RE: [Bug-gnubg] [Q] EvaluatePosition(), Utility(), and testing boards

[Thomas Hauk]

On Wed, 8 Oct 2003, Ian Shaw wrote:
> Joseph is apparently on holiday for a couple of months. However, I have
> a large chunk of his position database, since I did a lot of the
> rollouts. I have positions in the following categories: Race, Contact,
> Crashed, Doubles. Contact is probably the most general category.
> 
> Would you like me to send you a sample to play with?

Absolutely. Send me everything you have, actually (if you can).

> There is, however,
> a problem. The position ID used in the scripts is NOT the gnubg position
> ID.

Sigh, another instance of Not Invented Here syndrome!

> Joseph mentioned once that they could be converted with a script
> (python?). He offered to write one, but I didn't take him up on it. 
> 
> The rest of the file format is described here: 
> http://pages.quicksilver.net.nz/pepe/ngb/index-top.html
> http://pages.quicksilver.net.nz/pepe/ngb/src/index-top.html#GNU%20rollou
> t%20program

Ok, this doesn't make sense to me:

"The positions are encoded as 20 character in the range of `A' to `F'. 
Each letter stands for 4 bits (A == 0x0, F == 0xf)..."

If A is 0x0, then shouldn't F be 0x5? Err...

I downloaded the code to take a quick look, and I think he meant to write 
A == 0x0 and P == 0xf.

> Does anyone know how to convert Joseph's ID to the gnubg ID? This is a
> typical position LHNJCFAEAAOLAAAAMAAA
> It looks like the source is available, so perhaps you can extract the
> format from that.

Yup, looks like he's just doubled the 10-byte ID to 20 bytes by encoding a 
nibble at a time. I don't really understand why he didn't use the 10-byte 
key to begin with... 8P

I was also given this URL:
http://www.cs.arizona.edu/~gary/backgammon/positionid.html

-- 
Some people, when confronted with a problem, think "I know, I'll use 
regular expressions."  Now they have two problems.
  --Jamie Zawinski, in comp.lang.emacs