MATLAB Programming Contest BlogMay 14th, 2007Big Push Winner, 1000 Character ChallengeMarkus Buehren won the Big Sunday Push with two entries that made big improvements, one of which dropped the score more than any other player’s combined efforts for the day. Markus first made a showing in the Blockbuster contest, scoring two midcontest wins. Jan Langer, nathan q, Jin, and David Jones also contributed significant improvements. Coming up next is the 1000 Character Challenge. This mid-contest award goes to the highest scoring entry submitted on Monday before 9PM EDT. To make it easier to find these entries in the results, we’ve added a listing of the leaders to the statistics. 4 Responses to “Big Push Winner, 1000 Character Challenge”Leave a Reply |
|
|||
|
These postings are the author's and don't necessarily represent the opinions of The MathWorks. |
For those of you interested, I’ve posted another heavily commented version of the leading code as of mid-day today, Monday. The code is “Guided Tour Part 2″ (id: 40979).
Unfortunately, with the number of subfunctions and recursive calls the code has become much more difficult to understand than in my previous attempt at this. Hopefully you’ll find this of value / interest.
The code will become even worse. Since Nick Howe’s Early Bird submission, there has IMHO been no real algorithmic improvement on the results. The only improvements (maybe I not up-to-date right now) were the tweaking of certain numbers and factors and some huge timing tweaks (congratulations Markus :-)) resulting in subsequent readjustments of the numbers to invest the timing improvement on better results.
Hi!
As you have read, I won the Big Sunday Push. I just want to tell you the interesting story how lucky I were to win this prize.
My contribution markus9 introduced a significant speed improvement to the part of the code origininating from Yi Cao and his twilight winning contribution. Actually I had already put this improvement in an earlier submission called markus8 of 2007-05-12 07:58:42. If I had not made an algorithmic error when combining some different code versions, I had probably won the Saturday Early Morning Special (at 2007-05-12 08:00:00) with this contribution.
Well, as markus8 finished somewhere in no man’s land, of course no one took notice of the improvement. So I had the chance to use it in another attempt. markus9 - now without this f… error - made an improvement about 8 points. However, looking into the statistics later that day, I was disappointed that Jan Langer had already reached a cumulative improvement of more than 10 by numerous tweaking contributions (when does this guy sleep?!?).
But, to my great surprise, my speed improvement disappeared from the leading code again over the day! It was Nathan Q who found another, even larger way to improve the speed. He took the lead with his submission bigh_L8 at 06:12 h after my submisson markus9 of 05:17 h (no, I don’t work all through the night, I just live in another time zone). As the other contestants now started working on the code of the new leader, it happened that I had a third chance to use the same improvement again. This time the impact was even larger (about 18 points), lifting me up to win the Big Sunday Push.
I’d like to mention the contribution of Nathan Q here again, who introduced a fast update to the vector of possible moves instead of completely recomputing it. I guess in the next contests, I may not hope that Nathan gives me such assistance again in winning a midcontest prize. :-)
I wish you lots of fun during the remaining time of this contest!
Markus
Markus, this is a great story. With the volume of entries, it’s tough for observers to see this sort of action. We’d love to find ways to encourage more participants to post their experience, as you have here.