Wednesday, December 23, 2015

How To Inflate Your Resume: How Texas Tech Built A Nearly Perfect Schedule

Tubby Smith has talked in interviews about how important RPI is to his scheduling

The topic of RPI manipulation is a controversial one. Some coaches actively do it, while others deny that it's possible. The NCAA powers that be don't seem to believe that RPI manipulation is possible, or they assume that the Selection Committee can see through it.

The reality is that RPI manipulation is real, and surprisingly easy. Unlike complicated and well-developed computer ratings like Pomeroy or Sagarin (college football has ratings like S&P), the RPI is extremely simple. You can see the formula here.

There are two huge flaws with the RPI. The first is that home/road is poorly taken into account. In fact, home/road is not taken into account at all in RPI SOS. So you want to avoid road games as much as you can. The second gigantic flaw is that the RPI does not consider how good the teams you play actually are, but rather what their W-L record is and what the W-L record of their opponents is.

Here's an extreme example from last season:

Xavier went just 23-14 due to a very difficult schedule, which was why they finished 22nd in Pomeroy and earned a 6 seed in the NCAA Tournament. Albany went 24-9 despite not being a very good team thanks to a very soft schedule, which was why they were just 128th in Pomeroy and earned just a 14 seed in the NCAA Tournament.

On first glance, the RPI seems to have these teams rated properly: Xavier was 28th in RPI and 8th in RPI SOS, while Albany was 104th in RPI and 277th in RPI SOS.

The problem? Via the RPI formula, your RPI SOS is better off facing Albany than facing Xavier. In fact, facing Albany at home would be treated better than facing Xavier on the road. Imagine you were a team that just narrowly missed the NCAA Tournament, such as Temple. Via the Sagarin ratings, Temple would have been a 10 point underdog at Xavier and a 10 point favorite at home vs Albany. Yet they'd get more credit for the Albany game. That's a massive gap for creative coaches to exploit.
So let's say you want to manipulate your RPI. What are the three main rules to follow (along with caveats/clarifications):

1. Avoid Road Games Like The Plague*
      *1A. Road games are okay if they are against truly elite (Top 15 or so) teams.
2. Avoid playing the dregs of Division I (RPI 250+).
      2A. If you require a couple of true "cupcake" games, try to replace them with non-Division I opponents and with .500+ teams from bad conferences.
3. Schedule as many home games as you can against RPI Top 50 opponents.

So that brings us to a potential bubble team that has put together a nearly perfect schedule this season: Texas Tech. How did they do on each rule?

1. Avoid road games like the plague The only non-conference road game on Texas Tech's schedule is a non-voluntary game: Arkansas in the Big 12/SEC Challenge.
2. Avoid playing the dregs of Division I They do have one slip-up here, facing a bad Arkansas-Pine Bluff team. But besides that they have Tennessee-Martin (projected to finish ~.500), Sam Houston State (a contender in the Southland), and then every mid-major team on their schedule is likely to finish in the RPI Top 150. It's hard to find any other team in the nation that will do a better job of avoiding the dregs of Division I.
3. Schedule as many home games as you can against RPI Top 50 opponents Texas Tech has a slew of 2014-15 Albany-type teams in South Dakota State, UA-Little Rock, High Point, and Hawaii, all of which they have played at home. And in conference play they will get plenty of more traditional RPI Top 50 opponents in the very deep Big 12. To earn an at-large bid, Texas Tech will need "RPI Top 50 wins", but even last season, when they were the worst team in the Big 12, they still collected home wins vs Iowa State and Oklahoma State while coming up just short against Oklahoma (overtime), Baylor (5 points), and Texas (9 points). This is the significant built-in advantage that power conference teams have over mid-majors, who struggle mightily to get RPI Top 50 teams on their home court.

Just how valuable is Texas Tech's scheduling? We can compare their potential RPI to other Big 12 teams (since the Big 12 has a full round-robin schedule) using RPIForecast. Right now, Pomeroy projects Texas Tech to go 8-10 in Big 12 play while Sagarin projects 7-11, and Texas Tech has zero non-conference wins over teams likely to finish in the RPI Top 50, so that seems like a resume that shouldn't be particularly close to an at-large bid.

Let's compare Texas Tech to other .500-ish and potential bubble Big 12 teams: Texas, Baylor, Kansas State, and West Virginia. Texas has a win over North Carolina with only one iffy loss (Washington). Baylor has a win over Vanderbilt with no bad losses. West Virginia and Kansas State have no very good wins but no bad losses either. So looking at their non-conference performances alone, Texas and Baylor seem to be in a better spot than Texas Tech, while Kansas State and West Virginia are comparable. What RPI are these teams projected to finish with:

Go 7-11 in Big 12 play:
47 - Texas Tech
53 - West Virginia
56 - Texas
67 - Kansas State
69 - Baylor

Go 8-10 in Big 12 play:
37 - Texas Tech
45 - Texas
46 - West Virginia
54 - Baylor
54 - Kansas State

Go 9-9 in Big 12 play:
28 - Texas Tech
36 - Texas
36 - West Virginia
42 - Kansas State
42 - Baylor

So despite Baylor playing better than Texas Tech, they would be 15-20 spots in RPI worse than Texas Tech for the same conference performance. Even Texas, with that massive win over North Carolina and despite playing in a better Thanksgiving Week tournament than Texas Tech (important for RPI purposes) would still likely finish around ten slots lower in RPI than Texas Tech for the same conference performance.

How valuable is that? Well, in the history of the 64+ team NCAA Tournament, no power conference team with an RPI of 40 or better has ever failed to earn an at-large bid. All Texas Tech likely needs to finish in the Top 40 is go 8-10 in the Big 12, despite likely zero RPI Top 50 wins in non-conference play. That is staggering.

In other words, if Texas Tech goes 8-10 or 9-9 in Big 12 play, they are a near certainty to make the NCAA Tournament because of their schedule. And for that, Tubby Smith deserves a ton of credit. It's dumb that the RPI rules this sport the way it does, but as long as it does you have to take advantage of it.

1 comment:

Matthew Hayden said...

Great read! IIRC, last year, TCU was the worst team in the Big 12 with regards to MBB.