[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

Calculating confidence values

Posted By: Maik Stiebler
Date: Tuesday, 1 September 2009, at 12:15 p.m.

In Response To: Calculating confidence values (Timothy Chow)

I hesitate to guess how different the answer will be without doing some further estimates or calculations.

I did a bit of numerical experimenting. After 100 trials (two plays, DMP, no VR) the p-value of a one-sided t-test is an excellent approximation of the Bayesian estimate of the probability that play A is better than play B given a uniform prior distribution. The approximation gets better when more trials are done.

So, if this is a representative setup, what we have been doing to calculate confidences all the time is a pretty good replacement for the Bayesian analysis (with uniform prior) that you describe (granted, only as long the hypotheses are not too complex). Maybe some users and bot developers will find the current way simpler. I kept inquiring about this because I still believe that what you originally thought (that bots already use the Bayesian approach) is in some sense very close to being true.

Messages In This Thread

 

Post Response
Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:
 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.