High Variance Play

[mrface]

I have been playing on party poker (and lurking these boards) for the last 6 weeks or so and have been fairly successful. Started with a $250 bankroll in .5/1 and am now playing 2/4 with a $1500 bankroll. I think I play winning poker and my results seem to confirm that. However, I am worried about my standard deviation.

My pokertracker database of 11,000 hands (about 90% of all the hands i've played .5/1-2/4) says I win 5BB/hr and have a standard deviation of 17BB/hr. I dont know a lot about SD but this seems rather high. I would consider my play style to be what Sklansky describes as "fast" (limping with Axs from any position, raising with probable 2nd best hands; see my post on SS forum for a specific example) so a high SD does not surprise me, but it does worry me. Is 17BB/hr too high of a SD?

I play poker both for the money and for the intellectual challenge. Part of that challenge is moving up in stakes as I reach various bankroll goals. As I progress I have been trying to fine tune my game, but this is one of those things where I'm unsure whether or not it even needs to be altered. My main worry is that while I do make mostly +EV plays I may also be making many -EV plays whose effects are dampened by the poor quality of opposition. Would a high SD be indicative of this type of play?

My main goal is to become a better poker player and to continue to successfully progress through the limits (isn't it everyones?). To this end I'm hoping to catch any of my larger leaks before I run into opposition who will catch them for me. I really appreciate any help I get on this subject and I think it's an interesting topic to debate (merits of fast play) so hopefully this generates some good discussion.

In Summary:
Is a high SD necessarily bad? Does it indicate bad (-ev) play? What is a "normal" SD? Can I survive at higher limits with such high variance play?

[Nottom]

[ QUOTE ]
My pokertracker database of 11,000 hands (about 90% of all the hands i've played .5/1-2/4) says I win 5BB/hr and have a standard deviation of 17BB/hr. I dont know a lot about SD but this seems rather high. I would consider my play style to be what Sklansky describes as "fast" (limping with Axs from any position, raising with probable 2nd best hands; see my post on SS forum for a specific example) so a high SD does not surprise me, but it does worry me. Is 17BB/hr too high of a SD?

[/ QUOTE ]

I'm sure I won't be the first to say this, but its unlikely your true winrate is 5BB/hour (especially at the 2/4 level) 11k hands just really isn't enough to narrow it down. I'm not claiming that you aren't a good player, just that you are likely running good right now and I would expect that number to drop into at least the 2-3BB/hour range.

Also an SD of 17BB/hour isn't bad for that winrate. In live play a SD of 8-10 for a 1BB/hour winner is pretty good, so in a game with more hands and higher winrates your SD should be a bit higher and 17BB is certianly not unreasonable especially in some of the crazy loose Party games.

[Bozeman]

Are you playing multiple tables or shorthanded tables?

If you specified your WR and SD per hand or 100 hands it might make this question less relevant.

Craig

[mrface]

I almost always play only 1 table at a time and not the SH tables.

[BruceZ]

Unless you are playing 5 or more tables at once, or the games are moving very fast, your SD is not high, it's very low.

SD for 1 table live should be about 10 times your hourly rate. Here are the equivalent factors for online play, if we assume you get twice as many hands per hour per table online.

<font class="small">Code:</font><hr /><pre>

# tables SD/win rate (1 hr) SD for win rate = 5 bb/hr
1 live 10 50 bb
1 online 7 35 bb
2 online 5 25 bb
3 online 4.1 20.5 bb
4 online 3.5 17.5 bb</pre><hr />

To get the new rule of thumb for SD/win rate online, just take SD/sqrt(x*tables), where SD is for 1 hour of live play, and x is how much faster the online games are compared to live. I took SD = 10, and x = 2 to assume the online games are twice as fast.

[BruceZ]

Did you mean your hourly rate is 0.5 bb/hr rather than 5 bb/hr?

If it's 5 bb/hr, then an SD below about 35 bb for 1 hour online is good.

[mrface]

Interesting. I just assumed mine was high since most of the SD's I had seen quoted were around 7-10. Thanks for the good info.

[Louie Landale]

I think your main concern is your whopping 5bb/hour. You've been getting very lucky.

Your results are consistent with with a boarderline maniac who has gotten VERY lucky.

Your results are also consistent with a solid player who routinely pushes small EV high varience playes, who also got lucky.

Your results are NOT consistent with a knowledgable but inexperienced newbe.

The point here is this: [1] I'd be very conserned that you've been making numerous -EV high varience plays, but it hasn't come back to bite you yet. [2] You are in for a RUDE shock when the down swing comes and you drop $800 in your 2/4 game in 2 weeks.

I suggest that after every session, you take one sample high-varience play and figure out in detail exactly whether or not it really WAS +EV. You'll need to do such things as evaluating the possible distribution of possible hands the opponents may have, which is a nightmare, or you may want to run some simulations.

I also suggest you play less hands to start. That will tend to stop you from "improving" to high varience but -EV situations after the flop.

- Louie