RomiChess useing Dann Corbit's smooth formula for 'R'

[Dr.Wael Deeb]

With Dann's formula unchanged Romi is doing quite well even though she has gotten off to a very slow start.

snipped...

Dann's slightly tuned formula for Stockfish should not be very optimal for RomiChess so there is hope for a better result.

Now:
RomiChess96 - Olithinkwin32 : 23.0/31 19-4-8 (=10===11111101==10=1111110=1111) 74% +182

More later!

using formula:

Code: Select all

    double delta = max(h->eval - beta, 1.0); 
    double ddepth = (double)depth; 
    int r = (int)(0.25 * ddepth + 2.5 + log(delta)/5.0);

is better so far than the first test.

RomiChess96 - Olithinkwin32 : 25.0/31 22-3-6 (101=111111111=110101=111111==1=) 81% +252

To have fallen to 80% and then to claw her way back to 82% before allowing some draws is very promising.

Come on Michael,don't you twist it....this version is damn good,it's a release one
Dr.D

RomiChess96 - Olithinkwin32 : 71.5/100 61-18-21 (101=111111111=110101=111111==1=1=1010===1=11111010011101=1=011=1=0111111111=111010111=101110001====0) 72% +164

A new record by 1 point so I guess that I should be happy. But, I am too tired to care at the moment. Arena 1.1 defaulted to 2+2 from 1+1 that I had set. Olithink 5.22 did in the past even better against Romi at 2+2 than it did at 1+1 so this result may be better than it looks. Does anyone have a CCRL or CEGT real or adjusted rating for Olithink 5.22? Is Ot5.22 + 168 better than Romi's CCRL 2514?

Hi Michael,
From my rating list at long time controls:

RomiChess P3k 2339
RomiChess P3j q8 2261
OliThink 5.2.2 x64 q8 2082
OliThink 4.1.3 CCT6 1969

So both of the RomiChess versions are superior to OliThink 5.2.2 and hence the results that you're getting....
OliThink 5.2.6 is already released with LMR and it seems like a big improvement....
Dr.D

[Michael Sherwin]

I had to discontinue the tournament because Arena 2.0.1 was repeating position 32 again every 4th game when Romi had the white pieces. I really love the brown and tan canvas playing board. Oh well back to Arena 1.1 for me and its ugly board.

I am changing the formula to this for the next run.

Code: Select all

    double delta;
    double ddepth;
    int r;

    if(depth < 6) r = 3; else {
      delta = max(h->eval - beta, 1.0); 
      ddepth = (double)depth; 
      r = (int)(0.18 * ddepth + 3.1 + log(delta)/5.0); 
    }

Well, the above one was terrible.

The following one is doing fantastic!

Code: Select all

    double delta = max(h->eval - beta, 1.0); 
    double ddepth = (double)depth; 
    int r = (int)(0.25 * ddepth + 2.5 + log(delta)/5.0); 

RomiChess96 - Olithinkwin32 : 11.5/13 11-1-1 (101=111111111) 88% +346

Don't anyone dare say 'WOWIE!'!

WOWIE
And I'll dare even more to ask you when we will get our hands on this version
Dr.D

As usual, I am seconding Wael.



Best,

RomiChess96 - Olithinkwin32 : 12.0/13 11-0-2 (1=11111=11111) 92% +424

Hello DR. D & DR. T,

Do I dare say it? Do I do I? Well I just can't help myself!

WOWIE!

And now:

Wait .

Wait ..

Not quite ...

Almost ....

Just waiting for a loss .....

Romi made one of her famous rook for knight exchange sacrifices and almost won but Oli refused to die.

RomiChess96 - Olithinkwin32 : 15.5/18 14-1-3 (1=11111=11111=1110) 86% +315

Finally I can go to bed!

Happy New Year!

[Dr.Wael Deeb]

I had to discontinue the tournament because Arena 2.0.1 was repeating position 32 again every 4th game when Romi had the white pieces. I really love the brown and tan canvas playing board. Oh well back to Arena 1.1 for me and its ugly board.

I am changing the formula to this for the next run.

Code: Select all

    double delta;
    double ddepth;
    int r;

    if(depth < 6) r = 3; else {
      delta = max(h->eval - beta, 1.0); 
      ddepth = (double)depth; 
      r = (int)(0.18 * ddepth + 3.1 + log(delta)/5.0); 
    }

Well, the above one was terrible.

The following one is doing fantastic!

Code: Select all

    double delta = max(h->eval - beta, 1.0); 
    double ddepth = (double)depth; 
    int r = (int)(0.25 * ddepth + 2.5 + log(delta)/5.0); 

RomiChess96 - Olithinkwin32 : 11.5/13 11-1-1 (101=111111111) 88% +346

Don't anyone dare say 'WOWIE!'!

WOWIE
And I'll dare even more to ask you when we will get our hands on this version
Dr.D

As usual, I am seconding Wael.



Best,

RomiChess96 - Olithinkwin32 : 12.0/13 11-0-2 (1=11111=11111) 92% +424

Hello DR. D & DR. T,

Do I dare say it? Do I do I? Well I just can't help myself!

WOWIE!

And now:

Wait .

Wait ..

Not quite ...

Almost ....

Just waiting for a loss .....

Romi made one of her famous rook for knight exchange sacrifices and almost won but Oli refused to die.

RomiChess96 - Olithinkwin32 : 15.5/18 14-1-3 (1=11111=11111=1110) 86% +315

Finally I can go to bed!

Happy New Year!

Happy new year Michael and good luck with the upcoming release
Dr.D

[OliverBr]

OliThink 5.2.6 is already released with LMR and it seems like a big improvement....
Dr.D

Actually... Graham Banks replayed a tourney with OliThink 5.2.6 replacing 5.2.5 and it was much worse
But 30 games are nothing, I let OliThink playing 1000 Games against Crafty 20.14 and here the result:

Code: Select all

olithink525 - crafty20.14 : 345.0/1000 266-158-576 34% 
olithink526 - crafty20.14 : 373.5/1000 303-141-556 37%

Looks like an improvement, but it's always the same opponent, so I am not sure about this. I am only sure that you need a lot of games.

PS: If someone is interesting in some result agains Fruit... This is not even funny.

Code: Select all

olithink526 - fruit231_pg : 68.5/1000 41-55-904 6%
olithink411 - crafty20.14 : 104.5/1000 68-73-859 10%

[Edsel Apostol]

OliThink 5.2.6 is already released with LMR and it seems like a big improvement....
Dr.D

Actually... Graham Banks replayed a tourney with OliThink 5.2.6 replacing 5.2.5 and it was much worse
But 30 games are nothing, I let OliThink playing 1000 Games against Crafty 20.14 and here the result:

Code: Select all

olithink525 - crafty20.14 : 345.0/1000 266-158-576 34% 
olithink526 - crafty20.14 : 373.5/1000 303-141-556 37%

Looks like an improvement, but it's always the same opponent, so I am not sure about this. I am only sure that you need a lot of games.

PS: If someone is interesting in some result agains Fruit... This is not even funny.

Code: Select all

olithink526 - fruit231_pg : 68.5/1000 41-55-904 6%
olithink411 - crafty20.14 : 104.5/1000 68-73-859 10%

Hi Olivier,

Fruit has a somewhat complete eval. When you play Olithink against it, even if Olithink search is already strong, Fruit will always take advantage of the weaknesses of Olithink's eval (for example passed pawns).

[bob]

OliThink 5.2.6 is already released with LMR and it seems like a big improvement....
Dr.D

Actually... Graham Banks replayed a tourney with OliThink 5.2.6 replacing 5.2.5 and it was much worse
But 30 games are nothing, I let OliThink playing 1000 Games against Crafty 20.14 and here the result:

Code: Select all

olithink525 - crafty20.14 : 345.0/1000 266-158-576 34% 
olithink526 - crafty20.14 : 373.5/1000 303-141-556 37%

Looks like an improvement, but it's always the same opponent, so I am not sure about this. I am only sure that you need a lot of games.

PS: If someone is interesting in some result agains Fruit... This is not even funny.

Code: Select all

olithink526 - fruit231_pg : 68.5/1000 41-55-904 6%
olithink411 - crafty20.14 : 104.5/1000 68-73-859 10%

Hi Olivier,

Fruit has a somewhat complete eval. When you play Olithink against it, even if Olithink search is already strong, Fruit will always take advantage of the weaknesses of Olithink's eval (for example passed pawns).

Also there are versions of Crafty that are nearly +300 Elo stronger than 20.14 which is several years old. 23.1 is the most recent.