BBCOR and the New RPI
May 29, 2014
It's not going to surprise any fan that scoring remained ~10 runs per game this year, with an average score almost exactly 7-3. That's about 3 runs per game less than the last year before BBCOR. The final regular season numbers provide some insight into the scheduling effect of the new RPI in its second year.
Regular Season Scoring - 2014†
| All Reg Season |
| Cat | Gms | WR | LR | | MOV | TOT | SOV | | — | Vis Wins: | WR | LR | | MOV | TOT | SOV | Home Wins: | WR | LR | | MOV | TOT | SOV | | Home W% |
| D1 | 8054 | 7.07 | 2.99 | | 4.08 | 10.07 | 0.408 | | | 3334 | 7.16 | 4.15 | | 3.00 | 11.31 | 0.300 | 4720 | 7.03 | 2.18 | | 4.85 | 9.21 | 0.485 | | 58.6 |
| Conference | 4256 | 6.84 | 3.00 | | 3.84 | 9.83 | 0.384 | | | 1936 | 7.08 | 3.04 | | 4.04 | 10.13 | 0.404 | 2320 | 6.64 | 2.96 | | 3.68 | 9.60 | 0.368 | | 54.5 |
| InterConference | 3798 | 7.34 | 2.99 | | 4.35 | 10.33 | 0.435 | | | 1398 | 7.26 | 5.69 | | 1.57 | 12.95 | 0.157 | 2400 | 7.41 | 1.43 | | 5.98 | 8.84 | 0.598 | | 63.2 |
| Non-Neut |
| D1 | 7153 | 7.10 | 3.00 | | 4.10 | 10.10 | 0.410 | | | 2884 | 7.17 | 4.33 | | 2.84 | 11.51 | 0.284 | 4269 | 7.07 | 2.11 | | 4.96 | 9.18 | 0.496 | | 59.7 |
| Conference | 3859 | 6.86 | 3.00 | | 3.86 | 9.87 | 0.386 | | | 1739 | 7.13 | 3.06 | | 4.07 | 10.19 | 0.407 | 2120 | 6.65 | 2.97 | | 3.69 | 9.62 | 0.369 | | 54.9 |
| InterConference | 3294 | 7.38 | 3.00 | | 4.39 | 10.38 | 0.439 | | | 1145 | 7.24 | 6.27 | | 0.97 | 13.51 | 0.097 | 2149 | 7.48 | 1.26 | | 6.22 | 8.74 | 0.622 | | 65.2 |
| Neutral Site |
| D1 | 901 | 6.85 | 2.93 | | 3.92 | 9.78 | 0.392 | | | 450 | 7.04 | 3.00 | | 4.04 | 10.04 | 0.404 | 451 | 6.68 | 2.87 | | 3.81 | 9.55 | 0.381 | | 50.1 |
| Conference | 397 | 6.59 | 2.91 | | 3.68 | 9.50 | 0.368 | | | 197 | 6.65 | 2.93 | | 3.73 | 9.58 | 0.373 | 200 | 6.52 | 2.89 | | 3.63 | 9.42 | 0.363 | | 50.4 |
| InterConference | 504 | 7.06 | 2.94 | | 4.12 | 10.00 | 0.412 | | | 253 | 7.35 | 3.05 | | 4.29 | 10.40 | 0.429 | 251 | 6.81 | 2.85 | | 3.96 | 9.65 | 0.396 | | 49.8 |
|
Quick and easy bullet-points:
- Three cheers for early-season tournaments! More than one of every eight interconference matchups were at neutral sites. Sure, too much of that is still climate-challenged teams playing each other in Florida, but there was enough "hey, we're gonna be there anyway, got an open date?" so that at the end of the regular season more than 75% of the teams had more than half the field as either opponents or opponents' opponents, and 57.5% of all team-pairs were at least O-Os. That's the most of any D1 team sport.
- The fact that many of the neutral-site games are those "we both had to find a place to play" matchups may not be all that bad. The neutral site interconference games were generally as competitive as conference games overall (the lower the SOV value the closer the typical game.) Of course, for neutral site games "Home" just means "batted last."
- The new RPI formula's home field adjustment factor is just flat wrong, and you can see that by comparing the Home W% column to the MOV or (better) SOV columns. More below...
Requisite RPI Bashing Section
Actually, I don't bash the RPI in general. Every time someone who knows how it works says something like "couldn't the NCAA get the Stanford math department to come up with something?" they get an email from the NCAA that says "we did." So I don't say things like that.
I do think that the "new" RPI is not using the right adjustment factor for road wins/home losses. They based the factor on the home winning percentage for all non-conference games over a long time period. That's about the same as 2014's 65.2%. The problem with that is that it doesn't take into account the percentage of those games where the "home" team would've won against the opponent even if the game were played on Mars. Interconference games not played at neutral sites are loaded with visitors "not good enough to get teams to come to our place so we'll get paid to get beat by you."
The empirical data (in conference games the home team wins 54.9% of the time) keeps confirming Boyd's observation that the HFA should be 1.1. Over many years approximately equal teams' games result in the home team winning 55% of the time. The 65% or so used to come up with the 1.3 weight only results in a really, really bad measure of teams' body of work.
Anyway, looks like we had a lot of fun this year. The college baseball critics who said "it isn't like MLB" can't say that now that it is.
† Runs per 9 innings.
In memory of
Paul Kislanko