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

BGonline.org Forums

statsig and j.s.d. question

Posted By: Matt Cohn-Geier
Date: Friday, 31 October 2008, at 9:31 p.m.

In Response To: statsig and j.s.d. question (Barry Silliman)

Am I correct in assuming that 'statsig' is shorthand for statistically significant, and that you're referring to the j.s.d. measurement? Or is there another measurement I also need to look out for? And assuming it is j.s.d., it sounds like most of the time people are looking for at least 3 (maybe 2) j.s.d. to feel really comfortable about it, is that about right?

You are correct here. The classical confidence interval for X JSDs is given at http://www.bkgm.com/rgb/rgb.cgi?view+1238, but there is bias in the rollout so it will be somewhat lower than that.

This leads me to a question regarding gnubg rollouts...let's say I've set up a certain number of trials, e.g. 1296. If the j.s.d. number is less than my desired goal, can I extend this rollout in gnubg or do I need to start over. IIRC correctly I can extend a Snowie rollout but I'm not sure how to do this with gnubg. (I'm using the 0.90 build, I think that's pretty recent).

You can extend the rollout result. Just click rollout again.

My rollout default # of trials is 1679616 (the maximum), and I just stop the rollouts when they reach X JSDs.

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.