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

BGonline.org Forums

Further proof that you must roll it out

Posted By: Timothy Chow
Date: Thursday, 22 October 2015, at 8:30 p.m.

In Response To: Further proof that you must roll it out (Stick)

Let me remark that when this happens to me, often I get different results depending on whether

1. I analyze the move inside the match file

or

2. I copy and paste the position into a separate instance and analyze it in isolation.

XGR+ and XGR++, I believe, use a seed that is a hash function of something, and that "something" might differ depending on whether the position is part of a match file or part of a position file. (Even if it's not supposed to make a difference, it could actually make a difference if indeed this is a bug.)

I did save an instance of one of these anomalies not too long ago and I emailed to Xavier, but I didn't get a response. I assumed that this was a low-priority bug fix and didn't pursue it. However, I'll try to dig it up again and send it to you.

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.