Can Anyone Answer This?

I play limit Hold'em. My hourly Standard deviation is less than 8 big bets. What does it mean that my hourly standard deviation drops (to about 7 big bets/hr.) whenever I'm running bad and throughout a losing streak? This seems weird. I would think it should rise, since I'm experiencing results which are not normal (i.e. losing). On the other hand, maybe it makes sense, since I have a low standard deviation anyway and am able to fight off or at least cut big losses. But I would think that any time you experience results which are not outside the norm (for me, losing), standard deviation would rise.


btw- I AM a winning player. Trust me on that. At least + 1sb/hr. probably a little higher. Thanks for any input.

Are you using your overall average or your average just for the sessions on the losing streak to calculate SD? Obviously, if you were using the average from just the streak, your SD would be low because all of your results are similar (hence the streak).


If you are using your overall average, then your SD should be the same if you are winning or losing, unless you are a tilter. Remember, the session/hourly results are squared so it doesn't matter if they are +/-.

Thanks. I suppose this makes sense. Yes, if I use just the numbers during the streak they are very low, while the total average is not affected much. But still...


When you run a standard deviation or two outside the mean, shouldn't SD rise? For instance, I've been running flat for about almost 400 hours now. This is very unusual and I have to think I am at least between 1 to 2 standard deviations off my mean. I would think my SD would rise, not fall or stay the same. Of course, I am still very much looking for leaks in my game which may be causing recent poor results. Basically, I had two big slides and have yet to make up for them. (I do not tilt, at least not in any significant way).


Your mention that a low SD is expected during a bad streak (hence the streak), makes sense. But the way I understand it, the lower your SD the more likely it is that your results are accurate or at least not being subjected to a significant amount of luck. I also understand that about 30 samples is enough to get an accurate SD figure. So if you're running bad over 30 sessions and have a very low SD, it would seem to me an indication that your results are probably accurate. Yet, I know this is not the case. I'd just like to understand this a little better. Thanks.

"What does it mean that my hourly standard deviation drops (to about 7 big bets/hr.) whenever I'm running bad and throughout a losing streak? This seems weird. I would think it should rise, since I'm experiencing results which are not normal (i.e. losing)."


If you win 1sb/hr, there is almost no difference between winning and losing. If you lose 6BB for 5 sessions a row, your SD is not high, but you are losing. If you lose 1BB for 5 sessions in a row, your SD is very low, but you are losing. With an SD = 8BB/hr and EV = +1/2BB/hr, you can expect to win 8.5BB as often as you lose 7.5BB in an hour. Losing is not abnormal since SD is so much greater than EV.


Your chance of being -1/16SD's or less (even or losing) for an hour (from a z-table) is about 47.5%. Therefore, your winning sessions wouldn't really affect your overall SD much more than your losing sessions. Your data is consistant with this.


One reason your SD may be smaller when losing is that your EV will be higher in higher SD games because people are often playing too loose in these games. So it would make sense that you are making less money and experiencing less variance when the other players are tigher.


-g-

It seems poker is skewed, meaning you lose slowly and win quickly.


You lose a dollar every 10 minutes, except make six dollars once an hour.


El-Roi

Of course. This makes perfect sense since you can only lose what you put in the pot but you can win what all of the other players put in. Once you start talking about entire sessions or groups of sessions I think that this effect is somewhat nullified, but this too may account for the small difference between winning and losing SD, since you can only lose what you bet, but can win 10x that, making a huge winning session more likely than a big losing session.

Winning increases your SD much more than losing, as (except for blind vs. blind scenarios) you always win more (and typically much more) on one hand than you lose. Hands you fold preflop do not contribute much to your SD, since your average win per hand is basically zero (say, +0.1 BB for a world class player and -0.4 BB for a maniac on tilt), and the quadratic distance to 0 BB, -1/4 BB (small blind) or -1/2 BB (big blind) can be safely ignored.


It's the hands you play which make up your SD and esp. the hands you win. If you typically either lose 2 BB or win 6 BB in each hand you play, a single win increases your SD just as much as 9 lost pots.


Of course, it doesn't make a difference if you calculate your SD per hand or per hour (SD/h=sqrt(hands/h)*SD/hand) as long as you use a large enough sample. I wouldn't be surprized if the SD of all your losing hours, is only half or less than your winning hours' SD. If it isn't, then you have a serious problem with tilting. (You would have to take hourly or even finer grained notes of your stack size, for this calculation to be feasible).


cu


Ignatius

Sparky,


I too have quite a bit of logged hours in limit holdem and comparing different games I've played it seems to me that a SD of 8 BB/ hr means a relatively passive game (i.e not overly aggressive) a game which is easy to beat but in which players do not give a lot of action.


Is this the type of game you are playing? If so, a losing streak may correspond to below average cards. If this is resulting in losses they are probably not huge losses (like those you would experience in a loose-aggressive game ... hence a larger SD) and this would mean a lower SD than when you are winning (i.e you are raising more, winning/losing bigger pots i.e bigger swings).


BTW, the losing skid for which you calculated a lower standard deviation: how many hours is it and how much smaller is the SD.


It may just be insignificant.


Later,

Joe

oops, didn't see your statement of 7BB per hour when you are running bad...


how many different losing streaks produce this lower SD?