Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hash and first Fail High

Author: Robert Hyatt

Date: 20:14:45 08/16/03

Go up one level in this thread


On August 16, 2003 at 07:23:49, Gian-Carlo Pascutto wrote:

>On August 15, 2003 at 22:56:34, Robert Hyatt wrote:
>
>>On August 14, 2003 at 15:27:06, Peter Fendrich wrote:
>>
>>>1FH = The ratio where the first generated move in the node is a Fail High.
>>>
>>>I experimented with removing the the hash table and became a bit surprised.
>>>Normally I have about 1FH = 95-96%. When I removed the hash table that figure
>>>raised to more like 1FH = 96-97% while the search depth, as expected, was
>>>decreased.
>>>
>>>Increased 1FH when removing the hash table. Is this normal?
>>
>>Probably.  You increased the size of the tree.  And for the extra stuff you
>>searched, you searched it pretty accurately.  But you _did_ search it, which
>>slowed you down.  :)
>
>No! - That does not make any sense at all.

Sorry, but it makes _lots_ of sense.

If you don't try a hash table move first, then you will probably try captures
and captures have re-captures that work instantly.  Hence more cutoffs.  Just
try it...

The hash move is often _not_ a capture.  But if you try a capture first,
then the corresponding re-capture will be an instant cutoff that is easier
to find than a reply to a non-capture.

>
>"the extra stuff you searched, you searched it pretty accurately"
>
>If we disabled hashing and hence handicap move ordering, we'll do
>worse, not better, on average.
>
>Yes, we may generate extra nodes and do well there, but that's certainly
>not what we should expected in the average case, since we just made
>our move ordering worse!
>
>Do you see your contradiction?

Not at all.  See above.  Which is easier to refute, a non-capture or a
capture?

Answer:  A capture.

>
>--
>GCP



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.