Friday, March 19, 2010

Sometimes your Final Four odds get better, even if you don’t play

I used to track conference performance by listing wins and losses. Then I tracked Conference PASE which was Performance Against Seed Expectations. But this year I’m going to try something new. Today I want to track conference winners and losers in terms of the changes in Final Four probabilities.

I start by adding up the Final Four probabilities for the multi-bid conferences heading into the tournament.

0.931 Big 12
0.831 Big East
0.771 ACC
0.584 Big Ten
0.271 SEC
0.204 MWC
0.125 A10
0.082 Pac10
0.035 WAC
0.031 CUSA
0.026 WCC
0.109 Other

These are cumulative, so the best-case scenario for a conference would be for the total to reach 4.000. This would mean the conference has all four teams headed to the Final Four.

Not surprisingly conferences with many teams and good seeds such as the Big East and Big 12 have the best chance of putting teams in the Final Four.

Next I see how these probabilities changed on Thursday based on two things.
1) The wins or losses by the teams
2) Results in other games

As an example, Washington was a slight underdog against Marquette, but the game was nearly a 50/50 toss-up. So when Washington won, their Final Four odds increased from 0.042 to 0.086. This increase is "Self". But Washington’s Final Four odds also changed based on other games in the region. Because New Mexico advanced, Washington’s odds actually fell back to .078. This drop is "Other".

Initial = Initial probability
Self = Self-inflicted change
Other = Change in odds based on other results
New = New odds

At first glance, the Big 12 had a great day. Their teams went 3-1. But Kansas, Kansas St. and Baylor were heavily favored, so those wins didn't mean much. Instead, Texas losing meant the conference's Final Four odds went down.

The Big East had the worst day with Notre Dame, Marquette, and Georgetown losing. This allowed the ACC to move ahead of the Big East in terms of total Final Four probabilities. The losses by the Big East also improved the odds that the Big Ten and SEC teams would reach the Final Four.

-The big winners today included Butler and Tennessee who have the potential to be sleepers, but had to get through tough games today.

-Kentucky also won big today. The opening round win was part of it, but the key factor was Texas losing to Wake Forest. Texas had much better efficiency numbers on the season, so this really helps Kentucky’s odds of advancing.

-The other big winner today was Ohio St. Even though the Buckeyes didn’t play, Georgetown’s departure from the field was a big help. Ohio St.’s Final Four odds are now 22.7% instead of 19.0%

-The big losers today are the good teams that lost. These include Georgetown, Texas, Marquette, and Vanderbilt.

But Kansas St. was also among the day’s losers, despite beating North Texas. That’s because BYU advanced against Florida. BYU is a much stiffer test for K-State in round two.

Final notes: These probabilities use the Ken Pomeroy’s pre-tournament Pythagorean Winning Percentages and log5 odds, not my adjustments from earlier this week.

Quick Thoughts

Thursday was the most dramatic NCAA tournament day in several years. But since there’s a ton of coverage of these games, I’m going to limit myself to two quick comments:

-In the second half of the Georgetown game, while dribbling the ball against some limited full-court pressure, Ohio’s DJ Cooper picked his teammate’s mouthpiece off the floor and handed it to him. You just can’t make this stuff up.

-With the UNLV – UNI game tied, UNI had the ball with a chance for the last shot. It was the classic, win or go-to-overtime scenario. But rather than just sit back and let UNI run the clock down, UNLV pressure the ball and tried to force a turnover. It ultimately didn’t work because UNI held onto the ball and hit an amazing deep three to win, but I still thought it was a great strategy. Fear of fouling keeps most coaches from using pressure defense in a tie game, but why sit back? Teams practice getting a shot off in the final seconds all the time. And since that is such a practiced play, why not try to disrupt the timing? I also wonder what would have happened to UNI if the refs had started counting for a “closely-guarded” 5-second violation.