June 7th, 2009 by d000hg
For this year at least. I never expected to get invited back a 2nd time to cover this year’s TCO, I shall try and keep the same attitude so that should I be lucky enough to be invited again it is just as pleasant a surprise! Either way, now’s a great time (well actually it’s about two days late but I’ve been driving across the USA all that time) to remember to thank TC for allowing me to come and witness another great showing of some of the smartest guys in their fields - not to mention providing my accommodation and food! Read the rest of this entry »
Posted in Uncategorized |
June 7th, 2009 by Rustyoldman
I brought with me, and carried around the whole time in my laptop bag, my Kindle and the latest copy of Make magazine. I also had the new Algorithms in a Nutshell book in my other luggage back in the hotel room. None of them ever came out of the luggage. That’s how active things were at the TCO. Even in the airport, in the plane or in the car going home the priority was gather information, prepare blog, sleep, eat. It wasn’t so intense that I am burned out now. After one nice long nap I am already ready for the next one.
Now that I am back home am really glad to not be looking at that tiny laptop monitor and thrilled to have a reliable internet connection and ergonomic chair/keyboard combination again. If I ever do this again maybe I will figure out a way to get five monitors into the war room (well at least plug one more into my laptop). One of the most significant problems I had this time was with the wireless internet connections supplied by the hotel. I just could not keep the arena running in the war room and thus was cut off from a line of communication with the off-site viewers that I had planned to maintain. By the way, I received no questions from off-site during the event so I guess I won’t bother to make a big deal of that again (assuming I do this again).
A lot more stuff happened than I managed to record in this blog. I will be thinking about how I might improve my reporting system in case I am ever asked to do this again. I guess something was wrong with my twitter setup as apparently my twits were not searchable and were only visible to my followers. Actually I am not even sure they were visible to anyone but me. Maybe someone out there can help me debug that problem.
It was really cold in the war room and that makes my hands hurt. They didn’t have eggs for breakfast until Thursday. The curtains in my no-smoking hotel room smelled of smoke. My back hurt too much. But those things were nothing. It was a great time start to finish and I would do it again at the drop of a hat.
Long live TopCoder and all the wonderful people that make it work.
Posted in Around the Arena |
June 7th, 2009 by Rustyoldman
The most intense time was probably when the algorithm semifinals were running. At this point there were 18 active monitors mounted on three islands about ten meters apart. Each monitor mirrored the monitor of one of the competitors. Many people (but fewer than usual) were gathered around the islands trying to watch as many of the monitors as possible. At the start I tried to guess who would be the slowest reader and read the problem statement from that monitor.
After the start it is a constant dash between monitors trying to keep track of what everyone is doing interrupted by trying to study any interesting code I see (while only being able to see a part of it at a time), looking back at the global scoreboard and looking at the competitors to judge their moods (there were a lot of chins in hands this time). The urge to try to code up a solution to some of these problems instead of watching is pretty strong (you know you’re a TopCoder when…).
People off-site have a disadvantage in that the problem statements are not visible in as timely manner and feel that their viewing experience is greatly diminished because of this. But the situation on-site is only partly better. On-site we can only see the part of the problem statement that is visible on some contestant’s monitor. Once they start working on the problem we spectators have no way to go back and examine the problem statement again. This year once everyone had opened a problem then printouts of the problem statement were passed out on-site and it was posted in the forum. In that respect the on-site and off-site experiences were not completely different, but it is true that being on-site is much better (even though it is not perfect from the spectator’s viewpoint).
My personal opinion is that insuring the integrity of the competition is far more important than providing an entertaining viewing experience on-site and off-site. Just because no one has ever been caught cheating at an on-site (as far as I know) that is not a valid reason to lessen the security. Security is as much for deterrence as it is for detection. I think the current on-site security should be improved. The only reasonably cheap way I can see to preserve (and enhance) both security and viewer experience is to isolate the on-site arena in a separate room from the on-site spectators. Only the active competitors, TopCoder staff and some media would be allowed in the arena room. The monitors mirroring the contestant’s monitors would be in the spectator/exhibitor/announcement/scoreboard/main room. Perhaps a web cam watching each contestant could also be added, for the benefit of the spectators. This way there would be no possible communication channel (short of concealed radio equipment on the contestant’s person) into the arena. This would also remove much of the distraction that the contestants currently face. We could do the traditional introductions in front of the crowd and then each contestant runs through the door into the arena room. That would be almost the same. As soon as the contest starts the problem statements could be made available on-line and in the spectator room.
Aside from preventing communication to contestants while they are in the arena, there is the issue of making the contestant’s codes visible or available during the contest coding phase, the challenge phase or before the official results are announced. The possibility is that off-site testers might determine and announce the results or critical parts of the results before the official ceremony. I don’t have any suggestions on what to do about that problem other than 1) not making the code available until the official announcement or 2) make the codes available and only have administrative controls making such rule violations grounds for suspension. I think examining the code by hand and announcing “I think I see a bug here” should be OK, so the line seems a bit fuzzy. Anyone else have any ideas? This seems like a good topic to discuss in the forums.
Posted in Alogrithm Competition, Around the Arena |
June 5th, 2009 by Rustyoldman
At one point in the blog entry creation process (when previewing a post) I see something like this:
which I believe this the only time I have ever seen my handle officially rendered in orange anywhere on TopCoder.com. Me as an admin, now there is a scary thought. I post this screen capture to strike fear into your hearts and inspire my “followers.”
[Apparently just clicking the title of the post will give this view too. I wonder if that will change color Tuesday when my access is terminated.]
Posted in Around the Arena |
June 5th, 2009 by Rustyoldman
I happily wasted many an hour at http://www.logicgamesonline.com/netwalk/ over the last few years. I think this is because the game is very easy once you figure out the basic logic then you can play quickly and almost fall asleep. It is not very hard to see that the optimal moves are in nearly every game, but I had never gotten the best score in any of my games. This was (as far as I can tell) because I was playing on a Mac and even though I had a two button mouse, the right click just didn’t ever get to the game properly. Left and right clicks rotate in opposite directions. I tried different browsers, different settings, USB Overdrive, etc. Nothing worked.
I mention this to jdmetz (who is logicgamesonline) and he relates another “issue” (not a bug really but it is a feature if prevented). People frequently click too many times while rotating tiles, messing up their scores, even though they knew exactly what orientation they wanted. I do that all the time too. He has already thought up a solution that would work for the overclicking and as a side effect it would eliminate the Mac bias.
I run off into the war room to do some blogging and in an hour or so, the very next time I see jdmetz around the arena, he gives me a URL for the new enhanced version of the game so I can beta test it. And of course it works perfectly.
Now that is the way to get things done. A real life Mod Dash.
Posted in Around the Arena, Mod Dash |
June 5th, 2009 by Rustyoldman
I guess I forgot to press the “Publish” button on this one. My excuse is I was writing this post at 2:00AM. Well, just imagine you are back in time before the finals.
The semifinal algorithm match was very interesting. I am pretty sure I could not solve any of the problems in the allotted time. But that didn’t keep me from standing in front of the monitors watching the competitors code and trying to understand it all long after my back started to hurt. There weren’t as many high ranking observers as there were previously, since all the on-site algorithm competitors were competing at the same time. In the past with separate semifinals there were always lots of them not competing and available to watch. Psyho, who is an algorithm target but is only competing in the marathon matches, was the highest ranking observer and therefore was the designated analyst.
As the competitors lined up for introduction before the start you could see a little bit of tension. Most of the competitors were a bit figgity, except Petr. Petr was yawning and sitting down. It looked like he might fall asleep if he had a chance.
Read the rest of this entry »
Posted in Alogrithm Competition |
June 4th, 2009 by Rustyoldman
I was completely shocked to to hear my handle announced as this year’s spirit award winner. I guess I didn’t even think I was eligible, not being a competitor. In retrospect I suppose it makes more sense that the award can go to any member not just to on-site contestants. But even if I knew I was eligible there was no reason to consider the possibility.
I am very proud to be a member of the TopCoder community. And I am truly honored that they consider me to be a positive contribution to the community. The TopCoder community contains many people with very outstanding talent and spirit. There are many other TopCoder members that I admire and that I think make a greater contribution to the community, so it is truly touching that the on-site voters disagreed with me on this one. I will use this award as a reminder to always try to be a positive contribution to whatever I am doing.
Thank you. Thank you very much indeed.
Read the rest of this entry »
Posted in Around the Arena |
June 4th, 2009 by Rustyoldman
Well, the awards are all final now which I am sure you have already seen elsewhere in the site. For the algorithm and marathon competitions the previous champions with seemingly overwhelming rating advantages were overthrown.
crazyb0y (also winner of the TCHS) becomes the youngest TCO algorithm winner and is no doubt starting a long successful career at TopCoder. The coders faced a quite difficult problem set this year where strategy in problem selection order played a very important role. The problems in the finals seem to be getting harder and harder. Years ago the winner was expected to get all three problems. Now, even with the slightly extended coding phase, it has been rare for anyone to get all three problems correct in the finals. All top three coders opened the problems in the order easy-hard-medium in an attempt to beat Petr who was expected to use the easy-medium-hard order. Well, it worked. Next year I would not be surprised to see people skipping the easy problem completely to try to beat the people that use easy-hard order.
KOTEHOK takes over as the marathon competition TCO champion. KOTEHOK has been doing marathons at TopCoder for a while and It was easy for me to pick him to finish in the top four, and he didn’t surprise me there. He chose a more analytical strategy than the purely heuristic of Psyho. Psyho had a solution more highly tuned to the example cases but KOTEHOK’s solution was more powerful in the general case.
Congratulations to the these new champions and to the champions of the other tracks of TCO’09.
Architecture: Standlove
Assembly: pulky
Design: saaxixx
Development: assistant
Specification: AleaActaEst
AOL Sensations: Accessible walking directions
Mod dash: PE
Studio: djnapier
As for the sprit award, I think they are still recounting ballots since something must have gone terribly wrong there.
Posted in Alogrithm Competition, Around the Arena |
June 4th, 2009 by jwlms
Congratulations to Dale Napier (djnapier), our TCO09 Studio Champion!
djnapier won 3rd place in Round One and 1st place in Round Two, which gave him just a little less than a one point lead over our second place winner and reigning TCO08 Champion, oninkxronda. Round One was worth 60% of the finalists’ score, while Round Two was worth 40%.
See all of the great designs and the final placements for each Studio finalists in the full post.
Read the rest of this entry »
Posted in Studio Design Competition |
June 4th, 2009 by Rustyoldman
Right before the announcements the tension is high, It looks as though it will be the year of the new guys in the algorithm competition. But anything could happen still. Here we go…
Posted in Around the Arena |
