UNC Baseball
- Thread starter wmheel1287
- Start date
- Replies: 343
- Views: 7K
- UNC Sports
- Messages
- 8,208
I think RPI is still effective in college baseball because of its simplicity in accounting more heavily for strictly win – loss record in series across a long season, whereas in basketball there are much better and more advanced efficiency metrics that take into consideration things like margin of victory that RPI does not account for, which do a better job of indicating team quality relative to their competition. RPI does a better job of measuring a team by the overall quality of its schedule, whereas metrics like NET in basketball do a better job of measuring a team’s strength by how well it performs in individual games versus a three game series like baseball.
Last edited:
akillesheel
Distinguished Member
- Messages
- 394
yeah, a lot of RPI's issues in the other sports come down to the fact that single-game outcomes are really noisy. That gets fixed to a significant extent in a sport that's played mostly in 3-game series, where it's statistically a safer assumption that the better team won. it's not always true, of course, but it's probably true enough to not be egregiously wrong. it does still get tanked if a team loses a bad midweek, though... but that's also not something the committee seems to care about when it comes selection time.
it's kind of an aside, but i do wish college baseball had better publicly available numbers. in particular, judging defense simply by fielding percentage feels like something out of the stone age, and the only time we get access to pitchers' mix and distribution is when we're lucky enough that they get mentioned on national broadcasts.
- Messages
- 3,302
Others have already covered this but simply put...RPI works pretty well over a 56 game schedule...particularly once they tweaked for more credit to road wins.