UPDATED:
Chris's 2009 NCAA Bracket
Obama's 2009 NCAA Bracket
I just love that the bracket has its own presidential seal.
(And let me add, by the way, that he has Syracuse beating Temple, whereas I have both teams losing before that game even gets played - we'll see who wins this epic showdown of knowhow (of which the president apparently has quite a bit) vs. basic mathematics!)
Subscribe to:
Post Comments (Atom)
Posts
The president and I are both off to rough starts (one for the first three; my tournament champion was losing its first game with five minutes to go).
ReplyDeleteI'll put real-time updating links to our brackets in the article.
three out of three for me...
ReplyDeleteNo shit? Well done.
ReplyDeleteI am two for three so far. Damn Butler.
ReplyDeleteTwo complaints: 1. My family is stupidly doing our bracket competition through FOX Sports, perhaps the most poorly designed and buggy website in existence; and, 2. The scoring over at FOX Sports is ridiculous (correct pick in round 1 is 1 point, in round 2 is 2, etc.).
I think the only proper scoring of NCAA brackets is as follows: for each correct pick you get the winning team's seed number multiplied by the round of the victory. This way you are rewarded for picking upsets. Also, no matter what seed you pick you get more points as your chosen team progresses further in the tournament. For example, picking UNC to win the first round nets you one point if they win, but picking UNC to win the sixth (and final) round nets you six points if they win. Thus, if you correctly pick UCONN to win it all you would from that team only get 1+2+3+4+5+6=21 points because they are a favorite; correctly picking Chattanooga to win it all would net you 16+32+48+64+80+96=336 points, rewarding you for the risk of choosing a 16 seed (of which none have ever made it past the first round.
)
ReplyDeleteI agree with your theory but your practice is a little shaky. Let's say we get to the Sweet 16 and all regions have 1-2-3-4 teams left.
ReplyDeleteThere's no way the 1 seed wins 80% of the games against the 4 seed - in reality, it's probably between 50% and 60% at best - so I think you're overincentiving betting on the underdog.
I'm just nitpicking, I don't mean to criticize the idea. This issue is easily correctly by using Vegas odds or something like that.
That's a good point. Perhaps we can compete in a pool run in that manner next year. In the end the thrust of my point was that FOX Sports sucks just as much as most of the rest of the FOX family (including FOX Family).
ReplyDeleteI again would like to emphasize that my bracket is purely objective based on one simple (and greatly flawed) calculation. I only mean to show that as simple as flawed as it is, it'll still do better than "experts" in the long run.
ReplyDeleteI feel pretty good about this report from Nate. If you look through the charts, particularly the last one, it seems I'm not the only one who thinks that Memphis and Gonzaga (my final two teams - neither of them #1s) are underrated.
(Confounding variable check - their methodologies probably put heavy weights on the same statistics I used)
Man, I had a big analysis ready to go ripping into Nate's analysis of Obama's picks and whether he favors teams from swing states. Unfortunately, as I re-read the article to prepare my commentary, I realized we pretty much did the same thing.
ReplyDeleteLong story short, I used two variables. Variable 1 (think of as the "Going out on a Limb value") is the seed difference, which I made negative when Obama picked the favorite and positive when he picked the underdog. So picking a #16 over a #1 gets you a 15 on the Crazy-ness scale, picking a #8 over a #9 gets you a -1.
Variable 2 is a "Swingy-ness" variable, which is Log[Abs(UnderdogObama-UnderdogMcCain) / Abs(FavoriteObama-FavoriteMcCain)]. (The takeaway is that higher values go to matchups when the underdog is from more of a swing state.)
***
So I did all my (not at all) fancy regressions, and this is where I got upset. My p-value is 0.088. In statistics, this is the equivalent of "getting warmer" when what you want is to get down below 0.05, and really the key is down below 0.01. So how did Nate come to find that the president showed such a bias toward swing states? With his p-value of 0.052.
Now, to be fair to Nate this analysis was a great idea. And you can't really run it if the answer to your question is "no". (Just imagine listening to the 10 o'clock news, when just before the commercial the anchor says "And coming up, is a gigantic asteroid on a collision course with Earth this year? Keep it tuned to find out," only to come back after commercial and say "Nope. We just think it's a good question to ask from time to time." So I understand why he's taking that angle, but the fact is the evidence just isn't there, especially with so many other variables in the way.
This comment has been removed by the author.
ReplyDeleteWay to go Mizzou. Thank you for destroying (what was left of) my bracket.
ReplyDeleteHow do you know your bracket is busted? When the final four hasn't started yet, and you're guaranteed to go 0-3.
ReplyDelete