Ok, here is another property: all the records for N<27 are multiple of 3. Does that mean that the tops should alway be multiple of 3 ? JC [Non-text portions of...
162
Marcin Mucha
marcin_mucha
Jan 4, 2010 12:50 am
I don't think so, but I don't have a proof here. However, if you consider smaller values of C (the number of colors) the property you suggest breaks at some...
163
Dmitry.Kamenetsky@...
dimkadimon
Jan 7, 2010 7:02 am
Thank you James. I really like this problem! Just wondering, were you at all inspired by Gasarch's $289 puzzle? ...
164
James J Youlton Jr
youjaes
Jan 7, 2010 12:58 pm
The spark for this puzzle came from Paul Cleary using squares. I morphed it quite a bit in to it's current state over a period of about a week. This is the...
165
Al Zimmermann
azimmermann
Jan 9, 2010 5:16 pm
It just occurred to me. If I set up a Contest of Contests at AZsPCS, then James and I could collaborate and set up a Contest of Contest of Contests. Al, who...
166
Marcin Mucha
marcin_mucha
Jan 9, 2010 5:22 pm
I think two Contests of Contests are good enough :) Yours could have a different name though. In many competitive activities there are more than one major ...
167
Al Zimmermann
azimmermann
Jan 10, 2010 8:10 pm
Interestingly, Pyramid Power is like the AZsPCs Squares contest from 2003, with the following differences: Pyramid Power uses a triangular board. Squares used...
168
Mark
mmammel8
Jan 11, 2010 2:48 pm
Hmmm, this is a hard one, the best I can do using my standard simulated annealing approach is to leave 8 rectangles with the same color corners. I guess I'd...
169
Jean-Charles Meyrignac
jcmeyrignac
Jan 11, 2010 3:02 pm
The current record is 1 left rectangle: http://www.cs.umd.edu/~gasarch/BLOGPAPERS/17x17almost.txt It's probably possible to do an exhaustive search, because we...
170
Kevin Burfitt
zaph_torps
Jan 11, 2010 8:04 pm
Actually, I don't think that is '1 left rectangle' - for each of 0,1,2,3 I can see multiple squares created (thats just looking manually) - it is probably...
171
Dmitry.Kamenetsky@...
dimkadimon
Jan 11, 2010 11:24 pm
It seems that the contest of contests is biased towards people who compete a lot, rather than those who do well. The top is likely to consist of people who...
172
Dmitry.Kamenetsky@...
dimkadimon
Jan 11, 2010 11:29 pm
You are correct. 17x17almost.txt is a 4-coloring of 288 locations, the last location (+) cannot be colored without giving forbidden rectangles. I think the...
173
James J Youlton Jr
youjaes
Jan 11, 2010 11:58 pm
Let's say for example that someone participates in one contest and scores 100%. Should they be the CoC leader forever? James * From:...
174
Marcin Mucha
marcin_mucha
Jan 11, 2010 11:59 pm
I believe topcoder ranking is really bad for "recreational" contests like this. I take part in TC marathons and have some friends who do as well. The thing is...
175
Dmitry.Kamenetsky@...
dimkadimon
Jan 12, 2010 12:43 am
Not quite, because the ratings continue to change after every contest. Consider the following example. Contest1 PersonA 100 PersonB 90 PersonC 80 Contest2 ...
176
Dmitry.Kamenetsky@...
dimkadimon
Jan 12, 2010 12:54 am
I had a feeling that you are a TCer :) I agree completely with you. TC rating is not entirely suitable for these contests. In these contests, one might start...
177
Dmitry.Kamenetsky@...
dimkadimon
Jan 12, 2010 12:54 am
Oops. Contest2 should be PersonB 95 PersonC 85 PersonD 85...
178
Marcin Mucha
marcin_mucha
Jan 12, 2010 1:01 am
Sure, nothing wrong with that. All I was trying to say is that I would hate it if participating in a contest could decrease your rating... ... -- Dwell not on...
179
Dmitry.Kamenetsky@...
dimkadimon
Jan 12, 2010 1:38 am
I agree. The only solution I see: before the contest starts (before seeing the problem) each contestant chooses whether he wants to get rated for that contest....
180
Al Zimmermann
azimmermann
Jan 12, 2010 5:15 am
... Unfortunately, then any unscrupulous person could boost his rating in the CoC by creating bogus accounts and entering a single poor entry for each account...
181
Dmitry.Kamenetsky@...
dimkadimon
Jan 12, 2010 5:22 am
... But wouldn't that boost everyone's rating not just his? Creating bogus accounts should be illegal anyway and such people should be banned from the...
182
Al Zimmermann
azimmermann
Jan 12, 2010 5:33 am
It would only boost the ratings of those who'd entered the same contests he had. Also, I think (I haven't done the math) that if the 1st-ranked person in a...
183
Jean-Charles Meyrignac
jcmeyrignac
Jan 12, 2010 8:58 am
I think the current system is fair. I participated in 2 contests and got the top score for both of them, but I don't feel that I should be the first in the...
184
Dmitry.Kamenetsky@...
dimkadimon
Jan 12, 2010 12:30 pm
... If the 1st ranked person does this cheating in a single contest where he came 1st then his CoC score does not change, while the CoC score of the 2nd person...
185
Paul Cleary
paul2cleary
Jan 12, 2010 3:09 pm
There is another approach to this that would solve the cheating aspect and could also solve the rankings taking into account how many contests you enter. ...
186
Paul Cleary
paul2cleary
Jan 12, 2010 3:13 pm
Sorry about the look of the list, it was in a table when i posted it including boxes. ________________________________ From: Paul Cleary...
187
Marcin Mucha
marcin_mucha
Jan 12, 2010 4:05 pm
They sure do look similar :) So what you're saying is that it is still profitable to submit dummy entries to those contests that you're actually going to...
188
Paul Cleary
paul2cleary
Jan 12, 2010 6:12 pm
There are those who have taken a drop in the rankings and those that have gone up, It is biased towards those that have participated in more contests,...
189
Dmitry.Kamenetsky@...
dimkadimon
Jan 13, 2010 2:20 am
That seems to work. The only thing I don't like is that your CoC score is an average over all contest scores. This is bad, because one could compete once and...
190
Paul Cleary
paul2cleary
Jan 13, 2010 10:19 am
I placed a 'Once Only Guy' into the mix in the 3rd contest with a score of 100 and ran the numbers, his place in the CoC was Ranked 33rd place, his Normalised...
