Prodeo 2.2 SMP Krazy Klown

[carldaman]

Any hints on a release date?

No idea. I am currently wrestling with a bug. I have no clue, yet. Probably another blast from the previous Millennium like this one that stayed under the radar for decades but became visible and reproducible after version 1.86.

[d] 1rB2R2/5p1k/5Ppp/6P1/1p1R3P/1b2K3/2r5/8 w - -

1. Rdd8 with the announcement of a mate in 8.

It's only a draw, fixed it.

haha, decades!

Just don't take decades to get this new release out, I can't wait to test those new parameters!

BTW: Is there any way you can make the WB2UCI file use the correct mate-score? If that's the problem, that is.

Even in new Prodeo sometimes if he is getting mated, he shows a weird matescore and the GUI adjudicates FOR him.

I'm not very sophisticated about this type of thing, maybe there's an easy fix, but I'm using Fritz GUI if that helps.

Hi Brendan,

You can include this line

Code: Select all

MateScore = 32767

in your ProDeo wb2uci.eng file, under [options].

CL

[BrendanJNorman]

Any hints on a release date?

No idea. I am currently wrestling with a bug. I have no clue, yet. Probably another blast from the previous Millennium like this one that stayed under the radar for decades but became visible and reproducible after version 1.86.

[d] 1rB2R2/5p1k/5Ppp/6P1/1p1R3P/1b2K3/2r5/8 w - -

1. Rdd8 with the announcement of a mate in 8.

It's only a draw, fixed it.

haha, decades!

Just don't take decades to get this new release out, I can't wait to test those new parameters!

BTW: Is there any way you can make the WB2UCI file use the correct mate-score? If that's the problem, that is.

Even in new Prodeo sometimes if he is getting mated, he shows a weird matescore and the GUI adjudicates FOR him.

I'm not very sophisticated about this type of thing, maybe there's an easy fix, but I'm using Fritz GUI if that helps.

Hi Brendan,

You can include this line

Code: Select all

MateScore = 32767

in your ProDeo wb2uci.eng file, under [options].

CL

Oh Thanks Carl!

[Kohflote]

Hi Brendan,

Would you be sharing your Rodent III Kramnik personality?

Thank you & regards,
Koh, Kah Huat

[BrendanJNorman]

Hi Brendan,

Would you be sharing your Rodent III Kramnik personality?

Thank you & regards,
Koh, Kah Huat

Yep. Just as soon as I create the polyglot opening book (forcing it to use Kramnik's opening repertoire from 1995-2002) to go with it.

I plan in future to make my personalities with opening book and some other stuff together.

Otherwise you get people forcing my Karpov personality to play the Najdorf Sicilian (or something else which is like kryptonite to his style), losing and then saying "this plays nothing like Karpov!"

[Rebel]

Any hints on a release date?

No idea. I am currently wrestling with a bug. I have no clue, yet. Probably another blast from the previous Millennium like this one that stayed under the radar for decades but became visible and reproducible after version 1.86.

[d] 1rB2R2/5p1k/5Ppp/6P1/1p1R3P/1b2K3/2r5/8 w - -

1. Rdd8 with the announcement of a mate in 8.

It's only a draw, fixed it.

haha, decades!

Just don't take decades to get this new release out, I can't wait to test those new parameters!

BTW: Is there any way you can make the WB2UCI file use the correct mate-score? If that's the problem, that is.

Even in new Prodeo sometimes if he is getting mated, he shows a weird matescore and the GUI adjudicates FOR him.

I'm not very sophisticated about this type of thing, maybe there's an easy fix, but I'm using Fritz GUI if that helps.

Hi Brendan,

You can include this line

Code: Select all

MateScore = 32767

in your ProDeo wb2uci.eng file, under [options].

CL

Oh Thanks Carl!

In which interface does this happen and do you have a position for me to check?

Thanks in advance.

[Rebel]

Adam, didn't you create Rebel Century personalities about 15 years ago?

Yes! But Ed took the credit for all my hard work

As I will continue

Seriously.........

[Rebel]

Yes! There are many more parameters that could be exposed now (version 0.195):

Code: Select all

    // Switch off weakening parameters

    search_skill = 10;
    nps_limit = 0;
    fl_weakening = false;
    elo = 2800;
    eval_blur = 0;
    book_depth = 256;

    // Opening book

    use_book = false;
    book_filter = 20;

    // Timing

    time_percentage = 100;

    // Piece values

    values[P_MID] = 86;   // was 100
    values[N_MID] = 294;  // was 325
    values[B_MID] = 317;  // was 335
    values[R_MID] = 435;  // was 500
    values[Q_MID] = 1110;

    values[P_END] = 153;  // was 100
    values[N_END] = 340;  // was 325
    values[B_END] = 382;  // was 335 // decrease
    values[R_END] = 641;  // was 500
    values[Q_END] = 1214; // was 1000

    // Tendency to keep own pieces

    keep_pc[P] = 0;
    keep_pc[N] = 0;
    keep_pc[B] = 0;
    keep_pc[R] = 0;
    keep_pc[Q] = 0;
    keep_pc[K] = 0;
    keep_pc[K + 1] = 0;

    // Material adjustments

    values[B_PAIR]  = 51;
    values[N_PAIR]  = -3;
    values[R_PAIR]  = -5;
    values[ELEPH]  = -65;
    values[A_EXC]  = 68; // exchange advantage additional bonus
    values[A_MIN] = 71; // additional bonus for minor piece advantage
    values[A_MAJ] = 71; // additional bonus for major piece advantage
    values[A_TWO] = 53; // additional bonus for two minors for a rook
    values[A_ALL] = 4; // additional bonus for advantage in both majors and minors
    values[N_CL]  = 7;
    values[R_OP]  = 3;

    // King tropism

    values[NTR_MG] = 3;
    values[NTR_EG] = 3;
    values[BTR_MG] = 2;
    values[BTR_EG] = 1;
    values[RTR_MG] = 2;
    values[RTR_EG] = 1;
    values[QTR_MG] = 2;
    values[QTR_EG] = 4;

    // Varia

    mat_weight = 100;
    pst_weight = 80;
    pst_style = 0;
    mob_style = 0;         // 1 is only marginally behind

#ifdef USE_RISKY_PARAMETER
    riskydepth = 0;
#endif
    draw_score = 0;
    shut_up = false;           // surpress displaing info currmove etc.

    // Asymmetric weights - the core of personality mechanism

    own_att_weight = 110;
    opp_att_weight = 100;
    own_mob_weight = 100;
    opp_mob_weight = 110;

    // Positional weights

    threats_weight = 100;
    tropism_weight = 20;
    forward_weight = 0;
    passers_weight = 100;
    outposts_weight = 100;
    lines_weight = 100;
    struct_weight = 100;
    shield_weight = 120;
    storm_weight = 100;

    // Pawn structure parameters

    values[DB_MID] = -12;
    values[DB_END] = -24;
    values[ISO_MG] = -10;
    values[ISO_EG] = -20;
    values[ISO_OF] = -10;
    values[BK_MID] = -8;
    values[BK_END] = -8;
    values[BK_OPE] = -8;

    // Knight parameters

    values[N_TRAP] = -150;
    values[N_BLOCK] = -20;

    // Bishop parameters

    values[B_FIANCH] = 17;  // general bonus for fianchettoed bishop
    values[B_KING] = 0;    // fianchettoed bishop near king
    values[B_BADF] = -20;  // enemy pawns hamper fianchettoed bishop
    values[B_TRAP_A2] = -150;
    values[B_TRAP_A3] = -50;
    values[B_BLOCK] = -33;
    values[B_BF_MG] = -10;
    values[B_BF_EG] = -20;
    values[B_WING] = 10;

    // Rook parameters

    values[RSR_MG] = 16;
    values[RSR_EG] = 32;
    values[RS2_MG] = 8;
    values[RS2_EG] = 16;
    values[ROF_MG] = 14;
    values[ROF_EG] = 14;
    values[RGH_MG] = 7;
    values[RGH_EG] = 7;
    values[RBH_MG] = 5;
    values[RBH_EG] = 5;
    values[ROQ_MG] = 5;
    values[ROQ_EG] = 5;
    values[R_BLOCK] = -50;

    // Queen parameters

    values[QSR_MG] = -89;
    values[QSR_EG] = 85;

    // King parameters

    values[K_NO_LUFT] = -15;
    values[K_CASTLE] = 10;

Wow...

Care to explain some?

Specifically:

values[ELEPH] = -65;

values[N_CL] = 7;
values[R_OP] = 3;

// King tropism
values[NTR_MG] = 3;
values[NTR_EG] = 3;
values[BTR_MG] = 2;
values[BTR_EG] = 1;
values[RTR_MG] = 2;
values[RTR_EG] = 1;
values[QTR_MG] = 2;
values[QTR_EG] = 4;

I assume it's about King Safety?

values[DB_MID] = -12;
values[DB_END] = -24;

Double Pawn ?

values[ISO_OF] = -10;

Isolated pawn on open file?

values[N_BLOCK] = -20;
values[B_BLOCK] = -33;
values[R_BLOCK] = -50;

Example(s)?

values[B_BF_MG] = -10;
values[B_BF_EG] = -20;

BF = bad fianchetto?

Guess not.

values[QSR_MG] = -89;
values[QSR_EG] = 85;

Like the rook values no idea what this ia about.

Please explain so that I can continue to take all the credit for your hard work.

[Adam Hair]

Wow...

Care to explain some?

Specifically:

values[ELEPH] = -65;

values[N_CL] = 7;
values[R_OP] = 3;

// King tropism
values[NTR_MG] = 3;
values[NTR_EG] = 3;
values[BTR_MG] = 2;
values[BTR_EG] = 1;
values[RTR_MG] = 2;
values[RTR_EG] = 1;
values[QTR_MG] = 2;
values[QTR_EG] = 4;

I assume it's about King Safety?

values[DB_MID] = -12;
values[DB_END] = -24;

Double Pawn ?

values[ISO_OF] = -10;

Isolated pawn on open file?

values[N_BLOCK] = -20;
values[B_BLOCK] = -33;
values[R_BLOCK] = -50;

Example(s)?

values[B_BF_MG] = -10;
values[B_BF_EG] = -20;

BF = bad fianchetto?

Guess not.

values[QSR_MG] = -89;
values[QSR_EG] = 85;

Like the rook values no idea what this ia about.

Please explain so that I can continue to take all the credit for your hard work.

Keep in mind that this is all Pawel's work. I just copied and pasted a section of Rodent's params.cpp file that contains the parameter values to show what values could be modified.

I have privately modified his code so that all of the parameters are exposed for tuning, but that would have been beyond my ability if Pawel had not already done this for multiple parameters. I simply copied his example.

Here are explanations for the parameters (as found in Rodent III 0.195 source code):

[ELEPH] = "elephantiasis correction" for queen, idea by H.G.Mueller (nb. rookVsQueen doesn't help)
[N_CL] = Knight likes closed positions
[R_OP] = Rook likes open positions
Tropism terms = tropism to enemy king (based on Gambit Fruit); seems to be related to king safety
[DB_MID] = doubled pawn
[ISO_OF] = isolated pawn
[N_BLOCK] = Knight blocking c pawn
[B_BLOCK] = bishop blocked on its initial square by own pawn
[R_BLOCK] = rook blocked by uncastled king
[B_BF_MG] = seems to refer to fianchettoed bishop blocked by own pawn i.e. b2 bishop blocked by c3 pawn
[QSR_MG] = queen on 7th rank attacking enemy pawns or cutting off enemy king

[PK]

Added some more comments in https://github.com/nescitus/Rodent_III/ ... params.cpp

Generally the grand plan is to move all the parameters to one big table, and then apply Texel tuning. Unfortunately, this is a lot of rather uninteresting typing, and I am rather busy nowadays, so it will take time. I already did some limited, manual experiments with tuning, and it helped with piece values.

BTW, Adam's values did not pass the test of about 3.000 games. But the games were rather interesting, so please expect "Hairy personality" in the near future.

[Rebel]

Added some more comments in https://github.com/nescitus/Rodent_III/ ... params.cpp

Generally the grand plan is to move all the parameters to one big table, and then apply Texel tuning. Unfortunately, this is a lot of rather uninteresting typing, and I am rather busy nowadays, so it will take time. I already did some limited, manual experiments with tuning, and it helped with piece values.

BTW, Adam's values did not pass the test of about 3.000 games. But the games were rather interesting, so please expect "Hairy personality" in the near future.

Code: Select all

values[P_END] = 153;  // was 100 
values[N_END] = 340;  // was 325 

I was already wondering about the above but now I see it originates from Adam's mind palace. I suggest a Hairy Scary personality.