Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Deep Rybka 4.1 Personality
- - By Master Om (Bronze) Date 2011-08-18 16:47
Can any one give me a Deep Rybka 4.1 SSE42 personality for 3-5-10-16 min games but with a contempt that it evaluates Knights better than bishops and avoids Knights exchanges but exchanges opponent knights for bishops ?

Any help appreciated.

Regards
Om
Parent - - By activethinker (**) Date 2011-08-18 17:11
Just increase value of knights.
Parent - - By Master Om (Bronze) Date 2011-08-18 17:16
i know that but how much ?
Give an appropriate answer.
Parent - By Chaotic Chess (****) Date 2011-08-18 17:47
You can use Larry R4 parameter just adjust the values accordingly.
Parent - - By Uly (Gold) Date 2011-08-19 02:35

> games but with a contempt that it evaluates Knights better than bishops and avoids Knights exchanges but exchanges opponent knights for bishops ?


Do you mean "but exchanges opponent bishops for knights"? Otherwise it seems both sides would play equally.

Anyway, here it is:

[OPTIONS]
White Knight millipawns=50
Black Knight millipawns=50
White Light Squared Bishop millipawns=-50
White Dark Squared Bishop millipawns=-50
Black Light Squared Bishop millipawns=-50
Black Dark Squared Bishop millipawns=-50

That's simple to make, but making it behave exactly as you want, that's hard (these values are based on Mindbreaker's values, but may be too small, try bigger numbers like 100 and -100 for stronger effect).
Parent - - By Weirwindle (***) Date 2011-08-19 03:02 Edited 2011-08-19 20:55
He was correct saying "Exchanges opponent knights for bishops" meaning "sacrifices your own bishops to remove opponent knights".
This shows knights as being stronger for both sides.
And for personality options I think this would be best:
White Pawn = 30
Black Pawn = 20
White Knight = 130
Black Knight = 90
White Light Squared Bishop = 90
Black Light Squared Bishop = 70
White Dark Squared Bishop = 90
Black Dark Squared Biship = 70
White Rook = 170
Black Rook = 130
White Queen = 410
Black Queen = 340

This setting will evaluate knights as being stronger than bishops,
minimize knight trades (when the white knight would be lost),
and readily trade off a white bishop for a black knight.
Parent - - By Uly (Gold) Date 2011-08-19 04:21
The problem with your settings is that it also affects how these pieces are affected vs. Rooks, Pawns, or even queens, so that Rybka will be biased to keep Knights and Bishops, and will think the opponent will want to keep Knights and Bishops, agaisnt trading them for other pieces, and omprakash didn't ask for that.

My settings tried to avoid that, by using negative values instead of upping everything.
Parent - By Weirwindle (***) Date 2011-08-19 09:03
Sorry, I also meant that mindbreakers values for pawns, rooks, and the queen should also be used.
Parent - - By Master Om (Bronze) Date 2011-08-19 07:54

>Do you mean "but exchanges opponent bishops for knights"? Otherwise it seems both sides would play equally.


well let me explain you as i know only you can help me as i think.

Question is where is knight better than a bishop ?
Answer is a bit awkward as others think. i.e. Open positions within middle game. and for bishops its closed positions.
All think that closed positions are better for knights but thing is other way round. this is not my idea it's M Chigorin's.
Its the base of my new experiment for some specific openings.
That is why Knights must be preferred more than bishops.
second thing is  bishops are strong in endgames when 2 are there. In most computer games/Human too its rook endgames. But in terms of co-ordination Queen has the greatest when a knight is with her. So to make attack knight is a must. You can see that in games of Mikhail Chigorin and Petrosian who had weakness towards knights.
So that is why  in most opening systems bishops are exchanged for knights to open the system and not to close it !!like in winawer , Nimzo indian , Philidor ,Tchigorin defence, Baltic defence and more.
PM me the settings if u find any.
Parent - - By Uly (Gold) Date 2011-08-19 08:52
Are you going to try the ones I posted?
Parent - - By Master Om (Bronze) Date 2011-08-19 09:16
Already started :lol:
Parent - By Weirwindle (***) Date 2011-08-19 20:54
Edited my post above.
Parent - - By Moz (****) Date 2011-08-28 01:54 Edited 2011-08-28 02:17

>Question is where is knight better than a bishop ?
>Answer is a bit awkward as others think. i.e.
>Open positions within middle game. and for bishops its closed positions.
> All think that closed positions are better for knights but thing is other way round


No offense, but this is a ridiculous statement that flies in the face of a hundred years of chess theory and grandmaster praxis. It also defies logic -- knights have limited mobility and can leap other pieces. Hence, their increased value in the opening stages where pawns are plentiful and the position is closed. How, exactly, does a knight benefit from an open position? Put simply -- it usually doesn't unless there are other positional factors at play!  Oh sure, you can find a few specific positions where the case could be made but in general you have it exactly backwards. Bishops are typically terrible in closed positions which is why their value often increases as the pawns leave the board.

Our very own Larry Kaufman wrote one of the seminal works on the subject of evaluating material imbalances. Anyone interested in creating engine personalities would do well to study his analysis. Creating unbalanced engine personalities is always risky business because the value of pieces can vary considerably depending on the other pieces in play.
Parent - - By Master Om (Bronze) Date 2011-08-28 02:29 Edited 2011-08-28 02:33
No offense, but this is a ridiculous statement that flies in the face of a hundred years of chess theory and grandmaster praxis. It also defies logic -- knights have limited mobility and can leap other pieces. Hence, their increased value in the opening stages where pawns are plentiful and the position is closed. How, exactly, does a knight benefit from an open position? Put simply -- it usually doesn't unless there are other positional factors at play!  Oh sure, you can find a few specific positions where the case could be made but in general you have it exactly backwards. Bishops are typically terrible in closed positions which is why their value often increases as the pawns leave the board.

Did you read properly ? what I wrote ?
I said within middlegame not in endgames and I stand correct What i say. If position get closed in openings then it favours bishop as bad bishops hold good pawns
In the other hand if opening gets opened it favours knights. It is just the contrary what others think. few examples. Obviously none other than Karpov. watch the 4th move!!
Attachment: KarpovvsBrowne.aqs - Karpov vs Browne (98k)
Parent - - By turbojuice1122 (Gold) Date 2011-08-29 20:18
I think that this just shows that Karpov is better than Browne in using his minor pieces. :-)
Parent - - By Master Om (Bronze) Date 2011-08-30 01:04
Although you are right but the 4th move shows the strategy i was talking about. At 4th move he had a chance to close diagonals and he did it. Then opened the centre for his knights.
Parent - - By turbojuice1122 (Gold) Date 2011-08-30 01:20
I think that one could also make a case for the 4th move being to weaken the kingside pawn structure.  Now that ...g6 has been played, the capture gxf6 cannot be played to open the g-file for the rook.
Parent - - By Master Om (Bronze) Date 2011-08-30 02:39
I have the annotated game of this describing this strategy by IM watson
Parent - - By turbojuice1122 (Gold) Date 2011-08-30 14:55
If you can post it here, the rest of us can decide if we agree with the annotations. :-)
Parent - By Master Om (Bronze) Date 2011-08-30 15:42
Ok.
Parent - - By Master Om (Bronze) Date 2011-08-30 16:16
Whether we agree or not on that fact doesnot matter to IM Johhn Watson as its a fact ,so here it is..
Parent - By Uly (Gold) Date 2011-08-30 22:19
Whoa, that's huge, I'm attaching one for people with 15 inch monitor.
Parent - - By turbojuice1122 (Gold) Date 2011-09-01 02:38
What I see in these situations as most important is the worsening of the opponents' pawn structures due to exchanging the bishop for the knight.  I think that is the dominant factor, as opposed to the ideas going against conventional wisdom.
Parent - By Master Om (Bronze) Date 2011-09-01 04:10
there are various factor involved, what you said is one of them.
Parent - - By Pia (****) Date 2011-08-28 01:02
A different contempt:

1   Deep Rybka 4.1 w33  +16/=29/-5 61.00%   30.5/50 //Score Offset millipawns: +250
2   Deep Rybka 4.1 w32  +5/=29/-16 39.00%   19.5/50 //Default

TC: 1 min/game. CPU: Q6600. Hash: 256 MB. Nalimov: full 3-4-5 on USB.
Book: Perfect2011, "Optimize", changing colors.
Pondering: off.
GUI: Fritz 11. Draw: early. Resign: early.
Parent - By Master Om (Bronze) Date 2011-08-28 01:11
Ok I will try that.
Parent - By Uly (Gold) Date 2011-08-28 06:10
:eek:

That's something new, so avoiding draws at all costs seems beneficial?!
Parent - - By ernest (****) Date 2011-08-28 23:40
50 games??? :cool:

The 95% probability low end is 26/50 or 52%

omprakash said: Ok I will try that

Let's wait and see!
Parent - - By Pia (****) Date 2011-08-29 03:46
Here's another 77 :wink: games:

1   Deep Rybka 4.1 w33  +16/=45/-16 50.00%   38.5/77  0.00
2   Deep Rybka 4.1 w32  +16/=45/-16 50.00%   38.5/77  0.00

Changed: Rybka TB access freq. 256, GUI draw & resign: never.
Parent - By ernest (****) Date 2011-08-30 17:15
Now what? :cool:
Up Topic Rybka Support & Discussion / Rybka Discussion / Deep Rybka 4.1 Personality

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill