You really don't need anything that complicated to beat low-stakes no-limit poker short stacking. I used to know a few people who made decent hourly rates just mechanically following a very small spreadsheet. I did it myself for awhile and it seemed like an easy win, though not as profitable as simply being good at the limit games I was playing and nowhere near as challenging/fun. My personal sample size was small, but I still suspect a small set of if/then statements would have been a winner.
The real challenge is finding enough tables to play. The bad thing about short stack is that sometimes someone calls you and you win, and now you've got at least a medium stack. Most sites don't let you come back to a no-limit table with less money than you left with until some period of time (for instance 2 hours) has passed. Party Poker did not have this feature (maybe still doesn't) so it would have been easy back then, but for an American now I don't know of any site of substantial volume where you wouldn't quickly burn through all available tables and find yourself waiting.
Even for a bot, there might not be much profit in it.
I purchased 7M+ NL200 hand histories from a guy who mined them all day every day. I filtered that down to hands where someone was playing with 15 - 25BB stacks and then for every person that played more than 5,000 hands in that range, calculated their profits and winrates.
It's important to remember when looking at these numbers that the vast majority of people who try shortstacking never make it to 5,000 hands so this is heavily skewed toward the survivors. It would be like analyzing the revenue of startups that are over five years only and then using that average for the average revenue of all startups. There's only 80 people in the full ring group and 19 in the 6-max group. Just looking at these folks, only about 1 in 5 win more than 2bb/100 and in my experience that's not easy to do. The people that are profitable shortstackers would likely win a lot more playing with a normal stack.
I don't think there are many long term profitable shortstackers following rules. Again, I think the majority of the profits come from a few strategic plays that would be impossible for a spreadsheet, or a bot, to emulate.
Idk, those numbers seem pretty reasonable to me and don't indicate any real level of difficulty. 58% of people in it are even or worse, which means if there is any sample bias at all, it's very small. The fact that losers tend to drop out over time is probably mitigated at that level by the fact that winners will quickly rise from low stakes to at least middle ones.
2-4 bb/100 is great even for a human since you can do huge volume that way. The spreadsheet method can easily be applied to 8 tables at once, maybe more. 2 bb/100 (thats $4 as calculate by PT right?) could be $30/hr. 4bb is obv $60, which is more in line with what people I knew were making.
Also when short stacking, opponents at higher level games don't necessarily respond any more appropriately. In fact a lot of times they play worse against it because they have more idea what to do vs. normal opponents. Low level players like to call big bets too much, which makes them accidentally better vs. a short stacker. My friends claimed they were getting higher bb/100 at the $3/$6 and $5/$10 games than the $1/$2 because of that.
I just realized that I have the hand histories for all cash No Limit hands played on PokerStars from Dec 1, 2007 to Mar 12, 2008. I've got all the tools built to do the analysis, so I can pick apart the results from the lower and higher stakes as well. Any recommendations on how to do it?
Not PTBB/100, just BB/100. PokerTracker couldn't handle all the data.
I suppose you could possibly control for play style, thinks like what Poker Tracker tracks. VPIP, etc., and of course buy-ins if you do some technomagic to figure them out. (I'm assuming someone who short stacks all the time plays differently than a normal player who just lost half their stack.) Not sure how good your findings would be though.
The real challenge is finding enough tables to play. The bad thing about short stack is that sometimes someone calls you and you win, and now you've got at least a medium stack. Most sites don't let you come back to a no-limit table with less money than you left with until some period of time (for instance 2 hours) has passed. Party Poker did not have this feature (maybe still doesn't) so it would have been easy back then, but for an American now I don't know of any site of substantial volume where you wouldn't quickly burn through all available tables and find yourself waiting.
Even for a bot, there might not be much profit in it.