Winboard Forum

Hello,

if i remember correctly, it was Piotr who pointed out a division by zero
in a previous version of Romi.

As a result the engine crashed when playing using increments.
Looks like the problem still exists, at least it does here for me.

I hope that Michael still reads here and can fix this.

Btw: Thx for the new version, looks pretty strong

Peter

Sorry everyone,
I had to reformat my hard drive when I installed XP 64 bit and I lost my password to the Winboard Forum, so I have been gone awhile. About the divide by zero error -- I am a beginning self taught programmer and I failed to realize that the % operator would cause the same error. Versions 2f and 2g are the same except that the error was fixed in 2g. As far as strenth goes I've had too little time for testing, However. in a test tourney by Leo, 2f finished ahead of Ceng and Butcher while taking all four points from those two engines. I am stoping work on the prototype version, so, I can concentrate on getting RomiChess 1.0 up and running. If anyone wishes to make improvements to the prototype version then it will be posted at Leo's site with the new and improved source, so, that everyone can benifit.

Thanks
Mike

Hello Michael,

Romi can even blitz very well now,
works fine here.

thx and hope you make progress with the 1.0 as you wish.

Peter

Hi Peter,

Thank you for taking an interest in Romi. I think that there are other people that like Romi and think that Romi is interesting. But, to date no one has told me WHY. If you have the time could you please give me some insight into why you like RomiChess, so I can be sure to keep those same things in RomiChess 1.0. There are much stronger engins with far more features, therefore it must have something to do with style. If so, then the style between 1a and 2g is a little different. Which do you prefer?
Any info that you can give me will be greatly apriciated.

Best Wishes
Mike

Michael Sherwin wrote: Thank you for taking an interest in Romi. I think that there are other people that like Romi and think that Romi is interesting. But, to date no one has told me WHY. If you have the time could you please give me some insight into why you like RomiChess, so I can be sure to keep those same things in RomiChess 1.0. There are much stronger engins with far more features, therefore it must have something to do with style. If so, then the style between 1a and 2g is a little different. Which do you prefer?
Any info that you can give me will be greatly apriciated.
Best Wishes
Mike

Hello Michael,

I am sorry to confess that I dont have too exact an impression about the differences
between 1a and 2g and can hardly say anything useful about them
in the moment. Maybe later, but I cant promise, it would take some time.

But generally: I was very impressed by the relatively high strength of the
first version, it was certainly a great newcomer, quite fast, with the most
important knowledge implemented, quite mature for a newcomer.
The fact that you made the source available to the public did increase the
positive picture quite a bit , well and indeed I found the readme interesting, too,
because you gave some insight into the process of "The Making Of", to say so.

So my interest is only a fan's more general interest and is not very subtle
in this case, sorry about this probably disappointing fact.

(I failed to see a real progress from the first version to the last, but this is
based on few games played with short time controls, and there was
the issue about time management and some different versions,
so I really am not sure about this, could be totally wrong
and have to check the material more thoroughly.)

Peter
(oh, there was the new fun pack today,
thx for that, but I did not yet have a look at it.
So more about this later, maybe)

Thanks Peter.

That is more feedback in one place than I've gotten in total. The second version was because of the learning ability that I added not because of a strength increase. In a match with crafty 19.19 from the original position and with no opening books and with learning for both programs, if I remember correctly, Romi took the lead at about game 50 and after game 107 untill I stoped the match at about game 250, Romi won every single game.

Mike

For a new engine, RomiChess is very strong and very interesting.

If you want help to improve it, I would suggest that the Winboard forum or CCC is a good place to get it.

You will find that just making your code public is only a small help. However, if you have something specific (for instance "IID is slowing your engine down instead of speeding it up") then address that one single problem. Show what you have done and ask what should be done to improve it. If you throw out all the things you want fixed at once, I don't think you will find relevant speedup.

I think that for an early stage in development, it is one of the more promising chess engines.