Comments on R-Chart: Google AI Challenge: Scores/Rank by Language

The tall spike represents the "starter packag...

2010-12-17T14:57:45.243-08:00

The tall spike represents the "starter package" bot. This is an example bot that could be used by any contestant as a starting point upon which to improve. Many people created an account then submitted the starter package, then never came back.

Behind the scenes one of the contest admins produced a very similar chart which plotted every bot's rank on the x-axis along with its draw rate (the percentage of its games that ended up being draws) on the y-axis. One could easily pick out clusters of identical or very similar bots on this chart.

Not only was there a large cluster of starter packages, but there were also a few other small clusters. Our best guess was that these clusters represented the most common strategies that occurred to people as they say down to start writing their entries.

It's because a shitfuck of the submissions are...

2010-12-08T10:25:15.307-08:00

It's because a shitfuck of the submissions are the example bot.

Corrected Brian - thanks. As good a guess as any ...

2010-12-08T09:57:07.896-08:00

Corrected Brian - thanks.

As good a guess as any Diomedian. There were a number of universities listed - so some entrants may have been students. And in many such competitions, there are a lot of people that are interested initially, but other responsibilities crowd out work throughout the competition.

My guess (not knowing anything about this competit...

2010-12-08T08:30:15.097-08:00

My guess (not knowing anything about this competition) is that the really tall peak is due to programs whose output, for whatever reason, is not much better or worse than random. It peaks right around the middle and is bizarrely high. It may be explained by the incompetence of the general coding public at a very difficult task like AI.

s/your won vis/your own vis/

2010-12-08T07:12:39.963-08:00

s/your won vis/your own vis/