New UnseenSource for Reborn Engine.

[markchessman]

Hi ,I have not fixed an error in remainsworthy, but I did base the new engine RebornMainsworthy on its code and fixed the error. the souce code has never before been released.

reborn & reMainsworthy, are completely differrent engines from Mainsworthy

its available at Leo site in the Mainsworthy engine page
http://wbec-ridderkerk.nl/

the source code, has some easy to follow steps, it defines all move of each piece from all squares without a real-time move generator. it still has a generator but its basicaly just dropping moves from the list that are unavailable.

It dosnt check move legality.

it has very easy enpassant code to follow, eg if you move a pawn diagonaly but the square is empty then your useing enpassant, so delete piece.

it has an unusual book that you can edit in a text editor(it shows the position, and reply), and it uses a basic algo.

Mark Ainsworth

[markchessman]

I should of mentioned, its a winboard engine with only go move new & quit commands, also you cant change sides mid game.

[markchessman]

would somebody like to give the engine an ELO estimate? I havnt got a setup to run a tourny yet.

I would of thought there would be interest on a new engine on here No Replys so far.The engine is slightly based on an older engine, but it now has fixed errors and books, aswell as the source being released for the first time .

[mar]

Hi Mark, I remember about 5 versions of *mainsworthys running in last ChessWar's promo:
supermainsworthy, remainsworthy, tournamentmainsworthy, mainsworthy, son of mainsworthy...
So welcome rebornmainsworthy.

[mar]

Just played two games against rebornmainsworthy.
One was a draw where I totally messed up but rmw went for a 3-fold rep instead of a win.
In the second game, I won. rmw blundered in this position:

[d]2k4r/1pp4p/6p1/3rbpB1/p1B5/6P1/P1P2P1P/1R3RK1 b - -

Playing Rd7...
Are you sure that all bugs are gone? A 2-ply search (with quiescence) should see this.

Martin

[hgm]

I would of thought there would be interest on a new engine on here No Replys so far.The engine is slightly based on an older engine, but it now has fixed errors and books, aswell as the source being released for the first time .

In general there is, but quality has something to do with it too. Your previous engines were sort of 'asymptotically weak', all much weaker than a simple, slow, fixed-depth alpha-beta searcher with nothing but piece-square tables for evaluation. This can be 'forgiven' once, for a beginner, but you have been building sort of a reputation as 'repeat offender', cranking out one super-weak engine after another. You should not be surprised this dampens people's enthousiasm a bit to look at them.

Of course this new engine could be 2000+ Elo, I don't know, because I haven't tried it. But you can try it easily as well. The fact that you ask other people to determine its Elo doesn't raise hopes very much. If you are not willing to even do such a very easy thing yourself, how much effort can we expect that you put into debugging or tuning your engine? The fact that it still has the very rudimentary WB interface is another negative quality indicator.

So why don't you run a few test matches yourself, and tell us whether you new engine can beat Iota, Toledo nanoChess, TSCP or Fairy-Max?

Sorry for the harsh words. They are not intended to discourage you, but on the contrary to stimulate you into doing a better job.

[ilari]

would somebody like to give the engine an ELO estimate? I havnt got a setup to run a tourny yet.

I would of thought there would be interest on a new engine on here No Replys so far.The engine is slightly based on an older engine, but it now has fixed errors and books, aswell as the source being released for the first time .

Maybe there's not much interest because it's very frustrating to test engines that don't implement even the fundamental parts of Xboard or UCI properly. I tried to pit your engine against Fruit 2.1, and this is what happened:

Code: Select all

[Event "?"]
[Site "?"]
[Date "2012.08.03"]
[Round "1"]
[White "Fruit-2.1"]
[Black "RebornMainsworthy"]
[Result "1-0"]
[PlyCount "15"]
[Termination "time forfeit"]
[TimeControl "40/60"]

1. Nf3 {+0.10/12 4.1s} d5 {2.0s} 2. d4 {+0.10/11 1.7s} Nc6 {5.9s}
3. Nc3 {+0.12/10 1.5s} Bg4 {6.1s} 4. e3 {+0.32/10 4.2s} e6 {11s}
5. h3 {+0.29/10 1.5s} Bh5 {5.5s} 6. Bb5 {+0.32/9 1.2s} Kd7 {9.7s}
7. Ne5+ {+4.06/10 1.0s} Kc8 {19s} 8. Qxh5 {+4.10/10 0.99s, Black loses on time}
1-0

[Event "?"]
[Site "?"]
[Date "2012.08.03"]
[Round "2"]
[White "RebornMainsworthy"]
[Black "Fruit-2.1"]
[Result "0-1"]
[PlyCount "18"]
[TimeControl "40/60"]

1. e4 d5 {-0.09/11 1.9s} 2. Bb5+ {4.9s} c6 {+0.53/11 1.4s} 3. Ba4 {9.0s}
a5 {+0.91/9 1.8s} 4. exd5 {3.0s} b5 {+1.83/11 1.7s} 5. Bb3 {7.4s}
a4 {+1.94/10 1.5s} 6. dxc6 {6.5s} axb3 {+2.90/8 0.74s} 7. c7 {6.2s}
Qxc7 {+2.93/10 1.4s} 8. Qf3 {6.5s} Rxa2 {+8.99/10 1.5s} 9. Rxa2 {14s}
bxa2 {+12.59/9 0.68s, White's connection stalls} 0-1

I looked at the source code and I can now understand why you're having such a hard time fixing the bugs. It's just extremely difficult to maintain huge amounts of copy-paste code - I certainly wouldn't be able to do it. So my suggestion is to forget about hacking together chess engines as quickly as possible, and focus on writing clearer, more modular, reusable code; you'll end up making better programs.

[Sven]

would somebody like to give the engine an ELO estimate? I havnt got a setup to run a tourny yet.

I would of thought there would be interest on a new engine on here No Replys so far.The engine is slightly based on an older engine, but it now has fixed errors and books, aswell as the source being released for the first time .

I played a quick blindfold game with black against it, using only the console version without any chess board display. I sacrificed a knight by intent in the early opening to test whether it can win against a weak opponent. Later on the engine lost the piece again, then I stopped.

Code: Select all

[White "ReBornMainsworthy124"]
[Black "Sven"]
[Result "*"] 

1.e4 e5 2.Bc4 Bc5 3.d3 Nf6 4.Bg5 Bxf2+ 5.Kxf2 Nxe4 6.dxe4 Qxg5 7.Nh3 Qe7
8.Nc3 O-O 9.Qd5 Nc6 10.a3 d6 11.Ng1 Be6 12.Qb5 Bxc4 13.Qxc4 Nd4 14.Nd5 Qd8
15.Nxc7 Rac8 16.Kf1 Rxc7 17.Qd3*

So I believe there is still a bit of work to be done before estimating an ELO ... Currently I would think it is well below 1400 but I may be wrong.

Sven

[markchessman]

thanks guys at least Im not alone out here

I shall check your data and info and see what I can do !

[ZirconiumX]

Ummm... Have you ever thought about something called 'compatibility'???

I tried to compile it with

Code: Select all

g++ -o rbm *.cpp | tee errors

I now have a file called errors which is 412KB, and contains 4907 errors.

Matthew:out