Evaluating (and Improving) the MoneyBrackets System

Harvard v Michigan State

Happy morning after, everyone. If you’re like me, you’ve got a terrible case of basketball fatigue, some residual anger from all the potential money you’ve lost, and a bracket that makes you feel the same emotions men feel when reading “Dear John” letters. March Madness is wonderful and awful, and all I can really remember is that the first two days were the best I’ve ever seen, and the CBS commercials seem less annoying this year.

Now is the perfect time to tamp down our emotions and put on our analytical hats. Last week, I patented a system called MoneyBrackets that used research and market inefficiencies to help you gain an advantage in NCAA pools over your friends and enemies. I outlined 10 strategies, so let’s revisit them now to see how the system fared, and how we can improve it.

1. Don’t Pick Any No. 8/9 Seeds to Upset a No. 1 Seed

I wrote: This database from the Washington Post … shows that 1-seeds are 101-15 against 8- or 9-seeds. That’s 29 years and 15 upsets, or about one every two years … It is extremely unlikely we’ll have more than one 8- or 9-seed making it into the Sweet 16, and even if one does, you have a poor chance of picking the right one.

The truth: Well, one happened. Wichita State fell to Kentucky in what was, for me, one of the most depressing results not involving Duke in years. That means there have now been 16 victories for 8- or 9-seeds over 1-seeds in 30 years, or almost exactly one every two years. This advice, I think, is solid; you’re better off not taking the risk. As for Wichita? It felt like that game was in hand toward the end, but a series of awful bounces and great, uncharacteristic shooting by Kentucky destroyed the dream. That’s the cruelty of March.

2. Don’t Take Louisville Over Wichita State in the Sweet 16

I wrote: This one is tough, because like a lot of people, I think Louisville might be the best team in the country … [but] when you consider that everyone will be picking Louisville in that game, it’s like free money to go with the undefeated 1-seed.

The truth: Didn’t work out. Sorry. But at the risk of not accepting blame for a bad pick, I still think the nature of the advice is solid. Wichita is a good team that had never lost, and when you actually stand to gain points on everyone by picking a 1-seed over a 4-seed, well, that’s a market inefficiency. This one happened to fail, but the other very similar scenario I advocated, Virginia over Michigan State, will play out Thursday, and I still think Virginia is a great pick against the hordes (more than 50 percent, I’d guess) who took the 4-seed that lost seven of its last 10 regular-season games.

3. Don’t Get Fancy With Your Champion

I wrote: There are five teams you should be picking to win the whole thing: Florida, Virginia, Wichita State, Louisville, or Arizona.

The truth: Still stands. One thing I failed to note, but that I executed by accident, was this: Don’t get fancy with your champion, but if you truly don’t know which of the “very good” teams will win, pick the one nobody else is picking. This year, that champion is Virginia. The Cavaliers represent a HUGE market inefficiency. Despite their winning the ACC and earning a 1-seed, nobody knows exactly how good they are. I’m in a bracket group on ESPN with 128 other people, and two of us have Virginia winning it all. TWO! The other 1-seeds all have at least 10 takers, including Wichita, but everyone’s sleeping on the Cavs. What that means is that, no matter how bad my bracket is, I can still sneak a win despite a bracket barely over the 50th percentile due to the extremely high importance of picking the finalists and champion. In terms of MoneyBrackets, picking Virginia as a champion is the single best move anyone can make this year. The possible value gained is enormous.

4. Stay Away From the Trendy Conference Tournament Winner

I wrote: Last year, I ran the numbers … don’t be seduced by teams that won the conference tournament but were just so-so in regular-season play. Like Iowa State … Michigan State … or, on smaller scales, Providence (Big East) and Saint Joe’s (A-10).

The truth: Man, this advice was so close to being perfect. Providence and Saint Joe’s both got bounced by lower seeds (UNC, UConn) in the first round. Michigan State got a break (in my opinion) when Harvard beat Cincy, but even so, got tested by the Crimson in the round of 32. What really killed me, though, was Iowa State beating UNC. The Tar Heels were up 72-64 with 5:12 left, and then all hell broke loose. ISU finished 8-9 from the field, and 3-3 from three. It was brutal to watch as an ACC fan — almost as brutal as the Kentucky win. I thought Carolina was the better team, but, once again, we witnessed the cruel caprice of March. So we’ll have to wait a few days to know whether it was wise to stay away from Iowa State and Michigan State, but at the very least, they played to their seeds.

5. Ignore the Glimmering Snowcapped Allure of the Mountain West

I wrote: Fool me once, shame on you; fool me twice, shame on me; and if anyone gets fooled for a third time by the Mountain West, he or she should be shamed in the public square by Jerry Tarkanian.

The truth: New Mexico predictably loses to underdog Stanford, but San Diego State escapes in overtime against New Mexico State and then draws North Dakota State because … well, we’ll get to that. Anyway, trust me, the Mountain West is still hot garbage, and San Diego State will get rolled by Arizona on Thursday. (And no, I don’t care at all that SDSU kept is close with the ’Cats on November 14. People who bring up November games as if they mean anything in March are the worst … or on the selection committee.) You might have lost a few points picking the Aztecs to go down early, but the luck runs out soon.

6. Don’t Deviate From the “Three 1-Seeds, One Oddball” Final Four Strategy

I wrote: Since ’85, there have been 15 years when two or more no. 1 seeds reach the Final Four, and 14 years where one or none do so … because we’re not robots, and because there has been just one season when all four no. 1 seeds made the Final Four, we need an oddball … to me, the Sooners are the perfect oddball Final Four team.

The truth: Oops. This was the worst advice I gave, by far, and I apologize to anyone whose bracket it ruined. I didn’t want to admit this to myself at the time, but the actual, unsexy truth about picking your Final Four is that based on history, you probably shouldn’t ever pick anyone but 1- or 2-seeds — 3-seeds at a stretch — to reach the Final Four. Since 1985, 72 of the 116 Final Four teams have been 1- or 2-seeds. Add in the 3-seeds, and it’s 86 of 116, or 74.1 percent.

I hate this advice, because it makes for a boring bracket, but from a mathematical perspective it’s not smart to mess around with 4-seeds or below. There’s only a one-in-four chance an underdog actually makes it to the last weekend, and if you roll the dice like I did, your odds of picking the right one are infinitesimal. So Oklahoma was a stupid pick in two ways, and even though I really liked the Sooners’ chances against SDSU in a potential second round, I deserved that overtime loss to North Dakota State. Just as I deserved a completely destroyed Midwest Region for going too crazy. That’s not MoneyBrackets. Sorry!

7. In the Early Rounds, Never Choose the Inferior Game Coach

I wrote: Ride guys like John Beilein, Sean Miller, Gregg Marshall, Rick Pitino (seriously, I’m so angry those two teams have to play so early), Tony Bennett, and Bill Self. Be wary of icons like Coach K and Roy Williams who seem to have weird occasional lapses in big tourney games. Completely stay away from the likes of Drew, Travis Ford, Rick Barnes, and Mark Gottfried. Nobody, and I mean nobody, should be picking Oklahoma State over Gonzaga.

The truth: Other than failing to include Bill Self in the Coach K/Roy Williams “occasional lapse” category, this was pretty spot-on. Ford, Barnes, and Gottfried are gone; Beilein, Miller, Pitino, and Bennett are alive; and Self, K, and Roy had their lapses. Marshall ran into an extremely talented team peaking at the worst possible moment, and only Drew is still going among the “bad” coaches. On that note …

8. Never Forget That “Balanced Offense/Defense” Beats “Great at One, Bad at the Other”

I wrote: As I found here (ctrl-F “watch out for Maryland”), teams with great offenses and bad defenses rarely — almost never, actually — do well in the tournament … which means you need to be super, super wary of the following teams: Ohio State, VCU, Duke, Creighton, Michigan, San Diego State, Cincinnati, Iowa, and Baylor … take a really, really close look at Tennessee, with its great, balanced metrics, making an Elite Eight run.

The truth: This advice is rock-solid, and perhaps the best part of MoneyBrackets. Of the nine teams listed, only Michigan, San Diego State, and Baylor survive. We talked about SDSU (fluke, faced two imbalanced opponents), Baylor got to play two fellow imbalanced teams (Nebraska, Creighton), and Michigan has John Beilein. In fact, of all the games in which imbalanced teams faced balanced teams, there was just one — Michigan vs. Texas — that went to the imbalanced side. And that game featured Beilein vs. Rick Barnes.

Oh, and about Tennessee? I’d first like it on record that I wrote this before Ken Pomeroy’s piece on Deadspin was published, and second, that I almost took my own advice. I had the Vols beating Duke in my bracket, and was considering them in the Michigan game when I noticed a geographical quirk: The Duke-Tennessee game would be in Raleigh. I changed my mind, stupidly forgetting that unless a game is in the New York metropolitan region, Duke does not travel. Period. Not even the 20 minutes to Raleigh. The Mercer crowd, plus the neutrals, overwhelmed the small Duke presence in Round 1, and I never should have let this keep me from taking the Vols in Round 2. STUPID!

9. Don’t Be Too Impressed by a Mediocre “Major Conference” Team

I wrote: Controlling for teams seeded between 5 and 12, major conference teams are only 163-119 (58 percent) against minor conference teams. Change that to no. 6 through no. 11 seeds, and the record drops to 105-96. Seeds no. 7 through no. 11? 56-64 — a losing record! And in just the 8-9 games, major conferences are 26-30. This proves the whole “They played against stronger competition all year!” argument has always been dumb.

The truth: I don’t think I phrased this advice very well. Instead of implying that you should always take the mid-major, I should have just urged you all to consider the teams equal, since the numbers themselves are roughly equal. In other words, get rid of your major-conference bias when the seeds drop below five. This year, in matchups pitting big vs. little between seeds 5 and 12, major conferences went 4-5 (or 5-5, if you count the Big East as a minor conference and factor in UNC over Providence). Which is almost exactly what the numbers tell us to expect in those zones, and even a little better than expected for the mid-majors. Point is, evaluate these toss-up games on the merits of the teams, not by something meaningless like home conference or strength of schedule.

10. Don’t Have a Meltdown on Day One

I wrote: Be patient. Be calm. Be confident. You are a MoneyBracket Pro. (Unless you lose a Final Four team in the first round, in which case you should cry, set fire to your brackets, and threaten me on Twitter.)

The truth: I lost a Final Four team in the first round! My whole house burned down!

Final Analysis: Is MoneyBrackets Any Good?

Maybe I’m biased, but I think so. As we saw with the UNC and Wichita State games yesterday, and all the dramatic finishes on Thursday and Friday, it’s impossible to control for the insanity that’s poised to strike at every moment. Even teams like Louisville and Virginia can look jittery and vulnerable in first-round games against mediocre opponents, and then look like world beaters against much better teams a round later.

There were some mistakes. The whole “oddball” advice doesn’t make sense, and was a remnant of my love for making crazy picks that have little chance of panning out. Which I still maintain is really fun, but which doesn’t belong in MoneyBrackets. And I missed the obvious advice of finding the great team nobody is picking to win the title, even though I got lucky and took Virginia, one of the least-loved 1-seeds in tournament history.

Overall, though, MoneyBrackets identifies a few repeating patterns and exploits the poor analysis of the average picker. It’s subject to the whims of the basketball gods, sure, but on the whole it’s a useful tool with which to attack an uncertain bracket. I’ll be back next year with the new and improved version, and I’ll be watching the games in a padded room, far from anything flammable.

Filed Under: 2014 NCAA Tournament