March Statistical Madness aka Danny's Just Asking for It

Hey everyone...Tip-off for the Gator's biennial run towards an NCAA basketball championship begins in just about 30 minutes...OK, let's start over because that was an absurd statement...

Hey everyone...Tip-off for the Gator's biennial run towards an NCAA basketball championship tournament begins in just about 30 minutes, so just wanted to give you some last-minute information to chew on prior to gametime...

Let me first start off by saying that I am by no means a college basketball junkie. Frankly, I'd put the number of games I watch during the season at somewhere around 3.14 (happy belated Pi Day). The funny thing is that I do reasonably well in bracket pools because I transfer knowledge from other areas of expertise. Namely, football drafting.

If you've read any of my draft-related posts over the past couple of years here, you should know that I'm a firm believer in making sure a draft accomplishes 2 goals:

1. Avoiding busts
2. Finding diamonds in the rough

In the NFL and fantasy football, this means that you absolutely have to nail your #1 pick and you have to find a couple of sleepers in the later rounds. Therefore, relative to the tourney, my philosophy revolves around making sure I get my champion right (i.e., avoiding busts), and picking a few Cinderellas to advance in the first 2 rounds (i.e., finding diamonds in the rough).

From there, it's about the stats. Unlike the case with football stats, basketball stats are tracked much more effectively and tend to be more predictive. Essentially, although both football and basketball are both dynamic (i.e., multiple players are constantly moving) sports, basketball has fewer moving parts and most events (e.g., a made FG) are reliably attributable to a single moving part.

The rest of this post will be devoted to (a) giving you a few statistical resources with which to make your picks or last-minute changes, and (b) identifying my bust-avoiding and diamond-in-rough-finding picks. As a fun little experiment, I'm going to enter a bracket based on this article into our NN's tourney pool under the name "DamnedLies."

After the jump, some non-pretenders, stealth-contenders, and my picks...

First things first...there's a plethora of statistical info on the web looking at the factors affecting tourney success, so I'll limit this list to a few that I think are the most informative:

• John Ezekowitz from the Harvard Sports Analysis Collective (HSAC) has been running a series of posts about advanced team stats (from Ken Pomoroy's site) that influence upsets by lower seeds and early exits by higher seeds. Posts on the influence of tempo on upsets here and here, the identification of likely Cinderellas in this season's tourney here, and the identification of likely pretenders here.
• Peter Tiernan of Bracket Science has been running a series of posts over at CBS Sportsline. Article detailing the statistical history of every seed matchup since the tourney went to 64 teams in 1985 here. Article providing statistical guidelines for identifying upsets and choke-jobs here. Article identifying likely Cinderellas and pretenders here.

With respect to the philosophy I talked about before the jump, and without going into the gory details in this post, the essential things to take away from the resources above are as follows. From Ezekowitz:

1. Upsets that occur in the first 2 rounds are likely to be fast-paced games (i.e., have a higher-than-usual number of possessions).
2. The biggest predictors of a lower-seed upset are lower turnover percentage on offense and lower ratio of free throw attempts allowed and field goals allowed.
3. The biggest predictors of a higher-seed choke-job are higher turnover percentage on offense, lower offensive rebound percentage, and higher 3-point field goal percentage allowed.

From Tiernan:

1. Don't pick 1, 2, or 3 seeds to lose in the 1st round. They win 93.6% of the time (297-21) since 1985.
2. Don't pick a 1 to lose before the Elite 8. They get to the Elite 8 93.3% of the time (235-17) since 1985.
3. Relatedly, don't pick a 1 to lose to a 9 in the 2nd round. They win that matchup 94.4% of the time (51-3) since 1985.
4. 2 seeds are far more likely to lose to a 10 than a 7. They beat 7s 71.7% of the time (43-17), but beat 10s only 58.3% of the time (21-15).
5. However, if a 2 makes it to the Sweet 16, don't pick them to lose to a 6 or 11. They win those matchups 80.6% of the time.
6. High pointst-per-game and participation in the previous season's tourney are keys to identifying both Cinderellas and pretenders.

No doubt, these resources provide waaaaaaaaaaaaay more statistical detail than I have space for here, but that's pretty much the most important stuff you need to know.

So, who do I like and dislike? Based on the totality of the stats I've looked at, here's how my picking philosophy shakes out.

NON-PRETENDERS AMONG TOP 4 SEEDS

Kansas, Duke, West Virginia, Pittsburgh, Maryland

PRETENDERS AMONG TOP 4 SEEDS

Kentucky, Syracuse, Villanova, Kansas State, Ohio State, Purdue, Vanderbilt

DIAMONDS IN THE ROUGH

St. Mary's, Washington, Utah State, Murray State, Siena

MISCELLANEOUS FINDINGS

Among, the 5-8 seeds, Texas A & M, Notre Dame, Oklahoma State, and UNLV are most likely to be upset, whereas Temple and Xavier are most likely to be legit.

There you have it. That's what the stats have to say. Feel free to follow my lead on this one. Putting everything together, here are my picks:

FIRST ROUND WINNERS

Kansas, Northern Iowa, Michigan State, Maryland, Tennessee, Georgetown, Georgia Tech, Ohio State, Syracuse, Gonzaga, Butler, Murray State, Xavier, Pittsburgh, Brigham Young, Kansas State, Kentucky, Texas, Temple, Wisconsin, Washington, New Mexico, Missouri, West Virginia, Duke, California, Utah State, Siena, Old Dominion, Baylor, St. Mary's, Villanova

SECOND ROUND WINNERS

Kansas, Maryland, Tennessee, Georgia Tech, Syracuse, Butler, Pittsburgh, Brigham Young, Kentucky, Temple, New Mexico, West Virginia, Dule, Utah State, Baylor, St. Mary's

SWEET 16 WINNERS

Kansas, Tennessee, Syracuse, Pittsburgh, Kentucky, West Virginia, Duke, Baylor

ELITE 8 WINNERS

Kansas, Pittsburgh, West Virginia, Duke

FINAL 4 WINNERS

Kansas, Duke

NCAA CHAMPION

Kansas

Here's hoping I do well. No doubt going to be crucified on here if I don't. But, remember, I said I'm NOT a college basketball junkie.