New commercial engine soon: Onno

[bob]

Then why does random seed make the engine slightly stronger?

The theory, why randomness (not random seed) makes an engine stronger is as follows:

Static evaluation isn't exact. There might be a systematic error that over- or underestimates all positions that are examined in a search from the current position.

As the evaluation function is inherently inaccurate, a player is better off if there are more good positions he can reach. (He has several good moves rather then one.) So we want to add the number of good moves to the evaluation function.

As counting such moves is difficult to integrate in a search, instead the position values are randomly slightly modified. The more good moves a player has, the better are the chances that he can profit from the random modification.

This idea isn't new. It was discussed eithere here in the programmers forum or on the Winboard forum. (Also it is known for years that an engine with completely random evaluation beats a random mover.)

P.S. Now you can see what a mistake it was to mention this seed thing among the engine options.

Maybe it is in deed a too complicated option with only moderate user benefit.

This is essentially mobility, which can really be evaluated in other ways more accurately...

[menniepals]

Do you have a freeware version ?

[peter]

By the way, how did you come to Crimson as for your engine formerly known as Alderon?

Who is this Crimson and has he anything in common with the one we know together with clover or the one that Mark Knopfler would kill to get him? ("Kill to get Crimson" his latest CD as far as I know)

[Onno Garms]

This is essentially mobility, which can really be evaluated in other ways more accurately...

Mobility counts all moves, it's important to count good moves. Of course there is also mobility evaluation in my engine.

My tests didn't show a significant difference between randomized an not randomized evaluation. Scores of the randomized versions were better, but maybe that was just coincidence. So what could I do under the presumption that I don't have the resources to run 10 or 100 times as many games? Right, naiively use the version that performed better in my 1000 games.

Also I found it a nice idea to offer the user an option that might change the behaviour of the engine in a specific position but does not affect search or evaluation weights.

[bob]

This is essentially mobility, which can really be evaluated in other ways more accurately...

Mobility counts all moves, it's important to count good moves. Of course there is also mobility evaluation in my engine.

My tests didn't show a significant difference between randomized an not randomized evaluation. Scores of the randomized versions were better, but maybe that was just coincidence. So what could I do under the presumption that I don't have the resources to run 10 or 100 times as many games? Right, naiively use the version that performed better in my 1000 games.

Also I found it a nice idea to offer the user an option that might change the behaviour of the engine in a specific position but does not affect search or evaluation weights.

I have a random feature in the evaluation of Crafty. And when it is enabled, in any degree, it _weakens_ the program. I even ran some calibration runs on our cluster, 40,000 games at a whack, to discover exactly what effect the various levels of randomness had on actual Elo.

I've never seen any reported case of adding randomness to improve performance. I read Beal's paper, but that was for a program with a _pure_ random evaluation.

I'll run some more tests to see if there is any "win" from adding randomness, but there certainly was not the last time I exhaustively tested this idea.

[F. Bluemers]

The King :

[Graham Banks]

The King :

Some good music on that particular album!

[gerold]

The King :

Some good music on that particular album!

Hope the music is better than the pic.

Hows the testing going on the new engine with you.
I have been testing it. Not m uch to brag about.

[kranium]

By the way, how did you come to Crimson as for your engine formerly known as Alderon?

Who is this Crimson and has he anything in common with the one we know together with clover or the one that Mark Knopfler would kill to get him? ("Kill to get Crimson" his latest CD as far as I know)

Hi Peter,

i'm absolutely astounded that you that hinted here at 'Crimson and Clover'...
(for those that are unfamiliar with it)...it's an old song by Tommy James and the Shondells from the sixties...considered by many to be a true (psychadelic) rock classic.

i remember the tune very well, as if it was yesterdsay, as it had a profound affect on me...(i was a young teenager at the time).

as for the more recent Mark Knopfler tune...i'm not familiar with it, but remember very well the top albums of the 80s that Dire Staits produced, and Mark's crystal clear powerful ringing guitar riffs.

Norm

[kranium]

The King :

Hi Fonzy-
I'm not very familiar with King Crimson...but from what i have read they were very original and influential...