How baseball RPI works

[Bulldogg31]

I'm sure you guys know this, but it's still interesting how much our 20 straight home games to open a season hurts us:

"The RPI Formula used in NCAA baseball is the same as that used in basketball except for the adjustment of home and road records. Starting in 2013, college baseball "RPI formula will value each road victory as 1.3 instead of 1.0. Each home win will be valued at 0.7 instead of 1.0. Conversely, each home loss will count 1.3 against a team’s RPI and each road loss will count 0.7 against a team’s RPI. Neutral-site games will retain the same value of 1.0, but the committee is studying how to determine if a game should be considered a neutral-site contest. The adjustment is based on data showing that home teams win about 62 percent of the time in Division I baseball."[2] The change was made because of the discrepancy in the number of home games teams play. Some schools are able to play 35-40 of their 56 allowable games at home, while other teams, due to factors such as weather, may play only 20 home games."

So even if we started 20-0, RPI only gives us credit for about 13 wins. Wow.

 

[57stratdawg]

That's why when people say "the problem with our RPI are bad losses" I always shake my head. We have to learn that building a resume today is more than just banking wins.

 

[QuaoarsKing]

It's not a very well-thought-out formula, unfortunately.

It works fine (not great, but it's OK) for basketball. It's doesn't do a very good job of ranking baseball teams though.

 

[8dog]

It rewards road games as an accomplishment that really isnt an accomplishment. Winning on the road in baseball isnt like doing it in football or hoops. And it was done primarily to get teams like us to not play 20 home games and not necessarily because it is a great accomplishment. We showed them!

But yes the rpi doesnt care who you lose to. 1-1 at home against lsu and valley is the exact same rpi number regardless of which one you beat.

 

[Bulldogg31]

So to build RPI we should open the season with a 3 game home series, then go on a barnstorming tour of Jackson St, Miss Valley, etc. Call it a goodwill/recruiting tour.

Not sure it would hurt attendance much, since there's only about 50 people at the Tues night JaxST game anyway.

So the folks saying we need to stop playing low RPI teams have it wrong for baseball. In RPI terms for baseball, beating Mississippi Valley on the road is better than winning at home against Miami.

So we need less Florida St, and more Delta State (in Cleveland, of course).

Only half kidding here.

 

[QuaoarsKing]

Nah, that wouldn't really help.

We'd get 1.3 wins for beating Jackson State rather than 0.7, but it would hurt our SOS a lot.

Would be better than playing them at DNF, yes, but we should just not play them at all.

 

[dawgphd]

Ok someone help me with this.

UM SOS (1st) is credited with their just above .500 record yielding a 29 RPI.

Our SOS (29th) and 24-30 record manifested a 124 RPI.

Our SOS was shockingly only 9th in the SEC. RPI was last.

A 20's SOS seems reasonable. It's middle of the pack in the SEC.

Lets look at 3 similar teams: ( We should play these in a NIT type regional)
Baylor 21-30 27 SOS 125 RPI
Wichita State 24-31 34 SOS 111 RPI
Stanford 22-31 22 SOS 114 RPI

Now a couple with similar SOS and 10ish more wins
Upig 33-20 33 SOS 37 RPI
FAU 38-15 26 SOS 16 RPI

So don't we really just need more wins?

A harder schedule will aid. But it seems the SEC takes care of most of the SOS. With this year's SOS and 10 more wins and we'd be
in good shape. Frustratingly, it's not real had to find the 10.

 

[8dog]

Yes we need more wins which you get by playing road games. You have to look at adjusted win loss record to know a teams record. Boyds world usually has that info.

 

[RebelBruiser]

Basketball I believe uses 1.4 and 0.6 for its adjustments.

Baseball may have adjusted DOWN to the 1.3 and 0.7. I think they used to do it just like basketball.

I'm personally not a big fan of RPI, but I called this out earlier in the year. I'm just surprised no one has briefed Cohen on the importance of 1) playing someone good OOC, and 2) playing on the road OOC. He did the same thing at UK, and it cost some of his teams either a spot in a regional, a host, or a national seed on a couple occasions. You can't play a bunch of cupcakes at home and expect to get rewarded for putting up a gaudy record.

This is the same issue the SEC had in a BIG way in basketball, until Slive finally hired someone to come in and help us with scheduling (basically by telling the league to quit being wimps, play on the road, and play some good teams).

Again, I think RPI is a weak formula, but if the Committee is going to use it as a major metric, and it's easy to manipulate, you're an idiot if you don't schedule yourself in a way to best manipulate the formula.

 

[RebelBruiser]

Yes we need more wins which you get by playing road games. You have to look at adjusted win loss record to know a teams record. Boyds world usually has that info.

I pulled it from Warren Nolan using the records he had for home, road, neutral, and the 1.3/0.7 adjustments.

Here is how it stacks up to give you an idea where the difference came in:

MSU
Adjusted Record: 18.6-30.0 (0.383)
SOS: 0.552
Overall RPI: 0.5097

Ole Miss
Adjusted Record: 27.3-24.4 (0.528 )
SOS: 0.583
Overall RPI: 0.5692

When you consider that record is 25% of RPI, and SOS is 75% of RPI, the Adjusted record makes up 0.036 of the difference in our RPIs. The SOS portion only makes up 0.023 of the difference in our RPIs, with some rounding in there.

If you were to take an unadjusted win percentage, the overall record would only have a 0.025 difference on our RPIs, so the fact that we played more road games had a 0.011 impact on that RPI difference. A difference of 0.011 would've moved your RPI up about 20 spots. An adjustment down of our RPI by about 0.011 would've moved us down about 15 spots in the RPI.

That's why playing road games is important. I'd be curious to know if anyone is going to show figures to Cohen, or if Stricklin will step in on the scheduling, because not leaving Starkville does nothing but hurt your program's chances at seeding come tournament time.

Add in the 0.023 difference in SOS impact, and you've got a total of 0.034 difference in our two RPIs based solely on schedule (either home/road or quality of opponent). That kind of difference would've bumped you up from 124 to 63 in the RPI. Taking that much away from our RPI would drop us from 24 to 74.

As an aside, we played 6 more road games than you did this year. That doesn't seem like a lot, but we had 7 more road wins on the year, which makes up a good bit on the adjusted record.

In reality, you're rewarded for playing weaker teams on the road, because you can boost your adjusted record component a good bit, and the SOS hit overall won't hurt you as much. Anyone not taking advantage of RPI at this point is missing the boat. It's a weak and easily manipulated formula.

 

[Bulldogg31]

I'm no math wiz but getting 1.3 for a road win makes it nearly twice as impactful as a home win (.7)

Maybe an out-of-conference schedule looks something like 14 home games, 8 road games, and 4 neutral field (Gov Cup, tourney, etc).

 

[engie]

Yes, it's very perplexing that we don't exploit RPI more/better. That's why a couple of us melted pretty hard about our schedule in the preseason. There was never any margin for error this year, and once we lost a couple of those games, we put ourselves out of contention, and it snowballed. This year's schedule was another thing on a long list of glaring issues that have got to be fixed.

That said, I agree that baseball RPI is a BS formula due to the home/away ratings that give a false sense of home field advantage in the sport. Boyd and others try to establish true home field advantages for each individual school over time and use those in his their personal formulations instead. Which is why those are generally better formulas than RPI IMO. That said, RPI being as ridiculous as it is makes it easily exploitable if we would just pull our heads out of the sand to do it.