Romichess P3k test

[Jim Ablett]

Hi Mike,

As soon as I understand Jim's code I will try to make things a bit more robust.

I was hoping you'd say that. Coding is not my strong point.
While you're looking at it Mike, it would be nice to have modifiable
piece values in there. I couldn't find an easy way to adjust them.

regards,
Jim.

[Michael Sherwin]

Hi Mike,

As soon as I understand Jim's code I will try to make things a bit more robust.

I was hoping you'd say that. Coding is not my strong point.
While you're looking at it Mike, it would be nice to have modifiable
piece values in there. I couldn't find an easy way to adjust them.

regards,
Jim.

Hey Jim,

You cut the road through the mountain!

Time for me to put down some pavement.

[Tony Thomas]

Hi Mike,

As soon as I understand Jim's code I will try to make things a bit more robust.

I was hoping you'd say that. Coding is not my strong point.
While you're looking at it Mike, it would be nice to have modifiable
piece values in there. I couldn't find an easy way to adjust them.

regards,
Jim.

Hey Jim,

You cut the road through the mountain!

Time for me to put down some pavement.

I will owe you one if you put down some pavement. Makes it easier for me to meet goat at the other end of the mountain and boogie.

[Tony Thomas]

Here is the first Romi rating list calculated with Bayesian Elo. As you can guess, Romi didnt perform as well as expected against Djinn, she only managed 15.5/30, which was 2 points lower than the requirement to catch up with NG9. Still the rating differences are so minor that no conclusion can be drawn with a 30 point error margin.

Code: Select all

Rank Name                   Elo    +    - games score oppo. draws 
   1 ChessTiger2007.1 UCI  2725  101   85    60   87%  2432   17% 
   2 ChessTiger2007UCI     2721   39   36   390   87%  2404   19% 
   3 Naum 2.0              2652   29   28   600   80%  2396   17% 
   4 Zappa 1.1             2497   22   21   840   65%  2374   16% 
   5 Arasan 10.0           2469  208  140    26   94%  2086   12% 
   6 Francesca MAD 0.13    2449   21   20   870   58%  2388   20% 
   7 Arasan 9.5            2449   21   21   870   58%  2388   18% 
   8 RomiChessNG9          2438   30   30   420   57%  2384   18% 
   9 RomiChessP3k          2432   30   30   420   56%  2384   20% 
  10 RomiChessNG2          2427   31   30   420   55%  2384   16% 
  11 RomiChessNG8          2424   30   30   420   55%  2384   19% 
  12 Goatka v1.0 Beta.w32  2423   31   30   420   55%  2384   17% 
  13 RomiChessDK5          2419   33   32   360   61%  2334   16% 
  14 RomiAlphaNG1          2416   36   36   290   50%  2413   23% 
  15 RomiChessDK6          2415   30   30   420   54%  2384   18% 
  16 RomiChessNG3          2410   31   31   390   57%  2358   18% 
  17 RomiBetaNG1           2407   30   30   420   53%  2384   18% 
  18 RomiChessDK3          2388   32   32   360   57%  2331   18% 
  19 RomiChessDK4          2380   29   29   420   57%  2329   21% 
  20 Danasah 2.85          2378   20   20   871   49%  2390   23% 
  21 Phalanx Reborn        2374   21   21   840   48%  2389   15% 
  22 RomiChessP3j          2373   39   39   248   53%  2347   15% 
  23 RomiChessDK2          2359   32   32   360   54%  2331   16% 
  24 Delphil 1.6c          2342   21   21   847   43%  2392   23% 
  25 Djinn 0.925x          2329   21   22   780   45%  2368   17% 
  26 Horizon_4_3_173       2326   24   24   660   42%  2387   17% 
  27 Horizon_4_3_175       2313   32   32   360   46%  2338   18% 
  28 RomiChessNG6          2313  188  147    15   83%  2086   20% 
  29 Horizon 4.3           2291   29   29   420   43%  2343   19% 
  30 NanoSzachy 2.7        2287   22   22   780   38%  2381   15% 
  31 Zeus 1.28             2272   22   22   780   37%  2372   17% 
  32 GreKo 5.2             2216   22   22   830   29%  2384   17% 
  33 Lime 6.3              2086   25   26   911   15%  2401   10% 

[Tony Thomas]

From looking at this table, I have a feeling that Romi will be spending another few years in Leo's 4th division unless you come up with something rather innovative that fits in with the style of Romi.

[Michael Sherwin]

P3k will do very well at 40/40 with the DWD book. We hope!

[Dr.Wael Deeb]

P3k will do very well at 40/40 with the DWD book. We hope!

It will....It's improving rapidly in my rating list with long time control,I
ll post results later this spring

[ozziejoe]

I keep building her knowledge in the one caro kann position.

You can see in the last round she went undefeated against much stronger opponents. this is 4 min 2 sec time control. I will change the time again once she masters this

You'll notice romi started losing in middle of this tourni (see after rybka's last loss). this is because I changed time control from 1 min 2 sec to 4 min 2 sec

Hopefully, her knowledge will be general enough, eventually, to work at any time control

Engine Score Ro
1: Romichess 359.5/690 ···················································································································
2: Rybkav2.3.2a.w32 90.5/115 1=1111==111111111111=11111101110=1=1=11111111111111111111111=1010=0101010=010=0=00101===11=111=1=111=11=1111111111=
3: Movei00_8_438 79.5/115 111111=1=11=111111111111111111111011=11====11111==11101000000010=01010101010101010=0111111==1111=1111===1=1=1000=00
4: Slow 49.5/115 11111===0111=01010=011101=00=0110=01010111110=0100000=0=0000000=000000000=0000010100=101=111110=0=01010101010101010
5: List512 45.5/115 =011110001010100010101010101==00000=0000000101010=00000000000000010=010=0101010=0100111=111101011101010101010101010
6: Ruffian_210 35.0/115 11=10=0100=101=1=1=10101==0==1=0=10==1==0001=0000000=000000000=0000000=0=0=000=0000001111==10000001001000000000=000
7: Yace Paderborn 30.5/115 01011=0==10=0=010100==0=000==100000000000000=00=0=010000=000000000000=0==00001000=0===0=0=000=01====00011=0101==010

[ozziejoe]

by the way, michael, it would be fantastic if you could find a way to turn that learn file into an opening book file (e.g., in ctg or arena format.)

best
J

[Michael Sherwin]

I hope to be able to one day have enough time to ...

Work has all but stopped on the new program and I only have a little time to make small changes to the current program.

Before I made progress by sacrificing sleep, however, as a result I have become quite ill. Now I make sure that I get all the sleep that I need and still recovery from illness has been slow.