self deepening: an improved implementation of IID

[Gerd Isenberg]

Mike,
Bob was direct, straightforward and correct but probably not so perceptively to you with your frustration and hurted vanity. This is also quite understandable after so many years of reacting on hostilities from a lot of sides and people. You need somehow a thick skin

Thanks for the best wishes - after latest results of my program i don't feel so comfortable, but i hope for some nice matches and to steal one of the big irons a draw

Cheers,
Gerd

[Michael Sherwin]

deleted

[Volker Pittlik]

Gentlemen!

As I just wrote to someone else in a personal message it may be good to remember that this place is run to get information and to have fun.

If one disagrees with someone else or feels hurt by him it may be a good idea not to miss the standards of manners he use to respect if he talks to his colleagues at work. Disagreements and misunderstandings are normal. Name-calling does not help in that situations.

Regards

Volker

[Oliver Uwira]

It is way harder to indicate what the power is from Crafty than to indicate why Fruit plays so strong. Fruit is a very tiny engine. Crafty a very big monster whose code is difficult to read for most.

Hello Vincent,

when I started chess programming 2 years ago, I had a look into Crafty's source because I had no idea how to implement a static exchange evaluator with bitboards. I understood right away how to do it and the source was not ugly but very well commented and easy to read (that is, I didn't just copy and paste the SEE but could adapt the method).

I could not say that I would understand Crafty as a whole but that wouldn't be a surprise for me either because it encapsulates some 20 years of experience.

I don't know what's going on personally between you and Bob but as I often read about Crafty's alleged unreadability I felt that I could just throw my two cents' worth in.

Viele Gr??e,
Oliver

[Gerd Isenberg]

Ok, i understand. I was not aware of RomiChess. Maybe we have a different understanding of vanity - i think your reaction is a classical symphtom of vanity and it is ok - and it is good to talk about and to sensitize us with your issues. I think i am a vain a most here are to some degree. The world is inequitable - if you have no reputation nobody listens to you - but i think you are on a good way to change that.

Don't stop asking questions. Again there is a difference to study or even compile and debug a complete program or to react on relevant pseudocode snippets with concrete questions...

There are so many open sources around - and not everybody is really interested in it, until it becomes more knowledgable due to some spectacular results.

Gerd

[H.G.Muller]

Still, Michael, what you ask seems unrealistic to expect: that you post your source code on a web-site somewhere, and then hold it against people on a completely different website that they haven't solved any problems with it... While that other web-site does not even mention or specify the problem!

People on this forum are generally very helpful if you ask clearly formulated questions here. But just an "I tried it and it didn't work", without even posting what you tried... Bob is right, understanding someone elses code that you've never seen before is a hell of a lot larger job than 20 minutes. It would only be 20 minutes for someone that already intimately knows the code, like yourself.

So a much more fruitful approach would be post your search routine here, or better yet, a pseudo-code digest of your search routine, show how you tried changing it, and invite us to have a look at it. I am sure that would produce quick results. IID is conceptually very easy to implement. You have a depth argument to your seach(), and rather than using that argument directly in the body you replace it everywhere by an iter_depth that runs upto depth, as the loop index of a loop that you put around the bulk of your current search routine.

The only details to be added is how you would use what you learned in the previous iteration to improve the next one, a situation that is logically equivalent to when you get a hash hit of unsufficient depth.