Al Zimmermann's Programming Contests

[spurious_ai]

Glad to see the site back up Rémi.

I am stuck at this point, can't solve n=8. Just one point short. There are only small fractions left on my n=9 to n=14 scores.

Have you had time to work on the contest Rémi or are you done? Luke? Can anyone beat Hugo and Hermann?

[Rémi Coulom]

Glad to see the site back up Rémi.

Sorry for the downtime. A big maintenance operation at univ-lille3.

I am stuck at this point, can't solve n=8. Just one point short. There are only small fractions left on my n=9 to n=14 scores.

Have you had time to work on the contest Rémi or are you done? Luke? Can anyone beat Hugo and Hermann?

I have stopped working on the contest for some time. I am focusing on Go now. I will play two games against a professional player in Japan in September, and then travel to China for the Computer Olympiad. I may go back to that contest in October.

I don't remember very well how n=8 was. Maybe the search was lucky. My scores for n=9 to 14 are far from the best.

Rémi

[luke_g]

I am stuck at this point, can't solve n=8. Just one point short. There are only small fractions left on my n=9 to n=14 scores.

Have you had time to work on the contest Rémi or are you done? Luke? Can anyone beat Hugo and Hermann?

They seem to be doing very well, I hope they will share their algorithms after the contest. I have only looked at small N so far, up to 7 (although I've submitted for all N to get a sense of the records). I am very busy with my day job (also I was on vacation last week and did no programming), so I have limited time to develop my program(s) and limited chances of beating H&H. I haven't implemented many of my ideas yet. However, I am working on a new algorithm, which I hope will take care of smaller N.

I wonder how many CPUs other people use for this. I only have 2 spare cores to devote to it. Having a cluster could be a huge advantage!

EDIT: And best of luck to Remi in go! It's been great to see the developments in computer go the past year.

[spurious_ai]

With less than 30 days to go, the number of active participants has dropped dramatically.

The leaders tend to hold back results until the last days.

It appears the maximum result up to N=11 has been found.

Remi, I hope you can participate in the contest again.

I saw the results on the ICGA website, but no commentary. I would be interested in your impressions of the tournament.

[Rémi Coulom]

Remi, I hope you can participate in the contest again.

It is not likely. I am too busy with teaching right now.

I saw the results on the ICGA website, but no commentary. I would be interested in your impressions of the tournament.

Crazy Stone won the match against Kaori Aoba in Japan, with a 8-stone handicap:
http://remi.coulom.free.fr/FIT2008/
This is the best performance of a program against a professional player. MoGo failed at 7 stones twice against two other professional players.

I did not participate in the tournament of the Computer Olympiad, again because I am busy with other things. I was in Beijing for the conference, though, where I presented my WHR paper (http://remi.coulom.free.fr/WHR/). I also worked on updating the ICGA web site. Here are the links to the Go tournament results, for those who don't know:
http://www.grappa.univ-lille3.fr/icga/t ... php?id=181
http://www.grappa.univ-lille3.fr/icga/t ... php?id=180

As usual, the tournament showed big strength improvements over the previous edition. But since the Olympiad in Amsterdam, no really big idea was found. Strength improvement is mainly due to parallelization, big hardware, and implementation of more domain-specific knowledge, all of which is not very exciting. Go programmers will have to find another way in order to be able to challenge strong humans at 19x19 Go.

All in all, I enjoyed my visit to China very much. We had some time for a little bit of tourism, and it is always fun to meet our fellow computer geeks in real life.

There is no definite decision for the edition of next year. There is a small probability that it may take place in the US.

Rémi

[spurious_ai]

100 days is a long time for a contest!

Remi, you finished 25th out of 306. Seeing as you only worked on it for a few days, that is amazing.

I finished 6th.

They plan to reset the votes ( hopefully soon ), and then the voting will start with the results hidden so people don't start working on the next contest before it starts.

[Rémi Coulom]

100 days is a long time for a contest!

Remi, you finished 25th out of 306. Seeing as you only worked on it for a few days, that is amazing.

Thanks. This is not so amazing, really. My algorithm was a simple genetic algorithm. I added constraints to make sure that exactly one number is odd, and the others are multiples of 6, 30, 210...

I finished 6th.

Congratulations !

Rémi