A study on engines

Okay, if I want to include that many I'll need some help collecting games. If anyone would like to help out, give me a message. I'm keeping track of the games I need to play using an online tournament creator. If I add anyone in, I will need to regularly know every single game you have added -- to keep the statistics accurate, ya know. The plan right now is to eventually increase the maximum search depth to 20. For some chess programs that won't be a problem, but others will take a long, long time to reach it... and since this experiment lists each program search depth as a different engine, there would be 20 competitors multiplied by whatever number of engines I have.

To sum it all up, I have two options if I want to add more: add them selectively, and exclude some, or add a bunch and find someone else to help out.

P.S. I currently don't need the actual games played, just the results... but if we correspond by email, then you can, if you wish, email PGN files.

P.P.S. I should probably mention how to achieve what I am doing. To make this happen, download the GUI, Arena (for free) and connect each program to it. Then, load two Arenas at the same time on one computer, and make one engine play on each Arena. Then, press Control L, select the mode "Fixed Search Depth", and press the spacebar to make White make a move. Then switch to the other Arena, and input the move just made (make sure the "Edit" option is off, otherwise the engines will just sit there). Repeat, until a full game has occurred. 

My first search depth tournament (double elimination, depths 1-5) is almost  finished, and the results are . . . odd. You know the IPON rating list? The one that says Houdini's number 1 in the world, and that Komodo and Critter are fighting for 2nd place, and so on? Forget that. The engines pretty much created a whole new hierarchy here, and once they're done I'll post it all.

A provisional rating list has been drawn up from the results. I won't bother with the numbers, as that will change a lot over time, but here it is:

1. Rybka 2.2

2. Critter 1.4

3. Hermann 2.6 (?!)

4. Toga II 1.4

5. Fruit 2.3.1

6. Stockfish 2.2

7. Houdini 1.4a (?!?!??!)

8. Komodo 3

9. Gull 1.2

10. Deep Shredder 12

11. Anmon 5.75 (Well, at least I had predicted this one wouldn't make the top of the list . . .)

12. Dragon 4.6

13. Ruffian 1.0.5

14. Nejmet 3.07

15. SOS 5.1 for Arena

I'm quite amused by this. We'll see if this remains true at deeper searches.

Nah. Basements have really bad Internet.

Interesting post why u stopped?? Post more

Open-Chess.org forum and Chess2U.com forums are dedicated to chess engine comparisons, tournaments and rankings.  If you are really into watching engines play or even if you want to try and code your own, that's the place to look.  Apparently different engines have different strengths.  Additionally, the strength of an engine may be related to the opening book database and endgame tables being used (if any) and their respective reliance on those aids.  Obviously you want to compare engines on the same computer.

Personally, I use Fritz 13 to study the PGN after a game and only after personal analysis to check for missed ideas.

Fritz 13 is stronger than houdini 1.5a?

I have no idea which is stronger.  Both are far stronger than I am, and I can run either one within the Fritz UCI.  It really doesn't matter to me since I do not play computer chess, both are rated around 3000, and either one can help me analyze mistakes made in re-playing lost games.

Actually, I'm not sure what the Fritz 13 engine really is.  You seldom if ever see it used in computer vs. computer comparisons and tournaments, and since it comes from the Chessbase people, I suspect it is a fairly strong engine in keeping with their Houdini reputation.  I have Houdini 1.5a x64, Crafty 23.01, Fritz 5.32, Rybka 2.2 x64, and Stockfish 2.3.1 all installed.  I just don't think it matters much at my level if my engine is 3100 or 3150.  I only use them after the fact.

Another consideration is what opening book you are using and what tablebase for endings is being used.  Some engines work better within different time limits or plys than others, while others may be "better" at open ended (unlimited) analysis.  If Fritz gives me one "best" move and Houdini gives me another, they are both good moves, and the difference is both infintesimal and beyond my ability to discern the difference in tactical value.  So for me, at my pathetic level, it doesn't really matter.

hmm i agree but i only asked is fritz stronger than houdini not that which matter to you most or whats your pschology over engines anyway all the best

Houdini’s a perfectly decent engine but synchs poorly with most interfaces.  In the analysis lines you get repetitive best moves and spacing problems and it doesn’t support certain cool features, like the dial evaluator you get in Deep Fritz.  Make Houdini your third choice for an engine.  Also there’s no support to speak off – just Bob Houdart, over in Europe and not at home.