the table's average VPIP and PFR statistics...(as many exotic stats as PT can tell you about your opponents, I really think that VPIP and PFR account for 98% of what the typical player needs to know to get a good read on the players at his table.) If the table's average VPIP is under 26, I'll close the table without a second thought. If it's over 30, I'll nearly always take a seat or get on the waiting list; that's a pretty damn good table. If it's between 26 and 29, I'll take a closer look at the individual players at the table. Think about the following things: just how many fish are seated? Is the higher-than-normal table VPIP a result of a bunch of rocks and one gigantic fish with a 90VPIP, or rather is it 5 or 6 slightly-loose players with VPIPs of 28 to 35? More importantly, where are the open seats at the table? I typically require at least 3 fish at any table in order to consider it play-worthy, but I'll accept just two (or even one) if I can sit to the immediate left of a super-LAG. How big are the chip stacks of the fish? If they're down pretty low, that may be an indication that they won't be around anymore by the time your name comes up off the waiting list. Are the fish at the table passive or extremely aggressive? A 45/7 fish is a completely different animal than a 50/33. If the 50/33 is one of only 1 or 2 fish at the table, I'll typically only sit if I can get a seat to his left, whereas it's easier to exploit a the weaknesses of a 45/7 regardless of position. Those VPIP numbers will obviously differ at a short-handed table...at a 5 or 6-handed table, I typically look for average table VPIPs of above 38, although as before I'll accept a 35-37 if I can get good position on a LAG.
I'm sure some of the more experienced limit players reading this will find it all old hat, and a waste of their time...in fact, that's one of the concerns that previously stopped me from writing about table selection: I figured everyone knew all about it already. But you know what still stuns me: in searching for good tables to play, I'll frequently open up a 30/60 table, for example, only to see that it has an average VPIP of 18, with rocks all around. But the fascinating part is that I'll see 3 or 4 players sitting down who I happen to think are quite formidable opponents...and really not anyone I'd care to see at my table. I know that they must have played a ton of hands and in fact probably even use PT and PAHUD themselves...why then do they satisfy themselves with sitting at a table of 9 other TAGs?? It just doesn't make any sense?!! Even if they're not using PT/PAHUD, surely they've played enough hands to recognize the screenames of the tough opponents they're facing...so I close down the table and keep clicking around until I find a few juicy ones, which are never all that elusive. I can't help but wonder whether those TAGs are too lazy to look for better tables, or simply don't appreciate the importance of table selection...I find it very hard to believe the latter -- which made me think that maybe discussing this little tidbit isn't as worthless as it may at first have seemed. So take this little entry for what it's worth. Finding yourself a batch of below-average opponents, while it won't substitute for a solid techincal foundation, will do wonders for your winrate. Good luck.
I have managed to run at 3.5BB / 100 over my past 150,000 or so hands at 20/40 and 30/60, a pretty lofty winrate (and one that prompted one 2+2'er to christen me "maybe the best limit poker player around"). While I humbly accept the nomination, that 3.5BB figure has a huge asterisk next to it, one that I'm not quite sure most people fully appreciate: I'm a 3.5BB player when I sit in very juicy games. The importance of that last bit really shouldn't be underestimated -- how good a player would I be if I just sat down at the first table I opened? I have no idea -- 2BB? 1BB? Would I even be a winner at all?? I don't have the answer to any of these questions, but the cool thing about it is that I really don't care. The biggest strategical advantage of online poker over live play is not the speed the hands are dealt, nor the slightly lower rake or lack of tipping -- those are each dwarfed by the incredible advantage of being able to pick and choose from dozens of different subsets of potential opponents.
Poker is a game of relative strengths: it really doesn't matter how technically skilled you are -- the primary determinant of your eventual success is the discrepancy between your abilities, and those of the opponents you sit with. Accordingly, there are two primary avenues to improve your winrate: 1) improve your own play, or 2) do your utmost to ensure a poor ensemble of opponents. There are hundreds of paper and electronic resources dedicated to the former, and while some mention the importance of table selection, I have yet to see any author give the topic the due it deserves.
So let's get down to the nuts and bolts of just how you find yourself a table full of donators. First things first. If you don't have Pokertracker and PokerAce, you're doing yourself a disservice. Spend $75 and get them. It's embarassing just how valuable they are; you won't find a better bargain as long as you live. I would honestly estimate that the seventy-five bucks I dropped on those two applications has earned me upwards of $50,000. Pokertracker (PT) is software that lets you collect information on all the hands you play, and PokerAce (PAHUD for short) is add-on software that will display your opponents' playing styles and statistics right on top of the table as you play. Some poker sites even let you collect data on hands you just observe -- e.g. you don't even have to be playing in a hand to collect data on the players. Party Poker, however, appears to be putting an end to the ability to collect stats on observed hands...it's still possible for now, but most people believe that they will soon release a new version of their software which does not enable datamining observed tables.
So first step is building up a robust PT database full of statistics on your opponents. Even when I'm busy with schoolwork and other responsibilities that don't let me play poker, I make sure that my computer is actively datamining hands from many different sites so that when I finally do get around to playing, my database is as complete as I can make it. So what to do once you've got your PT database chalk full of statistics? Maybe I can best illustrate by giving an example of my typical table selection process. I login to the poker software, and open up as many tables as I can -- let's say I'm feeling like playing some full-ring limit games: I'll open up as many 20/40 and 30/60 tables as I can find. I've got PAHUD configured to show me the table's average VPIP and PFR statistics...(as many exotic stats as PT can tell you about your opponents, I really think that VPIP and PFR account for 98% of what the typical player needs to know to get a good read on the players at his table.) If the table's average VPIP is under 26, I'll close the table without a second thought. If it's over 30, I'll nearly always take a seat or get on the waiting list; that's a pretty damn good table. If it's between 26 and 29, I'll take a closer look at the individual players at the table. Think about the following things: just how many fish are seated? Is the higher-than-normal table VPIP a result of a bunch of rocks and one gigantic fish with a 90VPIP, or rather is it 5 or 6 slightly-loose players with VPIPs of 28 to 35? More importantly, where are the open seats at the table? I typically require at least 3 fish at any table in order to consider it play-worthy, but I'll accept just two (or even one) if I can sit to the immediate left of a super-LAG. How big are the chip stacks of the fish? If they're down pretty low, that may be an indication that they won't be around anymore by the time your name comes up off the waiting list. Are the fish at the table passive or extremely aggressive? A 45/7 fish is a completely different animal than a 50/33. If the 50/33 is one of only 1 or 2 fish at the table, I'll typically only sit if I can get a seat to his left, whereas it's easier to exploit a the weaknesses of a 45/7 regardless of position. Those VPIP numbers will obviously differ at a short-handed table...at a 5 or 6-handed table, I typically look for average table VPIPs of above 38, although as before I'll accept a 35-37 if I can get good position on a LAG.
I'm sure some of the more experienced limit players reading this will find it all old hat, and a waste of their time...in fact, that's one of the concerns that previously stopped me from writing about table selection: I figured everyone knew all about it already. But you know what still stuns me: in searching for good tables to play, I'll frequently open up a 30/60 table, for example, only to see that it has an average VPIP of 18, with rocks all around. But the fascinating part is that I'll see 3 or 4 players sitting down who I happen to think are quite formidable opponents...and really not anyone I'd care to see at my table. I know that they must have played a ton of hands and in fact probably even use PT and PAHUD themselves...why then do they satisfy themselves with sitting at a table of 9 other TAGs?? It just doesn't make any sense?!! Even if they're not using PT/PAHUD, surely they've played enough hands to recognize the screenames of the tough opponents they're facing...so I close down the table and keep clicking around until I find a few juicy ones, which are never all that elusive. I can't help but wonder whether those TAGs are too lazy to look for better tables, or simply don't appreciate the importance of table selection...I find it very hard to believe the latter -- which made me think that maybe discussing this little tidbit isn't as worthless as it may at first have seemed. So take this little entry for what it's worth. Finding yourself a batch of below-average opponents, while it won't substitute for a solid techincal foundation, will do wonders for your winrate. Good luck.
posted by Blog Administrator at 1:40 AM
20 Comments:
I'm a full-ring-game 1/2 player. I ten-table and make a good per hour return (obviously a lot less than a mid-to-high stakes player like you). I'm also looking to move up to 2/4 or 3/6 soon.
How important should PokerTraker, PAHUD, and, in general, table selection, be to my game?
Right now I don't use it, assuming that the sheer number of weak players at the low limits means i'll always be at a soft table.
But, reading your last posts (and a few others where you've mentioned such issues), I've come to suspect I'm wrong.
I'd appreciate your thoughts.
Why are you actively trying to make your opponents better? Delete this post now.
dude, erik, did you even READ his last post? How important do you THINK he'll say table selection is?
pahud and pt are infinately worthwhile.
Ha - I'm glad someone else said it so I didn't come across as a dick.
It's more important at the higher limits, I'd say, but still useful for you. (And no, I'm not deleting this post.)
To answer your point about why a table might be full of good players. Not everyone plays just for the money all the time.
Some people appreciate the challenge of playing against those they consider good players and seeing how they measure up. It's a good way to improve your game before your next fishing session.
I lost a 30 BB pot yesterday due to disconnection. ANy suggestions for a good multiple DSL backup system?.
Best
B
No, sorry B -- I saw your 2p2 thread though...hopefully that generates some ideas.
I use the same system, except I play short, so everything 40+ I jump on the waiting list. Every thing below 35 I dump and load a new table. I continue this process during the session since tables break up fairly often or the VP$IP goes south.
I also keep an eye out for tables with all players in the mid to high 30s, they can be real gold mines. Mostly, it's fairly effortless to find tables with two donators and two tight (often too tight) players.
Since you're in the sharing info mood, how do you handle a massive data base? My main db is getting close to 2 million hands and it's starting to get laggy.
If you're using seperate databases, how do you get PokerPatterns to make one graph?
I play 3/6 SH and have a similar table selection process. But the real question is, what on Earth will we do when Party stops storing the HHFs? There goes easy table selection. That could have a pretty big impact on winrate I think.
Great post. I've read a few of the poker books that are available, but I have never been exposed to this method of table selection. Thanks.
What sites allow for observed hands to be datamined, besides Party?
I believe there are tools to datamine observed hands at PokerRoom, UB, Full Tilt, and Absolute (according to a new post in the 2p2 software forum). Go to PT's forums, and I think there's a sticky devoted to each mine-able site.
2 Million hands in ur PT DB?? Geez, I create new ones after about 600,000, even though i know that postgres format allows for DBs of technically unlimited size, I find that performance slows down at 500K+, so I just create new DBs.
And as far as PokerPatterns - they key is having separate DBs for your own play, and for observed hands...that way you can create new DBs for observed hands when they get to big, and just have PokerPatterns point to your small to modest size personal-hands DB.
so if you're playing on Party now, will you continue to play there if datamining isn't possible? or will you move to a site you can mine.. which one? because i have a feeling i will move if mining is no longer possible, but i'm only familiar with party, don't know what other sites are "juicy" and which aren't so much..
Could you define vpip and fpr, so that I can better understand why these numbers describe a bad player?
Thanks for all your discussion and analysis.
Gordo
P.S. I suspect he won't be sitting across from any poker players that he makes better.
Gordo, just google VP$IP and PFR.
Even if Party stops allowing mining of observed hands, you'll still be able to get HHs for hands you played...so it will just make gathering opponent data much much much slower. For example, I can't mine Stars anymore, but I still play there when I find good games based on the stats that I DO have. Besides, creative programmers will probably figure out a way to datamine Party via some type of screenscraping function (even if that type of application is technically prohibited by the poker site.)
What do you think of Poker Office? Do you consider it as good as PT?
Great post !
I have a few questions about this subject.
Is it true that PT is more suitable for limit than NL games?
I only multitabling NL (full $100 tables) and like to know if PT combined with PokerAce is any help to my game.
What figures (VPIP/PFR/table average VPIP) are important for NL games.
After logging how many hands of a persons play do you have a trusty read on someone?
A nice extra for using PokerAce is that you can see mucked holecards.
Is there also an add-on for PokerAce to automaticly classify a player with icons like fish, rock, etc.?
thanks a lot
P
PT and PAHUD are must-haves regardless of whether you play NL or LHE.
Just FYI the answers to all your questions are scattered throughout my blog, so do some more reading but as way of quick answers for now, yes PT can automatically assign a picture (and you can have PAHUD display it) -- read the PT how-to section accessible from the help menu within the program.
I consider stats on opponents to give a reliable read after 150 - 200 hands...well, reliable enough to base my own actions on them, that is.
I think PFR and VPIP give you 95% of the information you need to know...for NL games I also really like looking at the continuation bet on flop/turn stats that PAHUD provides...lets you know how likely it is you'll be facing follow-up bets if you call a PFR.
Good luck.
9:16 AM
As long as you do not spend more then 10 minutes looking for a table it's OK. Remember, every second you are NOT playing is losing you money.
Get on the first table you can find and play. THEN look more another table.
Plus - tables turn over every 1/2 hour so a good table will always go bad nad a bad table will always go good.
So WHEN you find a good table - don't sit there and watch it if there are 5 guys ahead of you. It's silly!
Post a Comment
<< Home