Similarity Predictions: 3/13

Yesterday’s experiment went fine, so I’m going to post a few more of these predictions, for today’s major conference championship games – without commentary this time.

Tickets Punched: #15 - #17

Only one bid was sealed last night, as Lehigh took down Lafayette at home.  But there have already been a couple championship games today, including one of the biggest games from a Bubble-busting point of view – Houston upsetting UTEP in the CUSA final.

LEHIGH

This is Lehigh’s strength, and they’re led by CJ McCollum (average shooting line: 4/8 2FG, 2/5 3FG, 5/6 FT).  The two NCAA teams in their comps both lost in the first round when their offenses were completely shut down – Vandy because they couldn’t hit a three to save their life (4 of 20), and the Zags because they couldn’t hit a two (12 of 37) or grab a rebound.  I’m guessing it will be more of the latter case for Lehigh, assuming they run into an opponent with good D.

Similarity Predictions

Since there were no tickets punched last night, I spent part of the evening developing an Excel workbook that makes game predictions based on how a team has performed in games against similar opponents.  I’ll use today’s game between Wisconsin and Illinois as an example.

First, I found the ten Wisconsin opponents whose offenses were most similar to Illinois, using the same method I’ve been using in the Tickets Punched series.  I then calculated Wisconsin’s individual game adjusted defensive efficiency against each of those opponents, using this equation (derived from the standard adjusted efficiency formula):

Wisc Game Adj Def Eff = [Opp Game Raw Off Eff] * [NCAA Ave Eff] / [Opp Season Adj Off Eff]

[NOTE: I actually treated HFA the same way Pomeroy does, but I didn’t want to make the equation look even messier]

Finally, I took a weighted average of those efficiencies, with the weight being the similarity score.  This gives me an adjusted defensive efficiency for Wisconsin that is specifically for games against offenses similar to Illinois.

I repeat the above process for Wisconsin’s offense, and Illinois's offense and defense, and the results all go into these pretty charts (where, as usual, green=good and red=bad):

Tickets Punched: #13 - #14

Part … I don’t know, 5? … of a series comparing the offenses and defenses of current auto-bid teams to recent major conference teams, in order to get a feel for their style of play.

If you’ve managed to find this, you’re deep enough into the college basketball blogosphere that it’s impossible for you not to have heard about Anthony Johnson’s remarkable performance last night, in which he scored 34 points in the second half – including his team’s last 11 – to lead Montana past Weber State.  Did you realize there was another auto-bid locked up last night?  Quinnipiac was upset by some guy named…

ROBERT MORRIS

Robert Morris is one of the few teams where their FTR is more correlated with their offensive efficiency than TO% or OReb% is.  It’s all well and good to get to the line often, but they depend on free throws to win.  In games where they took at least half as many FT’s as FG’s, they went 16-1, while in games where they shot fewer than half as many, they went 7-10.  I thought there might be a cause and effect problem here, as teams that are ahead get fouled, so I looked at the same thing for team with a similar record and Pomeroy rating (Western Carolina). They were also better in the high-FTR games, but not to the same extent – 6-2 vs. 16-7.  They should be hoping they don’t draw Syracuse, Kentucky, or Ohio State in the first round.

Tickets Punched: #10 - #12

Three more championship games were held last night, with Butler preventing Wright State from stealing a a bid, and Oakland and North Texas getting in via their only possible means.  So, time for fourth installment of my series looking at the offensive and defensive comps for the auto-bid winners.  Previous posts can be found here.

BUTLER

You can see that this year’s Butler team is fairly similar to last year’s – they shoot better from 2 but worse from 3, and the turn the ball over slightly less but also rebound slightly less.  Overall, the differences are mostly a wash.  Their comps, for the most part, didn’t do a whole lot in the postseason, but the offenses weren’t to blame for their losses – the only case where an offense scored under 1 point per possession in a season-ending loss was 2007 Xavier, but that was against a great Greg-Oden-led Ohio State defense.

Tickets Punched: #7 - #9

Part 3 of my tour through the teams who’ve received automatic bids to the NCAA tourney.  (See Part 1 and Part 2.)

SAINT MARY’S

I think it’s funny that their top offensive comp is an old Zags team – apparently they’ve modeled themselves after the example of quality that’s closest at hand.  The Gaels can shoot from inside and out, and take care of the ball.  Their only subpar trait is that they don’t get to the line very much (not shown).

Tickets Punched: #4 - #6

An exploration of the NCAA automatic bid winners, continued.  Part 1 is here.

NORTHERN IOWA

The red in this case is just due to a slower pace – not a negative thing, but I had to choose a color.  Northern Iowa’s offense is very deliberate – extremely low turnovers, slow pace, above average efficiency.  I can definitely see the Butler or Wisconsin comparisons.  Their problem is, they picked the mediocre-shooting 2004 Butler, instead of the decent 2007 or 2008 Butler.  They’ll hold onto the ball, but they won’t hit a ton of their shots, and they won’t get second chances on the ones they miss.

Tickets Punched: #1 - #3

Due to an unfortunate incident involving my external hard drive, a parquet floor, and gravity, the files for my similarity scores system were lost this weekend.  I rebuilt most of it last night (though I haven’t had time to add the NCAA performance data for historical teams).  Given a chance to start from scratch, I made a couple changes – one is that I can now easily find comparables for only the offensive or defensive unit of a team.  I’m thinking I should be able to find better fits when looking at only one side of the ball, versus trying to find teams that play similar on both ends of the floor.  For example, I previously saw 2004 Wake Forest come up as a fairly good comparable for some very good all-round teams, despite Wake’s mediocre defense – simply because their offense was a perfect match.  Now the Deacons can move up to the top of the offensive comps list, and I won’t fret about the defense being a bad fit.  If the left shoe fits, wear it, and find another right shoe.

The other change is that I added an auto-chart-creation section that allows me to spotlight the categories which led to the matches, so the comparables are less of a black box.  You’ll see what I mean below, as I’m going to take these charts on a test drive through the teams that have clinched automatic bids so far.  Each of these charts will show the 5 categories that were weighted the highest when calculating the similarity score (SIM in the chart).  These will be areas where the team is particularly good or bad (e.g. 3P% below), along with categories that have a high weight just because they’re always important (e.g. Offensive Efficiency).  The values in these categories for the spotlighted team will be shown, as well as how many standard deviations above or below average they are.  This part will be color coded based on the standard deviations, so you can see at glance whether a team was good (green) or bad (red).  Last, the top 5 comps will be listed, along with their similarity score, and their values for the 5 key categories.  It might sound more complicated than it is; you’ll figure it out right away. (Oh, I should also note that I am only comparing teams today to major conference teams, plus a few exceptions like Gonzaga, Xavier, 2006 George Mason, etc.)

CORNELL

Vegas Watch was spot on when he compared Cornell’s offense to 2009 California – they both hit plenty of 3’s and take care of the ball well, but do little else.  Though Cal somehow managed to make it to the line much more than Cornell.