March Madness Math: UK All the Way?

So your "patience" is rewarded. It's Bracket Time! And if you plan to not pay attention, well, you're in the minority. This, from ESPN:

"The American Gaming Association estimates that 40 million Americans will fill out more than 70 million brackets and wager approximately $9 billion on the NCAA tournament, according to research released Thursday.

"The $9 billion is more than double what the AGA estimated was bet on the Super Bowl ($3.9 billion) in the United States. The total number of brackets expected to be filled out will be greater than the number of ballots cast (nearly 66 million) for President Barack Obama in the 2012 election."

That's a lot of attention (and a lot of coin). And it explains why CBS Corporation (NYSE:CBS) and friends are forking over $10.8 billion over 14 years to cover the entirety of March Madness.

So, how are we going to win some money this year? I heard a guy on CNBC call Kentucky a "negative alpha" trade, meaning it is a bit overvalued. Here's my first suggestion: Keep an open mind about Kentucky. I feel odd saying this about an overwhelming favorite, but they're not actually overpriced. Here's a look at the betting odds on winning the title, via 5Dimes:

So, if you bet $100 on Kentucky and they win, you earn $139. That implies it has a 41.84% chance to cut down the nets. And guess what? That's about right. KenPom.com says 33.8%, FiveThirtyEight.com says 41%, and the simulation I ran on PredictionMachine.com said 45%.

They are over-picked, however. At last check on ESPN, 49% of entrants picked Kentucky. That's high, but not outrageously high. And it does highlight the basic simplification of the poll this year. Your chances of winning your pool, or finishing in the money if there are multiple payouts, all comes down to your play on Kentucky. And the numbers suggest you have modestly more to gain by not picking them.

But the specifics matter. The smaller the pool, the more sense it makes to go "safe" and take Kentucky. It also makes more sense to take them in a format that has multiple prizes. If it's a large pool and/or one grand prize winner, it probably makes more sense to pick an undervalued team with a realistic chance. Why's that?

My favorite way of looking at it is like the "Birthday Paradox". Two people in the room aren't likely to have the same birthday. But each added person walking in makes it exponentially more likely there's a birthday "match." By the 23rd person, it's about 50%. Instead of birthdays, think of them as having filled out brackets. Odds are any individual person will have a lousy bracket. If you're going head-to-head, pick all favorites. But get 20 people together and it's likely one of them did it quite well. Picking all the favorites won't work nearly as well.

Anyway, that highlights the strategy I used this year. I went with teams that are under-picked. I defined that as when the percentage of ESPN entries picking them to advance to a certain round was significantly below the math odds of a team making that round. But that only goes so far. If a team has a 25% chance to reach Round 4, for example, but only 10% of entrants had them there, that implies I should take them, right?

Well, they're still a bit of an underdog. I may take them anyway, but I can't go with every profile like that, lest I have Valparaiso playing Buffalo in the Final 4. So, I used it more for potential upset picks here and there. I like it better when there's a team like Texas that's an 11 seed and a modestly better team than their potential opponent in Round 2, No. 6 seed Butler.

Fast-forward and I ended up with mostly chalk, or near chalk, in the Final Four. I have Kentucky, Arizona, Villanova and Gonzaga, with 'nova beating Kentucky in the finals. I doubt 'nova actually wins, it's just "value." But hey, once upon a time, Villanova took out an even bigger giant.

Disclaimer: Mr. Warner's opinions expressed above do not necessarily represent the views of Schaeffer's Investment Research.