View Full Version : Is a low SD indicative of anything?

12-13-2004, 06:21 PM
As I continue to analyze every facet of my play at 2/4, I recently noticed that my SD is significantly lower than it was at .5/1 and 1/2.

I have a 22K sample size at .5/1 where my SD is about 17.5

In my 13K at 2/4, it's significantly lower, at just above 15.

Over the last 5K hands or so, it's much lower at just above 12. This has been the worst part of a bad stretch, and I'm wondering what, if anything, this low SD indicates.

Other relevant stats:

.5/1: 16.5
2/4: 13.5

.5/1: 8
2/4: 7.8

.5/1: 2.8ish
2/4: 2.4ish

12-13-2004, 07:02 PM
IIRC, Mason Malmuth wrote in "Gambling Theory and Other Topics" that good players can have surprisingly low SD's.

On the other hand, it could mean you are playing too tight, including after the flop. Or you may be running bad and correctly dumping hands that don't pan out. How often do you see showdown? How often do you win at showdown?

Here are some thoughts (more or less as I've posted elsewhere):

If you play as tightly as possible (fold every hand) your SD will be zero. If you play every hand to the end, your SD will be quite high.

Aggression, both yours and your opponents', tends to force hands to end before showdown. Your good hands hold up more often, but you don't win as much. When you lose you don't necessarily lose the maximum because you folded to someone else's aggression. So variance may actually go down due to aggression.

Variance should be maximized when looseness is coupled with aggression. Say when muliple players are capping every street and seeing showdown. In this case you often lose a lot when you lose, but also win huge pots when you win.

As you go up in limit, aggression tends to go up and players are generally tighter. So if the above ideas are right, SD will tend to go down as a player moves up. This seems to be true in most instances where players have posted results at multiple limits.

12-13-2004, 08:36 PM
How often do you see showdown? How often do you win at showdown?

[/ QUOTE ]

Went to showdown (when flop seen) are similar at around 29%, if I remember correctly, but I'm winning at a higher % at showdown at .5/1... something like 55% versus 51.5%.

Thanks for the clarification... it makes sense.

I know I'm definitely not getting called down by as many players at 2/4 as I was at .5/1... so that probably has a lot to do with it.

12-14-2004, 12:32 AM
I should have read your original post a little more carefully.

Over the last 5K hands or so, it's much lower at just above 12.

[/ QUOTE ]

This is a very small sample. One of the guys here is working on trying to see how quickly SD should converge, but the chart below indicates that it can change significantly over this kind of sample size. [This chart represents my running SD over ~13K hands in a live kill game. Also, I went from raw beginner to raw whatever I am now over this time period, so this data isn't ideal.]