tag:blogger.com,1999:blog-1107147718367558732.post5418921320120339296..comments2009-09-05T16:02:47.307-07:00Comments on This Number Crunching Life: Data driven march madness predictionsDanny Tarlowhttp://www.blogger.com/profile/14670021337844708633noreply@blogger.comBlogger16125tag:blogger.com,1999:blog-1107147718367558732.post-15160603945064065962009-09-05T16:01:00.880-07:002009-09-05T16:01:00.880-07:00twangbio (and anybody else) -- I'm happy to se...twangbio (and anybody else) -- I'm happy to send you the code, but please leave a way for me to contact you. Better yet, send me email: dtarlow at cs toronto eduDanny Tarlowhttp://www.blogger.com/profile/14670021337844708633noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-32308164875778465312009-09-04T00:08:02.731-07:002009-09-04T00:08:02.731-07:00Can i get the code please? Interesting to look int...Can i get the code please? Interesting to look into it.<br />Thanks.twangbiohttp://www.blogger.com/profile/15901988820276028282noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-28280508175922934792009-03-27T11:06:00.000-07:002009-03-27T11:06:00.000-07:00It's the revenge of the nerds: Time to make some m...It's the revenge of the nerds: Time to make some money!Will Dwinnellhttp://www.blogger.com/profile/03379859054257561952noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-50558946303047393242009-03-19T10:43:00.000-07:002009-03-19T10:43:00.000-07:00Peter: my email is dtarlow at cs.toronto.edu.Peter: my email is dtarlow at cs.toronto.edu.Danny Tarlowhttp://www.blogger.com/profile/14670021337844708633noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-2680972874069446462009-03-19T06:47:00.000-07:002009-03-19T06:47:00.000-07:00Danny,<br><br>Care to share your email id, so I ca...Danny,<BR/><BR/>Care to share your email id, so I can contact you to get the code? I would love to take a look.<BR/><BR/>Thanks,<BR/>PeterPeter Prakashhttp://www.blogger.com/profile/01329534807597642854noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-15980592954846950872009-03-18T18:14:00.000-07:002009-03-18T18:14:00.000-07:00Hey Jeremy,<br><br>I added links to the learned 1D...Hey Jeremy,<BR/><BR/>I added links to the learned 1D offense and defense ratings for every Division 1 team. Let me know if there is something systematic that you think is off.<BR/><BR/>I will be the first to admit that I know very little about college basketball, so you should definitely be skeptical. I'm not doing anything more than interpreting data about scores over the course of the season Danny Tarlowhttp://www.blogger.com/profile/14670021337844708633noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-70195622870037055732009-03-18T17:54:00.000-07:002009-03-18T17:54:00.000-07:00I'd like to see the ratings you gave all of the te...I'd like to see the ratings you gave all of the teams. The ratings seem pretty arbitrary. I don't know if you know anything about college basketball, so why would I trust your predictions?Jeremyhttp://www.blogger.com/profile/03683212751489697843noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-75548954509524551222009-03-18T17:23:00.000-07:002009-03-18T17:23:00.000-07:00Can I get the code as well please? I've been tryi...Can I get the code as well please? I've been trying to do some learning algorithms like this as well.<BR/><BR/>Thanks.<BR/>MattGBOGHhttp://www.blogger.com/profile/10520674496641183603noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-91001789755642061142009-03-18T17:02:00.000-07:002009-03-18T17:02:00.000-07:00Hey Nick,<br><br>Thanks, but don't trust these too...Hey Nick,<BR/><BR/>Thanks, but don't trust these too much! It's just a fun little project I did to kill some time on an airplane -- I'm not about to put any money behind my predictions =P<BR/><BR/>And sure, send me an email and I'll send you the code: dtarlow@csDanny Tarlowhttp://www.blogger.com/profile/14670021337844708633noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-40530028692327070582009-03-18T16:42:00.000-07:002009-03-18T16:42:00.000-07:00Hi Zack,<br><br>I agree. It would be great if Nat...Hi Zack,<BR/><BR/>I agree. It would be great if Nate would take a crack at the bracket -- I wonder if there is anything he learned from his baseball days that would be applicable here.<BR/><BR/>In regards to your comment about strength of schedule, you're right that it's not explicitly included in the model.<BR/><BR/>However, it is implicitly taken into account.<BR/><BR/>For example, suppose Danny Tarlowhttp://www.blogger.com/profile/14670021337844708633noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-25852238092819136802009-03-18T16:32:00.000-07:002009-03-18T16:32:00.000-07:00Also...I don't have much of a background in Machin...Also...I don't have much of a background in Machine Learning, but I'm part of DGP here at UofT and I'd be excited to see the code and better understand how you did this. Really cool!nickhttp://www.blogger.com/profile/04843288610325263724noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-16381100997106693422009-03-18T16:07:00.000-07:002009-03-18T16:07:00.000-07:00I saw your post on 538.com, and I'm also hoping th...I saw your post on 538.com, and I'm also hoping that Mr. Silver gives us his bracket.<BR/><BR/>Regarding your model, it doesn't appear that you did anything to account for strength of schedule. It seems that your analysis would favor teams that built their data set against inferior competition. Indeed, it seems like this is the case, with teams from weaker conferences like the Pac10 an Zackhttp://www.blogger.com/profile/11929150389451374181noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-57274652833724215102009-03-18T15:58:00.000-07:002009-03-18T15:58:00.000-07:00This is awesome man! I just filled out my bracket...This is awesome man! I just filled out my bracket yesterday, and coincidentally I have similar results as you. Suddenly, I'm feeling a lot more confident about my picks! haha!nickhttp://www.blogger.com/profile/04843288610325263724noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-36452953472503286272009-03-18T15:12:00.000-07:002009-03-18T15:12:00.000-07:00Evan -- thanks for that reference. It looks like ...Evan -- thanks for that reference. It looks like the paper describing their method isn't up. I'll try to dig in and find it tonight and maybe post some comments about the differences in our methods.<BR/><BR/>In the meantime, for those who just want the predictions, here is the <A HREF="http://www2.isye.gatech.edu/people/faculty/Joel_Sokol/profspicks/profspicks09.htm" REL="nofollow">direct link Danny Tarlowhttp://www.blogger.com/profile/14670021337844708633noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-73929991161448965922009-03-18T14:46:00.000-07:002009-03-18T14:46:00.000-07:00A Professor at Georgia Tech has a model for predic...A Professor at Georgia Tech has a model for predicting the tournament, check out:<BR/><BR/>http://www2.isye.gatech.edu/people/faculty/Joel_Sokol/<BR/><BR/>It seems like you get pretty similar results, but there are a few differences.<BR/><BR/>May the best model win!Evanhttp://www.blogger.com/profile/16583349586517978225noreply@blogger.comtag:blogger.com,1999:blog-1107147718367558732.post-39648721041932118572009-03-18T14:13:00.000-07:002009-03-18T14:13:00.000-07:00haha, this is awesome!haha, this is awesome!Rolf Steierhttp://www.blogger.com/profile/04521713185343114657noreply@blogger.com