College Basketball 2

[Arbor Gold]

Trisk, I'll be honest - I don't give a fuck about my bracket.

And I probably care a little too much about mine. I spent a large part of Saturday running down the hall from the TV to the computer yelling "hang on, let me check my bracket!". The bracket makes all this so much more fun for me. Annoying to others, perhaps, but fun for me.

[Mack Kilimaro]

How often do people pick a perfect bracket (perfect bracket / all brackets)? It is clearly going to be more often than random odds, because the teams are weighted with seeds. The odds would change if you took as a maxim that 16 seeds would not defeat 1 seeds, for example. 2-15 games would be weighted less strongly.

I can tell you this. Several years ago to promote their bracket picking site (and they may still do this) Yahoo offered to pay $1 million to anybody who perfectly picked the whole tournament. No one was paid out that year and to the best of my knowledge no one ever has been paid out.

[S John]

And I probably care a little too much about mine. I spent a large part of Saturday running down the hall from the TV to the computer yelling "hang on, let me check my bracket!". The bracket makes all this so much more fun for me. Annoying to others, perhaps, but fun for me.

Definitely adds a lot. It takes me from only really caring about how far WV can make it to having a stake in every single game. I like college football pick 'em for this reason also.

[chionophile]

So...math people...

How many bracket combinations are there?

How often do people pick a perfect bracket (perfect bracket / all brackets)? It is clearly going to be more often than random odds, because the teams are weighted with seeds. The odds would change if you took as a maxim that 16 seeds would not defeat 1 seeds, for example. 2-15 games would be weighted less strongly.

My guess is that based on data regarding performance of various seeds over the years, you can vastly improve your odds over the random sample.

I'm asking so that I know how many brackets I actually have to fill out to win my pool.

I don't have the time or statistical ability to figure this out, but according to Peter Tiernen of ESPN, the odds are 1 in 9,223,372,036,854,775,808 (2 to the 63rd power. Which, in retrospect, makes sense) of perfectly picking a correct bracket. Mind you, that's picking every game completely at random, the odds are probably slightly better if you know something about the teams you're picking. Tiernen has apparently done a lot of research as to what attributes (like average point total, margin of victory, that kind of thing) an underdog usually has, although that doesn't really help a lot with picking the current bracket.

[S John]

I don't have the time or statistical ability to figure this out, but according to Peter Tiernen of ESPN, the odds are 1 in 9,223,372,036,854,775,808 (2 to the 63rd power. Which, in retrospect, makes sense) of perfectly picking a correct bracket. Mind you, that's picking every game completely at random, the odds are probably slightly better if you know something about the teams you're picking. Tiernen has apparently done a lot of research as to what attributes (like average point total, margin of victory, that kind of thing) an underdog usually has, although that doesn't really help a lot with picking the current bracket.

Nor does it predict possibilities like injuries, a monumental display of heart, or a loss of temper resulting in a technical foul at a really bad time. I'm sure someone will do it one day though. I know a guy at work who did one of his brackets using a magic 8 ball. The person who gets them all right will probably have pulled a stunt like that. Though this time he's in dead last out of 50 or so in the office pool.

[Gold Storm]

Even knowing a lot about college basketball does not help that much. Most of the experts had Pitt, UCONN, GTown and USC doing some damage. Picking at random would almost be as good I wonder how hard it would be to get every pick wrong.

[BLU-RAY]

Even knowing a lot about college basketball does not help that much. Most of the experts had Pitt, UCONN, GTown and USC doing some damage. Picking at random would almost be as good I wonder how hard it would be to get every pick wrong.

Picking at random, the odds of getting every pick wrong are 1/(2^32); after you get all the first rounders wrong, you're obviously wrong the rest of the way.

[Old Zog]

Picking at random would almost be as good [...]

I had a friend who did that one year - he flipped a coin 63 times to fill out his bracket. He did about as well as you'd expect. That is, significantly worse than everyone else. If you try this (with a fair coin) you expect 16 right in the first round, 4 right in the second round, 1 right in the third round, and essentially nothing beyond that.

I remember reading years ago (the year that Arizona won it all) that someone using ESPN's site got something like 60/63 correct. I'd be shocked if anyone's ever done better.