Explain confidence rates? (winrate/sample etc)

[NL Newbie]

So say i play 2PTBB with 32 Standard deviation 99% ocnfidence.

Says i could be between -3.83 PTBB up to +7.83 PTBB

So what can i conclude, im a winner? Or could be anything?


As the majority if positive, is it most likly im a winner???? [img]/images/graemlins/blush.gif[/img] [img]/images/graemlins/blush.gif[/img] [img]/images/graemlins/blush.gif[/img]

[DrVanNostrin]

You can't conclude anything at the 99% level. But you could conclude you were a winner at a lower confidence level.

[SunOfBeach]

Compile a larger sample, and wait until the bottom number is higher than zero. Then, you can conclude (with 99% confidence) that you're a winner. Kind of.

In any situation in which the mean result is 2 with sd of 32, it'll take awhile to get much confidence of anything...

[NL Newbie]

So i cant conclude that since the range is more in the +$ that i'm a winner more often than not?

Seems ill need 100k before i can get 'confident' with anything.

[jay_shark]

This post should help you .

http://forumserver.twoplustwo.com/showth...ue#Post12047089

[pzhon]

You have some evidence that you are a winning player. At the moment, it is not very strong evidence. As you play more hands, if your win rate stays the same, your evidence will gradually increase in strength. There isn't a set number of hands at which your results suddenly become meaningful.

It usually takes a much larger sample to get strong evidence that you are a winning player if your win rate is lower, like 2 PTBB/100, than if it is higher, like 10 PTBB/100.

If your win rate is low, you may have strong evidence that you are neither a big winner nor a big loser long before you determine whether you are a winning player.

[mykey1961]

Also something I learned the hard way.

You are only confident that you were a winner against opponents who played they way your opponents did at the time you beat them.

If you get new opponents (change levels) or your opponents improve, you're back to square one.