MATLAB Programming Contest BlogApril 11th, 2006Mid-Contest AnalysisLucio has posted a Mid-Contest Analysis which, along with the recent comments in the newsgroups , will help keep those creative juices flowing. Meanwhile, the Tuesday Leap challenge continues, with the biggest single-entry improvements so far today coming from Markus and David Jones. The Statistics now include a listing of the Leap leaders. 
7 Responses to “Mid-Contest Analysis”Leave a Reply |
|
|||
|
These postings are the author's and don't necessarily represent the opinions of The MathWorks. |
Due to the spammers the queue is so full at the moment that it takes more than one hour until one can see the result of an own submission. When submitting this comment, I have to read a spam prevention number from an image and type it correctly into a text field. Is it possible to include such spam prevention into the submission page quickly, I mean before the last hours of this contest? This change would not affect the rules in any way.
As a side effect, looking through the contest pages is getting less and less fun because the browser needs longer and longer to load the lists of entries.
I doubt this spam prevention technique, called a CAPTCHA, would slow down the inflow very much. We were kicking around the idea of only scoring one entry per contestant on each pass through the queue. That is, we’d only score one entry from each author until everyone waiting has had one of their entry scored. The drawbacks are that it is a little tough to explain and it would encourage contestants to submit under different names.
With respect to the long page, is there a shorter list you’d like to see? The most recently scored 20 entries or something?
I agree with Markus that the queue issue is making it very hard to participate. Regarding the Tuesday Leap, the announcement is a little unclear as to if it’s the best entry that RUNS by midnight or the best entry submitted to the queue by midnight. If the CAPTCHA idea won’t work (which I think it would help) another possibility is to limit the overall size of the queue to something like 10 entries. A new entry can’t be submitted until one finishes. That would alleviate the problem of people spamming the queue and then coming back hours later to see the results and spamming again. While the queue is likely to be always full, this way anyone at any given time should have an equal chance of submitting an entry and seeing it complete in a reasonable amount of time.
I clarified the language of the Tuesday Leap announcement. The code just needs to be submitted before midnight. Thanks.
The number of entries that some contestants are submitting has reached a new high in this contest. In particular, two players have submitted 1000 entries each. We do need to come up with some way to make it possible for less prolific participants to get more immediate feedback, so please keep the ideas coming.
The sample test suite is similar in character to the one we use to score the entries, so hopefully that compensates for a long queue somewhat. Contestants should be able to work in algorithmic improvements offline. It’s only for tuning (overfitting) the constants to the actual test suite that they need feedback from the queue.
18 Move CheckerBoardCompletion
[20 5 2
20 6 0
20 5 0
20 5 0
19 6 0
5 3 0
4 2 0
5 10 0
4 10 0
12 7 0
14 6 0
16 6 0
18 6 0
9 2 0
12 4 0
12 4 0
16 4 0
20 4 0]
16 Move CheckerBoardCompletion
19 3 3
19 3 0
19 8 3
19 8 0
19 8 0
19 3 0
20 5 0
7 4 0
9 5 0
9 5 0
12 5 0
12 5 0
15 6 0
17 6 0
19 6 0
18 6 0
It is now 2:41AM EDT (I am not a nightowl, just living in another timezone) and there are over fifty entries in the queue that were submitted before midnight. This means it takes around three and a half to four hours for an entry to get through the queue.