Coders Bracket: Write successful tourney code, win $2000 (algorithms and coding and stuff)
https://www.codersbracket.com/code_bracket
Write your own algorithm for picking your bracket (or hard coding a Michigan victory). You can pull in all manner of stats and rankings, plus add your own special sauce.
Has default examples like picking by seed or at random, or increasingly based on seed in later rounds.
Nothing as silly as picking by SAT scores, but a chance to test your hunches on what is really going to matter.
Based in JS. I figured there was a decent enough programming crowd here to pass it along. Plus, there are actually basic JS lessons there too, if you are both into winning at sports *and* learning marketable skills.
EDIT:
Forgot to mention the prizes, the key one being $2000 hard cash for the valiant victor.
Code on.
then yes. (Or 2^67 if you include the Play-In Round™.)
I got Harvard.
This is actually really, really awesome. Aaaand there goes my productivity at work. But boss, I'm coding.
this was the key to my code:
Hey Coach Schiano - that's pretty funny, don't we all wish it were that easy.
This is David from CodersBracket (actually a project of our students here at Fullstack Academy http://fullstackacademy.com). Feel free to reach out to me [email protected] if you have any questions and would love to see some more pro-Michigan brackets :)
yes indeed. I THOUGHT YOUR SITE GUARANTEED SUCCESS.
you should include other data like kenpom, etc.
cool site btw.
If anyone has an idea for an algorithm that they'd like to see coded, let me know. I'll try to get some done when I have some free time.
I'd like you to code an algorithm that codes algorithms for me in its spare time.
That's called "training your replacement" where I come from.
Soon, the machines will take over.
If there's a way to qantify "resume" (best wins/worst losses), I'm thinking something like maybe kenpom, minus the two best wins & worst losses, weighting more for stats in the last 10 games (and even more for the last 5), with an increasing bias towards highers seeds after each round.
...and also, Michigan always wins.
A "Plinko" version that is only 100% for 1 seeds vs 16s in the round of 64, but each step down in seed increases the influence of randomness?
Someone should reverse engineer the code after the tourney to see what it would have taken to get a perfect bracket (other than the one that I already entered - going to be nice being a billionaire, will be able to afford water at football games for the whole family)
March 18th, 2014 at 10:47 PM ^
March 19th, 2014 at 12:30 AM ^
a bracket generator back in the pre-internet days when getting info on this stuff was tough. It used a combo of the Sagarin rankings (which were about the only up-to-date rankings you could get via newspaper) and historical seed information that I generated personally by digging through all the tournament results since the field went to 64 teams in 1984. I used it to enter office pool contests and eventually the online tournament pools for about 10 years. In all those years I don't think I beat it with my own personal entry even once. I only wish I had the KenPom stuff to play with back then.
JavaScript + Michigan hoops...
Worlds...is...colliding.
March 19th, 2014 at 11:10 AM ^
Im going to put kenpom ratings into an array, use an array lookup function and go with the higher one.
March 19th, 2014 at 11:52 PM ^
It shouldn't be an array, but a property on the team object...
over Kansas, but I should probably correct that for injury.