Opening up the books

Kind of a lot has happened since my last blog post. Most notably was Poker's "Black Friday," on April 15th of this year. Here's a pretty good article about that, but the short of it is that the DoJ shut down PokerStars and Full Tilt Poker. Now, PokerStars was able to return money to their U.S. players pretty quickly, but Full Tilt Poker has been unable to do so thus far. So, currently I can't play online poker, and have about $3,000 locked up in there that I can't even access, and am starting to get a little worried that I might never see it again. But we'll see what happens.

So in this downtime I've decided to look at my recent stats from before the shutdown, to do some simple analyses and maybe learn something about my game and continue to try and grow in this way while I can't actually play. So... here we go.

First, my data that I'm looking at right now are from a recent 4 month window, and consist of roughly 50,000 hands played across three levels of stakes: $50, $100 and $200 no-limit hold'em (with the $ amount being the buy-in for the game). To be precise, it's 51,709 hands and the breakdown is:

13,056 hands at $50NL
26,538 hands at $100NL
12,115 hands at $200NL

So what is 51,709 hands of poker exactly? Well, in a brick and mortar casino, you can expect to get dealt about 30 hands per hour. So if I were playing only in a brick and mortar casino, how long would it take me to play 51,709 hands? If you do the math, it would take me about 43 weeks playing at 40 hours per week to attain this.

In contrast, by playing online I was able to attain this in 4 months playing an hour or two per night and only whenever I had the free time, due to the fact that hands get dealt faster online AND you can play multiple tables at the same time.

Ok, now how much did I make? Well my raw profit at these stakes was $2,986.75 over this time period. One meaningful statistic though is your winrate, in terms of BB/100. That is, big blinds per 100 hands. So for example, if you are playing $100NL where the big blind is $1, and your winrate is 2 BB/100, then that means that on average, you are winning $2 every 100 hands. Also, I will note that when I play, I average about 600 hands per hour. From this, I can also calculate my hourly rate. So here they are:

$50NL: 9.80 BB/100 -> $29.40/hr
$100NL: 5.88 BB/100 -> $35.28/hr
$200NL: 3.24 BB/100 -> $38.88/hr

Ok so my BB/100 decreases as the stakes increase, which is to be expected since higher stakes means harder opponents, typically. On the other hand, I do make the best hourly rate at the highest stake of $200NL.

It's only marginally better though; at each increase in stakes I'm playing twice as big (i.e. risking twice as much) but only doing slightly better. But, two things: 1) It's always good to try and move up and be able to win against better opponents and keep improving your game, so even if my winrate suffers in the short term, I should strive to attain good winrates at higher stakes in my continual quest for improvement as a player, and 2) These estimates are actually subject to a lot of variability, even after 10,000+ hands that I have at each level. The variance of poker is so much greater than one might think. I haven't done the math yet but conventional wisdom says that you need to have about 100,000 hands before you can have any reasonable precision on your winrate.

So, even though $29.40-$38.88/hr playing poker may sound pretty sweet, I wouldn't bet the farm on me yet (it's pretty clear that I do *beat* these games, but just by *how much* isn't very clear yet). Also, and more importantly, we already know that I hated playing poker full time when I did it the summer before starting my PhD.

There's a lot more I can look at, but I'll end this here for now and maybe take a look at some other things later.