March Madness bracket advice, adjusted for 2015

It’s that time of year again, and while you can go to just about any media outlet for March Madness advice, I’m fairly confident you won’t get most of the stuff that I’m going to write about here. I think that’s a good thing?

As preliminary thoughts, feel free to check out my two posts from last year:

Value and March Madness

What are the actual odds of someone picking a perfect bracket?

Okay, here are some thoughts and general strategies.

1- Your first round choices depend on your scoring system. And maybe even your second round picks, too.

Most pools can generally be separated into one of two categories – those with upset points or those without upset points.

Strangely enough, the vast majority of people entering picks in pools with upset points pick the same way as they would in pools without upset points. This is silly. In upset pools, for example, correctly backing a No. 13 seed to beat a No. 4 seed could be worth 5-10 times as much as the alternative. This makes taking Mercer over Duke, or Georgia State over Baylor, turn into reasonable choices.

If your pool does not have upset points, however, and is scored in standard 1-2-4-8-16-32 form, its usually not worth picking many upsets. In these formats, the only thing that really matters is picking the champion and finalist. So, it is not worth trying to be hip and taking Georgia State. And its not worth worrying about picking the right No. 12 seed to beat a No. 5 seed. Just take the teams that are favored to win by sportsbooks (lines here) and don’t fall behind.

2- Your champion depends on the size of your pool

Just like most events, there are usually no prizes in March Madness for finishing in 10th place. Your only goal is to finish first. So how does that impact your choice of teams?

Quite simply, it is a better idea to take riskier options as the size of your pool increases.

Over on Grantland, Ed Feng provides a useful example with the 2010 tournament, showing how Duke was undervalued, Kansas was overvalued, and the backers of Duke had much better chances of winning their pools as a result. Here’s his graph:

Screen Shot 2015-03-17 at 8.58.35 PM

The probability of winning in 2010 after backing Duke was roughly five times higher than after backing Kansas, despite the two starting the tournament with similar probabilities.

So, estimate the number of people in your pool, and vary your aggressiveness as a result.

3- You are not trying to pick games correctly, but trying to get more points than your competitors. 

An important March Madness strategy lies in identifying the teams you think your opponent is going to pick, and then picking the opposite teams.

I stole this from Jordan Sperber’s blog, which has some great insights on under and overvalued teams. Here’s a Table with 2015 Final 4 odds, using the expected values from Ken Pomeroy’s website and those chosen by the public.

Screen Shot 2015-03-17 at 8.55.39 PM

So, here are some Final 4 teams with decent 2015 probabilities that your opponents are not picking: Villanova, Gonzaga, Utah, Arizona

And here are some Final 4 teams with decent 2015 probabilities that your opponents are picking way too often: Duke, Louisville, Wisconsin, North Carolina

So, what does this all mean?

It looks like about 1 in 4 sheets on has Kentucky and Duke in the final game. My top advice is that you are not one of these people in 2015.

Of course, this is not because I don’t think UK and Duke can reach the finals; instead, it is because even if they do reach the finals, you still probably won’t win your pool. There’s just too much competition in picking those two teams.

In larger pools, Arizona, Villanova, and Virginia, are all being backed by small fractions of the public (no more than 6%, according to ESPN), but each have between a 9% and a 13% chance of winning the title, according to many of the sites that run through bracket simulations. One of those teams appears much more likely to give you a chance at capturing first place.


1 Comment

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s