Tuesday, November 3, 2009

ACM ECNA Regional 2009

The final standings are here. As usual the contest report doesn't mention any technical problems. I thought I would mention the ones I heard of here.

When the contest started the McMaster site had not yet handed out problems. So after PC^2 was started it was stopped again, then the contest was started over 8 minutes later. At least at my site we started and were never told to stop, and didn't check PC^2. I believe that the other sites except for McMaster did the same thing. Essentially McMaster was 8 minutes late to a 5:08 long competition. Its not fair, but it shouldn't change the results of the world finals teams.

The time left on PC^2 was wrong for about the last hour of the contest. It showed < 2 minutes left for about the last hour. We were done at that point so we were unaffected. This is a huge problem for teams still working because it makes it much harder to manage time. The other CMU team was printing every few minutes because on the printout was the current time.

The last issue may or may not be an issue. At this point I'm willing to give the benefit of the doubt to the judges. Most of the teams who got B incorrect (Waterloo, CMU Dragons, Purdue) weren't sure why they got it wrong. There are rumblings that the test data was wrong, or that there was floating point error. I haven't throughly checked our 4 wrong answers to that problem, but at this point I believe all 4 of them were incorrect. After the floating point debacle of last year I think the judges probably checked the problem very throughly before finalizing the results.

When the test data is released for B we'll be able to go back and figure out why teams were getting wrong answers. The problem was hard to code, and hard to debug so it is plausible that teams just had subtle bugs in their code.

The final results were CMU Dragons in 4th, Waterloo in 3rd, University of Michigan 2nd, and my team, CMU Tartans in first. My team had plenty of good luck on our side. We weren't affected any of the issues and potentially we got B right despite rounding errors in the test data. Either way we are headed to Harbin China in February.

No comments: