lucasart wrote:Updated ratings after 223/400 games. Now Cheng is seriously competing with Crafty for the 9th place
Thanks Lucas but I believe the score will still settle, cheng is nowhere near the strength of Crafty, I'm pretty sure of that
Martin
Sorry to contradict you, but the more games I play the more Cheng's elo settles around 2650. Perhaps you're overestimating Crafty ? Crafty is only 50 elo weaker then Fruit 2.1. This is all using 1 thread, and at 1'+1" time control. Of course if you use a quad core, and allow Crafty to use 4 threads SMP while the others use 1 thread, it would gain some elo, but probably not more than 100 elo. So Crafty is not the unbeatable monster you imagine
lucasart wrote:Updated ratings after 223/400 games. Now Cheng is seriously competing with Crafty for the 9th place
Thanks Lucas but I believe the score will still settle, cheng is nowhere near the strength of Crafty, I'm pretty sure of that
Martin
Sorry to contradict you, but the more games I play the more Cheng's elo settles around 2650. Perhaps you're overestimating Crafty ? Crafty is only 50 elo weaker then Fruit 2.1. This is all using 1 thread, and at 1'+1" time control. Of course if you use a quad core, and allow Crafty to use 4 threads SMP while the others use 1 thread, it would gain some elo, but probably not more than 100 elo. So Crafty is not the unbeatable monster you imagine
PS: Cheng crashed twice already, once against GNU Chess, and once against Crafty. I removed the games and will terminate these matches later.
Well the fact is that cheng is better at blitz. Perhaps that might be the reason.
But I thought Crafty is already better than Fruit. Strange.
The crashes are a disturbing. 2 crashes in 300 games is a lot. I have to admit that I had some very rare crashes since 1.05 (about 1 crash in 1500 games), but i was unable to track them. So I guess I will have to look into that... Thanks again Lucas. 2650 would be very nice.
Thank you for sharing your work with the open source community
My only rant against Cheng is that the source code is still quite messy and hard to read. But I'm sure it's all work in progress and improving day by day!
Thank you for sharing your work with the open source community
My only rant against Cheng is that the source code is still quite messy and hard to read. But I'm sure it's all work in progress and improving day by day!
You're welcome
Yes the code is very messy, I prevent cloning that way and I get more LCCRL points as a bonus
No really it's getting harder to read even for myself.
But as you said it's work in progress, i'm certainly not a guy who prints his wonderful code and puts it on the wall (=i'm not obsessed with coding style nor design patterns, I hate the word "Factory" and m_ prefixes).
Anyway cleaning it up seems inevitable, sooner or later.
Thank you for sharing your work with the open source community
My only rant against Cheng is that the source code is still quite messy and hard to read. But I'm sure it's all work in progress and improving day by day!
You're welcome
Yes the code is very messy, I prevent cloning that way and I get more LCCRL points as a bonus
No really it's getting harder to read even for myself.
But as you said it's work in progress, i'm certainly not a guy who prints his wonderful code and puts it on the wall (=i'm not obsessed with coding style nor design patterns, I hate the word "Factory" and m_ prefixes).
Anyway cleaning it up seems inevitable, sooner or later.
Martin
As a bonus, I've already found bugs while cleaning up my code. At some point, I just didn't understand parts of code, so I read them again and again, in order to rewrite them in a cleaner way... Only to realize the code was actually buggy
lucasart wrote:
As a bonus, I've already found bugs while cleaning up my code. At some point, I just didn't understand parts of code, so I read them again and again, in order to rewrite them in a cleaner way... Only to realize the code was actually buggy
Sure I hope to find bugs that way. There's a lot of stuff i would like to get rid of, especially legacy board.cpp and search.h coming from cheng2 (never released). But I've also experienced the opposite in the past, "fixing" bugs where I didn't fully understand my own code and thought there's something wrong with it (usually too late and tired). So I think that I'll get rid of some and produce some new. Happens all the time