Wednesday, February 08, 2006

Poker Stop Losses

After further reflection (and nudging from comments), I think my arbitrary number of 40 BBs for a daily stop-loss at 10/20 short is way too small. But if that number sucks, what's a good replacement?

If the last month of playing 10/20 short has taught me anything, it's that your bankroll is always quickly moving in one direction or another. Like, seriously quickly. I honestly don't think the 1000 BB bankroll number is conservative at all for that game, especially if you're playing aggressively and multi-tabling.

So that's a $20K bankroll. Okay. You'll see different estimates from different pros, but the numbers I can remember seeing tossed around are that you should never lose more than 5-10% of your bankroll in any one day (which I think comes from the mouth of Lederer), if you're practicing good bankroll management. So that's $1,000-$2,000/day, as far as a stop-loss, which I suppose you should adjust as your bankroll grows/dwindles.

On the conservative 5% end, that's 50 BBs. Hmm.

I imagine there's a very good chance that the above estimates weren't taking the nature of shorthanded games into account, and speaking more broadly. Even so, if you look at the actual numbers, you're running in dangerous waters when your stop-loss is 10% or more of your total bankroll, methinks.

I think the real truth, lurking beneath this, is that if I'm even pondering setting a dtop-loss, I might instead rethink the whole idea of playing 10/20 short. I tend to be stubborn at times, especially when things seem readily apparent and I've worked hard at them, but the simple truth may be that I'm much more suited for full ring play. I'm barely breaking even at the 10/20 short games and definitely not fully situated in my comfort zone.

Those games are very juicy, indeed, but not necessarily juicy for all types of players, especially those who tend more to the tight side, like myself, instead of towards the gleefully-flinging-chips side. I think it's a reasonably telling thing that I've gotten much more upset in regards to poker than I ever have before, in the last month, verging on complete disgust at times. Which isn't much fun, really, in any way.

So, err, yeah. Rambling. I'll probably stick with a 50 BB stop loss for now, even though I know that's likely too small for the variance of shorthanded games. I'm still looking at all this as a working experiment, as far as setting myself up to grind away at whatever limit/game I settle at in a few months, when I slip the surly bonds of Cubelandia.


d said...

Over how big a sample size are you "barely breaking even"? It seems like the numbers that people throw around for a good SH player is 3+BB/100. If your sample size is big enough, then do you think that your tilt losses after passing Caro's threshold of misery are on the order of magnitude of 3BB/100? If not, you need to do more analysis as to what is affecting your results. As you touched on, maybe your style of play needs to be tweaked more for SH play. Maybe your game selection isn't as ideal as you think.

cc said...

I think 40 BB is about right, with a hard rule that you never jump up after being down 20-30. 60 is probably too loose. Also, what do you do if you hit your 50? You may want to have a detailed diagnostic afterwards to identify top causes with written game plan for next session. Probably a bit anal, but my two bits.

ScurvyDog said...


I'm nowhere near a valid sample size for 10/20 short, just barely at 30,000 hands or so.

As far as 3BB/100 win rates, I don't doubt it's possible, but I'm not sure I'd ever expect to personally get there. I don't push thin edges hard enough to get there and, honestly, simply am not that good yet to achieve that.

That said, I should be in the 1-2BB/100 range, and I'm not close to that. Some of that is due to running horrendously, epicly poorly over a stretch of 10,000 hands, but not all of it.

Long story short, yeah, definitely more tweaking is needed if I'm going to continue to play the short game.

ScurvyDog said...


You definitely have a point, as I could stand to be much more anal about keeping and reviewing records, and acting accordingly. Far too often I lump bad play/bad beats in the same money-losing boat, but they obviously come from different places and should likely be treated differently.

While I remember individual hands where I was doomed to lose many chips (i.e. set over set, boat over boat), I don't take the time to make notes on sessions as a whole, as far as pinpointing the nature of losses, with x percentage due to my donkey play, and y percentage due to unavoidable situations, and z percentage due to passive play, etc.

d said...

If you set your stop loss limit proportionally to your BR size, then there is should be some mathematical way to come up with a number based on your estimated BB/100 numer and your standard deviation. SD should start to converge by 30k hands even though win rate would not. One of the probability guys at 2+2, like Homer, probably made a post sometime about calculating BR based on those 2 numbers. (notice they are independent of SH or full).

Personally, the few times that I have put meaningful table hours into SH, the number of simultaneous tables I was playing did affect how I felt about any swings I took. A 100BB swing when 4 tabling seems pretty common. BR requirements shouldn't be significantly impacted by the degree of multitabling (except for the impact that multitabling has in decreasing your BB/100), but I think that a stop loss limit (if you choose to have one) should definitely be adjusted for larger session swings in multitabling.

Mark said...

I'm skeptical of stop-loss plans unless you feel like you're tilting away a lot of money. It seems more likely that the idea of a stop loss is more to limit your own misery rather than save money.
Shorthanded games are tough, as you probably know better than me. I wonder if they're even worth the trouble.
I don't know how to quantify the increase in variance in shorthanded play, but it certainly seems to me that it's a good deal higher than full ring than I expected.
I don't know why.
Good luck.

BSN said...

Supposedly, Howard told Annie 30 BBs for a stop loss. Of course, this was live play at one table. If he/they subscribe to the 300BB bankroll rule, that would be 10%.

kurokitty said...

Yeah, Annie Duke says you're unlikely to recoup more than 30BB when you're stuck in a session.

Keating said...

I play for a living (in part time grad school as well) and don't like the Party 10/20 6-max game. Are the edges sometimes huge? Definitely. But the variance is a supreme bitch, and in my opinion the best players in that game are easily as good (if not better) than the regulars in the 15/30 and 20/40 full games, so you have to table-select a lot. The games break up a lot more often which is a pain. And it's tough for me to play more than 2 tables at a time without losing reads - reads which are more important in 6-max than in full, esp. given the frequency of heads-up blind battles.

I'm confident that it can be beaten EV-wise for 3BB+/100 by an excellent player. But there aren't nearly as many excellent players as people think - and I am good enough at this game to know that I'm not excellent.

Just my thoughts, of course.

cc said...


One other thing. Luke Kim writes about Podcast with Chris Ferguson on FullTilt regarding his experiment of running up $1 to $20k. He says he busted out once at $1 and had to rebuy, that he never put more than 5% of his bankroll on the table at any one time. It is more broadly about bankroll management, patience, and discipline with your own rules than about stopgap, but if you haven't listened to it there are some great points.

ScurvyDog said...

Many thanks for all the comments.

Still not sure where my head is, as far as using a stop-loss, but I think the more pressing question to answer is just how dedicated I am to the high variance of shorthanded play to begin with. That seems to be what's really lurking behind the consideration to implement a daily stop-loss, more than anything actually related to the mechanics of a stop-loss, what percentage of your bankroll it should be, etc.

sm0kelm said...

I don't mean to sound stupid with my question, and forgive me if I do, but why not move down a limit?

Apart from playing that limit because you have the bankroll for it, if you are breaking even, why not move down to the 5/10 short, wait until you have definitely established a solid win rate there (2BB/100 or so), and then move up?

Isn't that what you did in the full ring games? Start low, and move up? Why not 6max?

SeanSkill said...

I have played about 150k Hands of 6max in the last 6 months, and while I cant offer any mathematical evidence, dropping 75bb's in a 1500 hand session shouldnt bother you. If it does I would move down. I would also suggest a 2000 bb bankroll if you were going to try to make your living at 10/20 short. I dont play as high as you do but the varience at 3/6 and 5/10 is still pretty significant. FWIW