A short post on something I'm doing these days

[ QUOTE ]
Understand, I play 2 tables and think that playing much more than that is probably a bad choice for the long term development of your game, so I'm not talking to anybody trying to play 12 tables.

[/ QUOTE ]

You have to realize that even a very good player only has a small edge over a bad player in NL cash games. The rewards have to be squeezed out over hundreds of thousands of hands to have any meaning. I am trying to win money while there is money still to be won, so I need to get in as many hands as possible.

I learn the same today as when I was playing 1-2 tables or playing live. It is about how much work you put into the game and the thought process of the game that will help you improve your game. If you want to put extra work into reading players, I suggest adding live play which you may or may not already do.

[ QUOTE ]

Maybe you guys just don't understand or realize how a little tiny bit of information lets me imagine exactly what type of player my opponent is, compared to two numbers that can skew your reads considerably.

I'm talking about NL.

[/ QUOTE ]

And what you don't understand is how un-important that information is in the short-run. You need to gain your value over these opponents by playing hands, many of them.

[krishanleong]

[ QUOTE ]

This is less about specifically being able to beat 3/6 than it is about just building what I would consider "good" habits as I progress. Even though I don't play often, I approach everything I do at these low limits as learning for some pipe dream when I play higher and not just an exercise in making money.

[/ QUOTE ]

Excellent. Wish I had a more of this in me somtimes.

Krishan

[krishanleong]

[ QUOTE ]
No hard feelings really, but I disagree pretty much 100%. I'd be very very very skeptical that you could "define how a guy plays" with 1k hands worth of observation better than by using 1k hands worth of data.

I'd also think it might be a bit more time effective to accumulate the data without having to observe. Though most of the sites I play I can only get information on hands I'm dealt in.

My point is that probabilistic thinking for me rules. you can't do sh!t for bayesian analysis with observation unless you are rainman squared.

[/ QUOTE ]

Not suprisingly I agree totally with Stox. I rely on stats a huge amount as a serious multitabler. I think I understand when sample sizes become significant and identify player types accurately based on combinations of stats.

I believe in stats so much that I'm compiling a shorthand high limit (15/30 - 30/60) where I only have 4-6 handed games filtered. It's taking forever... [img]/images/graemlins/smile.gif[/img]

Krishan

[B Dids]

I don't say this to be a dick, but don't both of you sell and/or aspire to sell large mined databases of player stats?

I understand what you're saying, and fully believe you'd say that if even if you didn't, but it's probably something that merits mentioning.

If I played to pay the rent, I would totally datamine. As it is, given the approach to poker that I've taken, it simply isn't that critical to me, and I enjoy working on gathering this information in other ways. I understand what stox is getting at, and have made arguments very similar to his in other contexts (baseball). Clearly my sample size sucks, and clearly obvseration is a flaw method of gathering information in some ways. However, I still think you get a bit extra from that bit of focus that you might grant to each new table.

[jason_t]

[ QUOTE ]
This seems awfully weird coming from somebody who makes his money in games that tend to run short, where how a guy plays seems even harder to define in specific numbers.

[/ QUOTE ]

It depends on what type of player (even human being) you are. Analytical / mathematical / statisical types will favor the data and the incredible amount of information contained therein. The opposite type of person will favor more specific reads acquired from a few observations. Analytical types will tell you they get a better picture out of 1k hands than those that aren't do. This is because of the encapsulatory nature of their statistics.

[stoxtrader]

krishan used to seel stats for party, he does not any longer. I sell my DB for limits 50/100+, the purpose is as a teaching tool and nobody that has bought it has asked any question to lead me to believe they are using it as anything otherwise. It is certainly conceivable my DB has been bought by opponents of mine for data purposes, but thats not the thrust of the offering - it's as a teaching tool and a number of players playing 5/10- 30/60 have bought it.

I am li ke you'ared loca sion!!

[krishanleong]

[ QUOTE ]
I don't say this to be a dick, but don't both of you sell and/or aspire to sell large mined databases of player stats?

[/ QUOTE ]

I use to. But I don't believe stats are valuable because I use to sell them. I decided to sell them because I think stats are valuable.

I currently mine about 80ish tables 24 hours a day just for myself.

If you start every session with no stats on half the players at your table, you are going to be at a severe disadvantage to someone with data on everyone(me) given the same skill level.

And of course, having reads that support the stats is a good thing. No reason you can't have both.

Krishan

[Grisgra]

[ QUOTE ]
maybe I'm being shortsighted, but I'm really a seller of observed information vs compiled data. I guess the biggest reason is sample size. It's impossible to get more than 2-3 "observations" on any substantial amount of players, so there is a sample/size and scaleability effect their.

secondly, I can't for the life of me see where something you observe could be more valuable than a stat telling his play over a much larger amount of hands....definitely not pre-flop, and surprisingly, even for me, post-flop.

pokertracker really is an amazing tool.

How about this, I want to be proved wrong, and imagine I will be with this challenge, but why dont you come up with a few observations that can prove more valuable to me than I could get from a corresponding PT stat of my choice describing the situation. The objective here is to takeaway information on a player from a previous hand that will help you play better vs him on subsequent hands. So observation better vs, say 1k hands of PT data. or whatever number of hands you choose, because I think any sample size arguement that refutes PT data, can be used against observational information....

but just try to give me examples and I will try to refute, If i cannot I will graciously admit defeat and mend my ways....it may even be in this manner that we do discover there is something different between limits, but I can't see what...

[/ QUOTE ]

I sat down to a 20/40 3-handed table recently (I was the 3rd person), and went into bully mode, raising like a MF. On the 4th hand I button-raised with 76s, and some regular schmoe called in the BB. Flop was 754 and to make a long story short, he ended up 3-betting the turn with K5o.

I have his PT stats. Nothing in there suggests that he's going to 3-bet the turn with 2nd pair . . . fact is, though, given the way I had been playing -- lots of aggression -- he felt it was time to make a stand. I figured my hand was probably good even after the 3-bet, while if the situation had been different I would have thought my hand was utter crap.

Because I had been paying attention previous hands, I could read his 3-bet for what I thought it was (frustration/fighting back against a perceived LAG), and I got a read on him that could prove useful down the line (how he reacts to bullying).

Give me the PT stat for how a player reacts to bullying/aggression, please.