Why College Baseball RPI is Flawed

[Todd4State]

Someone on here told me that 50% of RPI is your opponents won/loss record. That makes RPI a hugely flawed stat because it hurts teams that play West Coast teams.

Baseball is a little different in the South than it is out West and also in the North as well. One major difference is we have some really, really BAD teams here. I'm not talking about Sun Belt or C-USA teams at all. There are definitely plenty of GOOD teams. The BAD teams I am talking about are SWAC teams. Most every SEC team plays a few SWAC teams and some UT-Martin's which are basically almost programmed wins and glorified scrimmages.

Out West, you have a lot more quality mid majors that are more comparable to a South Alabama or USM. And people like USC play them pretty much every week. Cal State Fullerton, Long Beach State, Cal Santa Barbara among others can more than hold their own with any PAC 12 team. Basically you have a smaller gap between the mid majors and the PAC 12 teams. And they ALL want to beat the USC's and UCLA's every year- which means those teams are basically someone's World Series every Tuesday. And there aren't very many SWAC-esque teams out there.


The result is everyone out there beats each other up, which results in about a 30 win season if your team is good. However the problem is RPI perceives that as your opponent being "down" which thus hurts your RPI. If USC or Oregon played SEMO State this year in a three game series I would imagine that the PAC 12 team would likely win a three game series because the PAC 12 team is the better team. RPI says that SEMO State is "better" based on their math though.

The good news is the selection committee doesn't base EVERYTHING on RPI. It's only one component of the equation.

Now, I do like playing PAC 12 teams because they do play a different style and we may very well see that style in a regional and have to deal with it. I do think that one thing we could do is maybe schedule an ACC team as well as a PAC 12 team. That might help us keep from getting burned by RPI when our PAC 12 opponent inevitably can't hold up their end of the deal.

Or they could do like Shotgun said and just actually pick the best teams.

[RocketDawg]

Or they could do like Shotgun said and just actually pick the best teams.

Problem with that is that it's subjective, which leads to all sorts of arguments ... like the polls in football do. Baseball has polls too but they're not all that consistent. I suppose they're trying for a system that is definitive but maybe it's not working very well either. It'll all work out in the end in Omaha.

[messageboardsuperhero]

My biggest issue with RPI is that it weights all home games equally, as well as all road games equally... So a game against Texas A&M in mid-April is weighted the same as a game against Alcorn in February. I think RPI should put more of a premium on elite wins (i.e., wins over top 10 teams) and be more forgiving on a couple of bad losses a year.

What is happening with us right now is that we are being held down about 7-8 spots because of a couple of bad mid-week losses before SEC play started. Does any really think the team we have should be defined by those fluke losses or by the 8 elite, top 15 wins we have? Take UC Santa Barbara for example- what elite teams have they beaten? They have no top 10 wins, and their best wins are single game wins vs Kentucky and Bryant and also going 1-1-1 against Charleston. Hell, UCSB even lost a series to Oregon, who we dominated. Does anyone honestly believe that they are deserving of being 8 spots better than us because they haven't had any bad "that's baseball" type of losses? It's asinine.

I think RPI needs one or two RPI "drop games" and put more emphasis on wins over elite teams- because that is much more indicative of the quality of team you have.

[BulldogBear]

My biggest issue with RPI is that it weights all home games equally, as well as all road games equally... So a game against Texas A&M in mid-April is weighted the same as a game against Alcorn in February. I think RPI should put more of a premium on elite wins (i.e., wins over top 10 teams) and be more forgiving on a couple of bad losses a year.

What is happening with us right now is that we are being held down about 7-8 spots because of a couple of bad mid-week losses before SEC play started. Does any really think the team we have should be defined by those fluke losses or by the 8 elite, top 15 wins we have? Take UC Santa Barbara for example- what elite teams have they beaten? They have no top 10 wins, and their best wins are single game wins vs Kentucky and Bryant and also going 1-1-1 against Charleston. Hell, UCSB even lost a series to Oregon, who we dominated. Does anyone honestly believe that they are deserving of being 8 spots better than us because they haven't had any bad "that's baseball" type of losses? It's asinine.

I think RPI needs one or two RPI "drop games" and put more emphasis on wins over elite teams- because that is much more indicative of the quality of team you have.

Wow, I've thought RPI was overrated for many many years now, and that includes roundball and football as well. But I didn't even realize it was that flawed. I thought they DID somewhat factor in things like good wins on road. So, you're telling me they count winning 2 of 3 at Florida the same as if we had played an early season weekend series on the road against say, Tulsa, with the same result? Good grief. Now, I really hate RPI.

[messageboardsuperhero]

Wow, I've thought RPI was overrated for many many years now, and that includes roundball and football as well. But I didn't even realize it was that flawed. I thought they DID somewhat factor in things like good wins on road. So, you're telling me they count winning 2 of 3 at Florida the same as if we had played an early season weekend series on the road against say, Tulsa, with the same result? Good grief. Now, I really hate RPI.

My understanding is that a road series against Florida helps SOS, which is a factor of RPI, but it is not weighted any differently than any other road game.

[Ralph]

Wow, I've thought RPI was overrated for many many years now, and that includes roundball and football as well. But I didn't even realize it was that flawed. I thought they DID somewhat factor in things like good wins on road. So, you're telling me they count winning 2 of 3 at Florida the same as if we had played an early season weekend series on the road against say, Tulsa, with the same result? Good grief. Now, I really hate RPI.

Yep. When we beat fla 2/3 on the road RPI doesn't factor WHO you win/lose to or what that teams RPI was. It factors your SOS and wins/losses.

Our RPI jumped this week bc we won and Fla's SOS is #2, not bc their RPI was #1.

[smootness]

Yep. When we beat fla 2/3 on the road RPI doesn't factor WHO you win/lose to or what that teams RPI was. It factors your SOS and wins/losses.

Our RPI jumped this week bc we won and Fla's SOS is #2, not bc their RPI was #1.

Eh. That's correct and incorrect. No, RPI doesn't technically factor in Florida's RPI. But it does heavily factor in their winning percentage and their opponents' winning percentage...so it basically does factor in their RPI.

[Ralph]

Eh. That's correct and incorrect. No, RPI doesn't technically factor in Florida's RPI. But it does heavily factor in their winning percentage and their opponents' winning percentage...so it basically does factor in their RPI.

I get your point but no where does a team get credit for beating a high RPI team based solely on their RPI. SOS and win% would obviously be high if their RPI is high.

A team will get more mileage out of beating the #25 RPI team with #2 SOS, vs beating #2 RPI team with #25 SOS.

[maroonmania]

Someone on here told me that 50% of RPI is your opponents won/loss record. That makes RPI a hugely flawed stat because it hurts teams that play West Coast teams.

Baseball is a little different in the South than it is out West and also in the North as well. One major difference is we have some really, really BAD teams here. I'm not talking about Sun Belt or C-USA teams at all. There are definitely plenty of GOOD teams. The BAD teams I am talking about are SWAC teams. Most every SEC team plays a few SWAC teams and some UT-Martin's which are basically almost programmed wins and glorified scrimmages.

Out West, you have a lot more quality mid majors that are more comparable to a South Alabama or USM. And people like USC play them pretty much every week. Cal State Fullerton, Long Beach State, Cal Santa Barbara among others can more than hold their own with any PAC 12 team. Basically you have a smaller gap between the mid majors and the PAC 12 teams. And they ALL want to beat the USC's and UCLA's every year- which means those teams are basically someone's World Series every Tuesday. And there aren't very many SWAC-esque teams out there.


The result is everyone out there beats each other up, which results in about a 30 win season if your team is good. However the problem is RPI perceives that as your opponent being "down" which thus hurts your RPI. If USC or Oregon played SEMO State this year in a three game series I would imagine that the PAC 12 team would likely win a three game series because the PAC 12 team is the better team. RPI says that SEMO State is "better" based on their math though.

The good news is the selection committee doesn't base EVERYTHING on RPI. It's only one component of the equation.

Now, I do like playing PAC 12 teams because they do play a different style and we may very well see that style in a regional and have to deal with it. I do think that one thing we could do is maybe schedule an ACC team as well as a PAC 12 team. That might help us keep from getting burned by RPI when our PAC 12 opponent inevitably can't hold up their end of the deal.

Or they could do like Shotgun said and just actually pick the best teams.

The formula is basically:

The RPI is calculated by adding three parts.

Part I (25% of the formula): Team winning percentage. For the 2005 season, the NCAA added a bonus/penalty system, where each home win or road loss get multiplied by 0.6 in the winning percentage calculation. A home loss or road win is multiplied by 1.4. Neutral games count as 1.0. More on the effect of these changes can be found here.

Part II (50%): Average opponents’ winning percentage. To calculate this, you must calculate each opponent’s winning percentage individually and average those figures. This is NOT calculated from the opponents’ combined record. Games involving the team for whom we are calculating the RPI are ignored.

Part III (25%): Average opponents’ opponents’ winning percentage: Basically taking all…

So the flaws are IMO, at the risk of offending the RPI defenders on the board (and there are quite a few), that, first of all the weighting it totally arbitrary. Somebody subjectively decided to break it down as a 25, 50, 25 weighting. It could have been done any number of ways. Secondly, its based solely on winning percentage for all 3 factors. Nothing in the formula takes into account who the winning percentages were built against. I'm sure people will say it eventually all washes out but if you wanted to be smart the best strategy in scheduling would appear to play opponents that generally play weak competition whose competition also plays weak competition. After all its all about your opponent's winning percentage (the heaviest weighting) first, and your winning percentage and your opponent's opponent's winning percentage next. Doesn't really factor in who the quality of who those winning percentages are built up against that I can tell. My complaint is that while its a worthwhile tool its become relied on way too heavily as the measuring stick for post-season selection and seeding.

[Taog Redloh]

Any poll based strictly on W/L will always be flawed, because there is no baseline for it. Theoretically that would be the polls, but that is flawed too, as it is subjective. So, at this point in time, when we do not have a committee-based ranking system like football, we just take the best we can get, and that is a mix between the polls and RPI. RPI is about as good as we'll ever have I guess to filter through the BS. You can definitely game it though, which is why it is very important IMO to add some type of subjective intuition to the sheer numbers (sheer numbers is all RPI really even is).

I'd like to see some advanced stats tied into college baseball. Sort of like what Bill Connelly does for football.

[Jarius]

Any poll that has Ole Miss ahead of MSU in baseball right now is pretty damn flawed.

[smootness]

I get your point but no where does a team get credit for beating a high RPI team based solely on their RPI. SOS and win% would obviously be high if their RPI is high.

A team will get more mileage out of beating the #25 RPI team with #2 SOS, vs beating #2 RPI team with #25 SOS.

Not really. 50% is opponents' winning %, 25% is opponents' opponents' winning %. So the #2 RPI team would obviously have a great record, which technically counts more than their SOS.

I'm fine with the formula not counting the team's actual RPI, that wouldn't make a lot of sense.

[Ralph]

Not really. 50% is opponents' winning %, 25% is opponents' opponents' winning %. So the #2 RPI team would obviously have a great record, which technically counts more than their SOS.

I'm fine with the formula not counting the team's actual RPI, that wouldn't make a lot of sense.

Just to be clear you're saying 75% of RPI is based on opponents' or opponents' opponents' winning %, so therefore the #2 RPI team will obviously have a great record because their opponents won against other teams? This doesn't make sense. High RPI teams often don't have the best records nationwide bc as you said 75% is based on what some other team does while not playing against you. This is exactly why OM's RPI is higher than ours yet we beat them 2/3 H2H.

And Im not advocating for including the RPI in the formula at all so not sure who you're arguing with. I was confirming what another poster said using examples. You actually pointed out the flaw in your response.

[HSVDawg]

I agree with the West Coast teams argument, but I think its hedged a bit by the fact that the 25% component for the opponent's opponents winning percentage should be much higher than for the SWAC / UT Martins / etc. for the reasons that you stated. It doesn't totally balance the scale, but it is worth mentioning.

At the end of the day though, we choose to play those bad RPI teams. It's been stated before, but we can easily schedule SE Missouri State instead of UT Martin, play more games against ULM / Troy / Austin Peay / MTSU / Tulane / La Tech, etc. and ditch the SWAC teams, UT Martins, Nicholls States, and UMASS Lowells. Like you said, there are plenty of good mid majors to choose from. Historically, we know that SWAC schools and UT Martin for sure going to be sub-250 RPI teams year in and year out. There's no reason to schedule more than one or two games against them. We did a great job of this in 2013 when we really didn't have a "sexy" nonconference schedule, but we played a lot of solid, top 150 mid majors (several of which made the tourney) and didn't have many games at all against truly bad teams. And we were rewarded with a solid nonconference SOS and got to host a regional over Arkansas who didn't have those things on their resume.

And I'm not an "RPI defender" by any means, but it must be acknowledged that it is weighed heavily by the NCAA in the selection process. Therefore, its in our best interest to manipulate it to our advantage whenever possible.

[Taog Redloh]

And I'm not an "RPI defender" by any means, but it must be acknowledged that it is weighed heavily by the NCAA in the selection process. Therefore, its in our best interest to manipulate it to our advantage whenever possible.
I'd rather spend our time lobbying to the committee about how stupid RPI is, so we can schedule fan-friendly games, rather than try to game the system. By fan-friendly I mean playing more games against SWAC teams in Pearl or Biloxi.

[maroonmania]

I agree with the West Coast teams argument, but I think its hedged a bit by the fact that the 25% component for the opponent's opponents winning percentage should be much higher than for the SWAC / UT Martins / etc. for the reasons that you stated. It doesn't totally balance the scale, but it is worth mentioning.

At the end of the day though, we choose to play those bad RPI teams. It's been stated before, but we can easily schedule SE Missouri State instead of UT Martin, play more games against ULM / Troy / Austin Peay / MTSU / Tulane / La Tech, etc. and ditch the SWAC teams, UT Martins, Nicholls States, and UMASS Lowells. Like you said, there are plenty of good mid majors to choose from. Historically, we know that SWAC schools and UT Martin for sure going to be sub-250 RPI teams year in and year out. There's no reason to schedule more than one or two games against them. We did a great job of this in 2013 when we really didn't have a "sexy" nonconference schedule, but we played a lot of solid, top 150 mid majors (several of which made the tourney) and didn't have many games at all against truly bad teams. And we were rewarded with a solid nonconference SOS and got to host a regional over Arkansas who didn't have those things on their resume.

And I'm not an "RPI defender" by any means, but it must be acknowledged that it is weighed heavily by the NCAA in the selection process. Therefore, its in our best interest to manipulate it to our advantage whenever possible.

Hasn't bothered UNM. They've played UT-Martin as well plus a 3 game set against Grambling. Also played another 3 game set against FIU (in the 160s). Our schedule this year SHOULD have been very solid RPI wise. A couple of issues have factored in. First, we SHOULD have not lost at home to EKU and ORU. If we had taken care of business there most of this is a moot point. Second, some of our top end competition (UCLA, USC, OK, OR) have not been as good as expected but all of their RPIs are on the uptick because of their conference play so that should help. FAU continues to be strong so they are helping as well. What we can't have happen is something like a loss to Memphis tomorrow night. We've already stumbled against a crap team this year (EKU) and if we want to keep the RPI high it better not happen again.

[HSVDawg]

Hasn't bothered UNM. They've played UT-Martin as well plus a 3 game set against Grambling. Also played another 3 game set against FIU (in the 160s). Our schedule this year SHOULD have been very solid RPI wise. A couple of issues have factored in. First, we SHOULD have not lost at home to EKU and ORU. If we had taken care of business there most of this is a moot point. Second, some of our top end competition (UCLA, USC, OK, OR) have not been as good as expected but all of their RPIs are on the uptick because of their conference play so that should help. FAU continues to be strong so they are helping as well. What we can't have happen is something like a loss to Memphis tomorrow night. We've already stumbled against a crap team this year (EKU) and if we want to keep the RPI high it better not happen again.

It hasn't affected OM because they have played 7 nonconference games against top 50 teams compared to our 4, and have only played 5 games against teams outside the Top 200 compared to our 7 (plus agame against #193 EKU). OM will also add their 8th game against a top 50 team this week, where we will add our 8th against a team outside the top 200. Part of that is luck in that Louisville was as good as they were expected to be whereas the west coast teams and Oklahoma weren't for us. However, another part is that OM doesn't have as many RPI killer games on the schedule as we do. Even with their Grambling series, they didn't have as much exposure as we did. That is why their nonconference SOS is 33 and ours is 147.

Also, your point about our schedule not being solid RPI wise because of us losing to EKU and ORU doesn't make a lot of sense. Whether or not we won those games doesn't really impact SOS as a whole, it just affects our winning percentage component of the RPI. Actually, if we had won those games, our SOS might actually be marginally worse because it would decrease the winning percentage of two of our opponents. All that being said, the Oral Roberts game is being lumped in with EKU, but it wasn't nearly as bad of a loss. And you are right that a repeat of EKU does not need to happen again.

[smootness]

Just to be clear you're saying 75% of RPI is based on opponents' or opponents' opponents' winning %, so therefore the #2 RPI team will obviously have a great record because their opponents won against other teams? This doesn't make sense. High RPI teams often don't have the best records nationwide bc as you said 75% is based on what some other team does while not playing against you. This is exactly why OM's RPI is higher than ours yet we beat them 2/3 H2H.

And Im not advocating for including the RPI in the formula at all so not sure who you're arguing with. I was confirming what another poster said using examples. You actually pointed out the flaw in your response.

You compared a team with a #25 RPI and the #2 SOS with a team with a #2 RPI and the #25 SOS. In that case, beating the team that is 2nd in the RPI is still going to help you more. Because they obviously have a much better record themselves than the team with the #25 RPI, and opponent winning % counts more than opponents' opponent winning %, which is where their SOS comes into play.

And yes, if the #2 RPI team has the #25 SOS, they will absolutely have a phenomenal record. And definitely a much better record than the #25 team with a #2 SOS.

[HSVDawg]

I'd rather spend our time lobbying to the committee about how stupid RPI is, so we can schedule fan-friendly games, rather than try to game the system. By fan-friendly I mean playing more games against SWAC teams in Pearl or Biloxi.

I'd certainly be open to more neutral site games against teams like that. If we're going to take the hit for playing them, at least get the full 1.0 multiplier on the W as opposed to the 0.67 multiplier for a home win. However, I think lobbying to the committee is relatively pointless. We need to focus on things we can control, and chief amongst those things is the nonconference schedule.

[maroonmania]

It hasn't affected OM because they have played 7 nonconference games against top 50 teams compared to our 4, and have only played 5 games against teams outside the Top 200 compared to our 7 (plus agame against #193 EKU).

The EKU loss at home was the one really bad loss we have. ORU is still a Top 100 team. In fact, if you move to the top 100 RPI teams, we have actually played 10 of those in non-conference and the Bears have still played only the same 7. I mean who would have guessed preseason that SE Mo State, Coastal Carolina and USM would be top 50 teams but UCLA, USC and Oregon wouldn't (but they are Top 100). Sometimes its a crapshoot. But as stated if you take the full Top 100 we've actually had several more of those than UNM has. Difference, in part, though is still that the Bears have taken care of business against weaker teams and have won every game played outside the Top 50 while we have dropped several of those.

[Ralph]

You compared a team with a #25 RPI and the #2 SOS with a team with a #2 RPI and the #25 SOS. In that case, beating the team that is 2nd in the RPI is still going to help you more. Because they obviously have a much better record themselves than the team with the #25 RPI, and opponent winning % counts more than opponents' opponent winning %, which is where their SOS comes into play.

And yes, if the #2 RPI team has the #25 SOS, they will absolutely have a phenomenal record. And definitely a much better record than the #25 team with a #2 SOS.

So you think the higher the RPI the better the record will be? SMH. You're confusing a theoretical example and make way, way too many assumptions in these statements. Curious how you explain these situations...

Maryland- 18-15, RPI 43, SOS 16, 54.5%WP
Arizona - 20-12, RPI 64, SOS 73, 62.5%WP

These 2 teams are 21 RPI spots apart and see nothing obvious about the higher RPI team having a better record.

Another...

Bryant - 20-6, RPI 29, SOS 165, 77% WP
FSU - 22-8, RPI 5, SOS 1, 73% WP

[DawgPoundtheRock]

My whole reaction to this thread and RPI is, huh? WTF? Just win and everything will be okay.

[chef dixon]

Road games matter: Minnesota is 17-10, highest RPI team they've played is 36 one time and they lost, and they are ranked 19th in RPI. But 22 of their 27 games have been away from home. They also have some pretty ugly losses mixed in.

[smootness]

So you think the higher the RPI the better the record will be? SMH. You're confusing a theoretical example and make way, way too many assumptions in these statements. Curious how you explain these situations...

Maryland- 18-15, RPI 43, SOS 16, 54.5%WP
Arizona - 20-12, RPI 64, SOS 73, 62.5%WP

These 2 teams are 21 RPI spots apart and see nothing obvious about the higher RPI team having a better record.

Another...

Bryant - 20-6, RPI 29, SOS 165, 77% WP
FSU - 22-8, RPI 5, SOS 1, 73% WP

What on earth gave you the idea that I think RPI is based 100% on your record? Of course it isn't. I was comparing the two examples that you yourself used. You said it would be better to beat the #25 RPI team with a #2 SOS than the #2 RPI team with a #25 SOS, and that is incorrect. It is definitely better to beat the #2 RPI team in that case. Why? Because the #2 RPI team will absolutely have a much better record in that scenario than the #25 RPI team, and your opponents' record counts more than your opponents' opponents' record. That is simply a fact.

Use the exact same examples there to logically assess your prior example. If SOS is so important that a team with a worse record can have a much better RPI based on SOS, then it also follows that if you have a worse SOS than another team but are much higher in the RPI, you have a much better record than they do.

[Ralph]

What on earth gave you the idea that I think RPI is based 100% on your record? Of course it isn't. I was comparing the two examples that you yourself used. You said it would be better to beat the #25 RPI team with a #2 SOS than the #2 RPI team with a #25 SOS, and that is incorrect. It is definitely better to beat the #2 RPI team in that case. Why? Because the #2 RPI team will absolutely have a much better record in that scenario than the #25 RPI team, and your opponents' record counts more than your opponents' opponents' record. That is simply a fact.

Simply not true at all. A team with a #2 RPI can have the same or worse record as the #25 RPI team. See Bryant/FSU example. FSU's RPI is #5, but Bryant's record is 4% better while being RPI #29. Why is FSU's RPI higher than Bryant's? Bc of FSU's SOS.


Use the exact same examples there to logically assess your prior example. If SOS is so important that a team with a worse record can have a much better RPI based on SOS, then it also follows that if you have a worse SOS than another team but are much higher in the RPI, you have a much better record than they do.

This is obvious. What you're not considering, using the Bryant/FSU example again, is the difference between FSU's win% and Bryant's win% is only 4% in Bryant's favor. While the difference in their SOS's is more like 15-20% in FSU's favor.