Tuesday, October 24, 2006

Signature Stupidity and Top-Level Impressions

Alright, confessions of a secret poker sin --- I'm not as ardent of a note-taker as I probably should be. One of the reasons for this is that short-term interactions are rarely reliable as long-term trends, so seeing a player dive in on seven out of ten hands --- or 45 of 100 --- may mean nothing more than a little run of hot cards. The truth is, most people who use PokerTracker also misuse it, because they make decisions based on numbers that are not only subject to short-term variations, they're also categorized according to arbitary, percentage-oriented boundaries. And, yes, I know that you can set the boundaries yourself, but the fact is that no matter where a given boundary lies, a player with behaviors just to one side of that boundary is almost the same as a player just to the other side of the same boundary.

So I rely less on notes and stats and more on feel than I probably should, though I do note what I call as signature behavior. For instance, I was playing 2/4 razz at Full Tilt yesterday, and saw a player make a river bet of $4, be reraised by the other remaining player, who had up cards of 2-7-8-K, or something like that. The first player called his re-raise, showed his A-2-3-4-5 wheel, and raked the pot. And he received a note in my record stating that he was an idiot --- the other player had at best a seven-low, so there was no danger of a chop; the re-reraise was mandatory. But there are many bad razz players out there right now, the majority of whom seem to be no-limit-only kids who now want to be H.O.R.S.E. champions, but lack a mathematical clue. Still, I give them credit for at least trying to expand their borders, even as I'm happy for the bankroll padding they provide.

So, if you make steal attempts from the cutoff five straight times, I notice it and note it. Overly fond of the check-raise? I'll catch on to that as well. I trust signature behaviors far more than I trust mathematical-derived boundaries.

There's always a second reason to my posts, and this one comes courtesy of a post today from Juice over at the The Pain of Texas Hold Em' Poker (sic) blog. In it, Juice mentions a new site that puts some PokerTracker-type findings on the web for all to see, along with limited recommendations as to the type of basic improvements that the searched-upon player could make to better his game. So, I'm always curious as to what these sites say about a low-level player such as me. There are two names I've played under more often than any others, and both of them should have thousands and thousands of hands out there.

First, let's see what it has to say for the name 'ChayseTilton':



And now, for the name 'CawtBluffin':



A distinctly different player, despite the fact that in both cases, I'm me. One thing I discovered in subsequent searching is that one of the two profiles shown above is based on a very limited hand sample, despite the fact that I've played all those hands that should have shown up at each site. So I'm suspicious of the database used for these findings --- when one combines that with the fact that those symbols that seemed to be freely borrowed from PokerTracker, what this looks like is a massive data-mining project that's been converted to a query-able database and put onto this site. Sounds a bit shady to me, and certainly not in the spirit of poker.

I'm not much for data-miners, so I'm pretty sure I can resist the temptation to sign up for this service.

No comments: