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

BGonline.org Forums

Inconsistent XGR++ Evaluations

Posted By: Timothy Chow
Date: Monday, 19 June 2023, at 4:47 a.m.

In Response To: Inconsistent XGR++ Evaluations (Bob Koca)

Bob Koca wrote:

Is XGR ++ essentially a truncsted rollout? If so could the seeds being different for some reason explain the different results?

Well, Xavier has said that the seed for XGR/XGR+/XGR++ is supposed to be a combination of a hash of the position and a fixed user seed. So I think that the intent is for the result of XGR/XGR+/XGR++ to be deterministic and reproducible. But evidently, that is not always the case.

I'm assuming that Nack and David have checked a lot of positions and found that XGR++ gives consistent results in most cases, but maybe it's worth repeating the experiment (of soliciting XGR++ results from many different users) for some randomly chosen position to confirm that there's something peculiar about the Nackgammon starting position.

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.