How long does it take to know you're winning in LHE?

[basementproject]

[pzhon]

Here is an excerpt from a past post:

<ul type="square">
There is a useful unit I call the long run, which I define to be how long it takes for breaking even to be 2 standard deviations below average. If your win rate is WR/hour and your standard deviation is SD/hour, this is 4 (SD/WR)^2. For example, if you win 1 BB/hour with a standard deviation of 10 BB/hour (reasonable for many [live] midstakes limit games), the long run is 400 hours.

If your win rate is closer to 0, the long run gets longer. Cutting your win rate in half increases the long run by a factor of 4.

The normal approximation is good at that point, so the probability of being behind when you reach the long run is about 2.5%.

The long run is also the point at which there is a 50% chance that 0 will not be in the rough 95% confidence interval of results +- 2 standard deviations.

At 4 times the long run, breaking even is 4 standard deviations below average, and there is about a 97.5% chance that 0 will not be in your rough 95% confidence interval.

You can be confident that you are a winning player with fewer hands if you pay attention to more than the results. If you notice that your opponents make particular common and costly mistakes, you might be confident that you are a winning player after relatively few hands. For example, when you see players show down garbage from early position, or routinely cold-call with hands that aren't worth limping, you know the game is soft, and you may be able to judge that it is softer than other games you have beaten. [/list]