Unlucky with our non-conference cupcakes...

[Jay Bee]

We can all agree MU probably played one of our hardest non-conference schedule in many years.  However, according to espn's RPI, MU had the 162nd hardest non-conference SOS out of 351 teams.

Unfortunately, five of our cupcakes turned out to be really terrible which dragged it down... Grambling (349), NH (344), IUPUI (306), Samford (307), and Ball State (285).  Last year, our worst was UMBC (301) and fifth worst was Savannah State (167).

I have to imagine these five awful cupcakes bring our RPI down at least 10 spots.

Anyone smart (Henry Sugar?) easily able to extrapolate or guess what MU's rpi would be if these five teams would have been in the 200-250 range?

Disclaimers...
- I understand we're still 4 straight wins from being near the bubble... I'm just curious what effect this has.  
- I realize it's hard to know in advance who the truly worst cupcakes in D1 will be.
- I prefer kenpom, but it seems RPI is still what drives the media/committee.

It's not an issue of being "200-250" vs "300". It has to do with the team's win loss record.

The impact is larger than 10 spots. MU's schedule hurt them.

The sad thing is MU played more competitive games than most.. but the RPI calculation can't tell that.

Fortunately, people can. Will they? We may find out.

Indeed, the RPI calculation results in a gloomier view of Marquette than a reasonable system or normal person would.

[bilsu]

Make me the AD and I would never schedule a team that is expected to be ranked 300 or higher. Given that these games are almost always played at home, we should expected to beat teams in the 200-299 range. While they are not great for RPI they are not as detrimental as a 300+ team.

[Coleman]

Make me the AD and I would never schedule a team that is expected to be ranked 300 or higher. Given that these games are almost always played at home, we should expected to beat teams in the 200-299 range. While they are not great for RPI they are not as detrimental as a 300+ team.

Easier said than done. These ADs don't have a crystal ball

[bilsu]

Easier said than done. These ADs don't have a crystal ball

I said expected to be ranked 300+. I realize a team due to injuries, defections, etc. can end up being 300+. There is simply no reason to schedule a Grambling, who had no chance of being better than 300.

[Jay Bee]

Stupidity of the RPI, part 9,494

As things stood yesterday (some RPI's have changed slightly)...

MU at around #77 RPI...

Take out MU's worst 4 opponents this season:
335    New Hampshire
289    Ball State
320    IUPUI
349    Grambling

Replace with 4 different crappy teams:
305   Texas Southern
322   Georgia Southern
299   Austin Peay
284   Sacramento St

...and MU would be sitting at approximately #60.

Or, replace with 4 alternative crappy teams:
273   Portland St.
266   SE Missouri St
251   Alabama St.
255   Norfolk St.

...and MU would be sitting at approximately #50.

--------
"SOS", which is a name of something the RPI calculates, but is NOT a reasonable measure of strength of schedule, is not going to be good for MU despite a very challenging schedule.

Adjusted winning percentage (AWP) is currently sitting at .5714 - which is bad.

I believe Nova was got in last year with a .5903, but they had a stronger SOS.

For MU, a couple of home losses would brutalize their AWP and put their RPI in an awful position. Strictly from an RPI calculation perspective, if you have to lose games the rest of the way, you'd want them to be on the road instead of at home.

3-1 at home, 2-1 on the road and 1-1 in BE tourney puts MU at .5957... not good, but they'd have a prayer.
4-0 at home, 1-2 on the road and 1-1 in BE tourney puts MU at .6015
2-2 at home, 3-0 on the road and 1-1 in BE tourney puts MU at .5906

Win the next two and MU is fully in the bubble discussion.

[ChicosBailBonds]

It's not an issue of being "200-250" vs "300". It has to do with the team's win loss record.

The impact is larger than 10 spots. MU's schedule hurt them.

The sad thing is MU played more competitive games than most.. but the RPI calculation can't tell that.

Fortunately, people can. Will they? We may find out.

Indeed, the RPI calculation results in a gloomier view of Marquette than a reasonable system or normal person would.

Everyone knows the calculation going into it.  Schedule accordingly.  With the team we had coming back this year, picked to win the Big East, we should have had better cupcakes.

As for the RPI, its applied equally to everyone.  The reality is that the RPI tells you the same thing your eyeballs do.  We haven't lost to anyone bad, we haven't beaten anyone good.  We are what we are.  Opportunities in front of us, but the RPI for MU illustrates nicely where we are.  A team that lost to Butler, a team that has only three road wins and a team that has played 3 teams that are above 300 in the RPI.  Including the worst team in Division I the last two years.  Cupcakes  can still be cupcakes without making them 5 guys that play rec ball.

The NCAA selection commitee has said time and time again that they will reward teams that play good competition.  For some reason we did a really nice job of that with half of our non conference schedule, but we went way overboard on the other ones to the weak side and didn't have to.  Scheduling is tough, but how we have Grambling on the schedule as often as we do is a head scratcher. Games like that come back to bite you in the heine if you fail to win against a few good teams, which is our situation today.

[ChicosBailBonds]

Stupidity of the RPI, part 9,494

As things stood yesterday (some RPI's have changed slightly)...

MU at around #77 RPI...

Take out MU's worst 4 opponents this season:
335    New Hampshire
289    Ball State
320    IUPUI
349    Grambling

Replace with 4 different crappy teams:
305   Texas Southern
322   Georgia Southern
299   Austin Peay
284   Sacramento St

...and MU would be sitting at approximately #60.

Or, replace with 4 alternative crappy teams:
273   Portland St.
266   SE Missouri St
251   Alabama St.
255   Norfolk St.

...and MU would be sitting at approximately #50.

--------
"SOS", which is a name of something the RPI calculates, but is NOT a reasonable measure of strength of schedule, is not going to be good for MU despite a very challenging schedule.

Adjusted winning percentage (AWP) is currently sitting at .5714 - which is bad.

I believe Nova was got in last year with a .5903, but they had a stronger SOS.

For MU, a couple of home losses would brutalize their AWP and put their RPI in an awful position. Strictly from an RPI calculation perspective, if you have to lose games the rest of the way, you'd want them to be on the road instead of at home.

3-1 at home, 2-1 on the road and 1-1 in BE tourney puts MU at .5957... not good, but they'd have a prayer.
4-0 at home, 1-2 on the road and 1-1 in BE tourney puts MU at .6015
2-2 at home, 3-0 on the road and 1-1 in BE tourney puts MU at .5906

Win the next two and MU is fully in the bubble discussion.

Think about what you just said....merely replace number 349 with number 251...that's nearly 100 spots different.  That is a BIG improvement.  You make it sound like they merely had to swap out a bunch of teams in the 330's with ones in the 315's.   It does make a difference....you play that many bottom feeders you best not lose in every one of your other games against quality teams, which we did. 

Hate on the RPI all you want, it's the NCAA's baby.  They created it, they use it, they will continue to use it.  Everyone knows the formula, use it to your advantage and schedule accordingly.  Crystal balls don't need to be in play, Grambling has no business being on the schedule. 

[Jay Bee]

Think about what you just said....merely replace number 349 with number 251...that's nearly 100 spots different.  That is a BIG improvement.  You make it sound like they merely had to swap out a bunch of teams in the 330's with ones in the 315's.   It does make a difference....you play that many bottom feeders you best not lose in every one of your other games against quality teams, which we did.  

Hate on the RPI all you want, it's the NCAA's baby.  They created it, they use it, they will continue to use it.  Everyone knows the formula, use it to your advantage and schedule accordingly.  Crystal balls don't need to be in play, Grambling has no business being on the schedule.  

The "spots" don't matter - the opp's win-losses and their opp's win-losses matter. You and most people are confused about the specifics of the calculations.

MU is absolutely at fault for scheduling as they did this season. It was a poor job of scheduling relative to the RPI. Their strength of schedule as any reasonable human would describe it was very good - but people do pay attention to the RPI (how much? we don't know and it depends on the individual) and therefore I believe schools should do the best they can to schedule optimally from an RPI perspective -- or at a minimum avoid absolutely awful teams who are projected by most to win 5 or fewer games.

MU failed to do so this season. That's unfortunate.

edit: the original point on the impact to RPI is this - the RPI is stupid. Whether MU plays Grambling or Alabama State at home.. or swaps out the four teams mentioned with another four crappy teams... the schedule is still the same --> they are games MU should easily win and have pose no significant challenge. However, the RPI cares a lot about it. It's dumb. But MU didn't play along with the dumb rules of the game this year and they're paying for it (via their RPI ranking... how much it matters - we don't know - that's up to a small gathering of individuals).

[wardle2wade]

Thanks for doing the work, Jay Bee.  That's what i was afraid of and was the reason for my original post... the cupcakes being truly awful cost MU about 15-20 rpi spots as it stands today. Disappointing and hopefully they'll try to be more cognizant of it in the future... may cost us a bid.

[Jay Bee]

Thanks for doing the work, Jay Bee.  That's what i was afraid of and was the reason for my original post... the cupcakes being truly awful cost MU about 15-20 rpi spots as it stands today. Disappointing and hopefully they'll try to be more cognizant of it in the future... may cost us a bid.

My pleasure... getting back to what chicos said.. and this is key - because (1) many websites (including the 'big named' ones and others who 'specialize' in RPI) get calcs wrong.. and (2) many people assume playing a team with a better RPI means it's a better game for YOUR RPI.. not true.

We'll do one example for chicos sake...

Assumptions:
MU is around #75 in RPI with a score of .5569.
An increase of .5569 by .0200 (i.e., to .5769) would put them at around #57 today.

Let's take just one game:
Ball State, #289.. and switch out Portland State (not Portland.. but Portland STATE) who is team AFTER Ball State at #290.. what would be the impact to MU's RPI currently had they played Portland State - a team with a worse RPI than Ball State - instead of the Muncie boys?

Answer: an increase of approximately .0063.

This results from the net of (a) Portland State having a better true win-loss record (not adjusted for location of games) compared to Ball State, partially offset by (b) Ball State's opponents' having a better true win-loss record than Portland State.

So again.. it's not necessarily that Marquette played teams with RPI's of X or greater than some magical threshold... it's the opponents' true win-loss record (mostly) and that of their opponents (to a lesser degree).

The weighting of the components and the adjusting winning percentage factor being used only for an individual team's RPI score calculation cause the many issues you'll run into if you start saying, "well we should have played teams with an RPI of X instead of Y." It may be true, it may not.

Remember, when someone talks about a team's "strength of schedule" when they're talking about the RPI, they are NOT including any consideration as to where any of the games were played. AND strength of schedule does not directly consider the RPI's of other teams... only certain of their components, differently calculated (win-loss) and weighted differently.

Stupid stuff, but it's true that MU's schedule made this situation (having an unimpressive SOS) likely. MU's RPI "SOS" will rise significantly over the next few weeks due to who they are playing vs. who others are playing, but it still probably won't look a lot better than top 50 (from probably mid-70s now?)

[ChicosBailBonds]

The "spots" don't matter - the opp's win-losses and their opp's win-losses matter. You and most people are confused about the specifics of the calculations.

MU is absolutely at fault for scheduling as they did this season. It was a poor job of scheduling relative to the RPI. Their strength of schedule as any reasonable human would describe it was very good - but people do pay attention to the RPI (how much? we don't know and it depends on the individual) and therefore I believe schools should do the best they can to schedule optimally from an RPI perspective -- or at a minimum avoid absolutely awful teams who are projected by most to win 5 or fewer games.

MU failed to do so this season. That's unfortunate.

edit: the original point on the impact to RPI is this - the RPI is stupid. Whether MU plays Grambling or Alabama State at home.. or swaps out the four teams mentioned with another four crappy teams... the schedule is still the same --> they are games MU should easily win and have pose no significant challenge. However, the RPI cares a lot about it. It's dumb. But MU didn't play along with the dumb rules of the game this year and they're paying for it (via their RPI ranking... how much it matters - we don't know - that's up to a small gathering of individuals).

That was my point Jay Bee, everyone knows the rules.  You may think they are dumb, doesn't matter...that's the criteria.   

Now, if people think the RPI doesn't matter as so  many people here every year opine, it won't matter one bit.  Seems to me that some are a little nervous it matters a bit, which is why they claim it is dumb, etc.

The RPI is the NCAAs creation, they are going to use it.  Everyone knows how it works.  For MU not to play by the "dumb rules" is MU's fault, not the dumbness of the RPI.  Coaches rig the schedule every year on every team to try and get a magical 20 wins, try to get a scalp here or there at home, prop up the SOS, etc.  It's not that hard.  Grambling, as an example, should not be on MU's schedule.  Presbyterian, who finds there way on our schedule every once in a while, should not be on our schedule.  UMBC, NJIT, Longwood, those types of teams you just have to avoid...or at least don't schedule more than one of them.

We are hurt this year partly because Fullerton sucks donkey balls and we didn't have a choice in that one, but some of the others is simply not playing the game right.

[Jay Bee]

That was my point Jay Bee, everyone knows the rules.  You may think they are dumb, doesn't matter...that's the criteria.   

Now, if people think the RPI doesn't matter as so  many people here every year opine, it won't matter one bit.  Seems to me that some are a little nervous it matters a bit, which is why they claim it is dumb, etc.

The RPI is the NCAAs creation, they are going to use it.  Everyone knows how it works.  For MU not to play by the "dumb rules" is MU's fault, not the dumbness of the RPI.  Coaches rig the schedule every year on every team to try and get a magical 20 wins, try to get a scalp here or there at home, prop up the SOS, etc.  It's not that hard.  Grambling, as an example, should not be on MU's schedule.  Presbyterian, who finds there way on our schedule every once in a while, should not be on our schedule.  UMBC, NJIT, Longwood, those types of teams you just have to avoid...or at least don't schedule more than one of them.

We are hurt this year partly because Fullerton sucks donkey balls and we didn't have a choice in that one, but some of the others is simply not playing the game right.

No - you don't understand the rules and the calculations. Having UMBC or NJIT on this year's schedule instead of some others would have helped the current SOS and RPI. It's a very specific analysis that must be performed. Your high-level commentary on MU's scheduling is fine, but you're way off the mark when you try to speak specifics.

[ChicosBailBonds]

Remember, when someone talks about a team's "strength of schedule" when they're talking about the RPI, they are NOT including any consideration as to where any of the games were played. AND strength of schedule does not directly consider the RPI's of other teams... only certain of their components, differently calculated (win-loss) and weighted differently.

But you are ignoring that it does, indirectly, consider where the games are played because a team earns more points for a road win vs a home win.  It impacts, indirectly, .25 of the score where a home win is worth .6 and a road win worth 1.4 (inverse for losses).  That flows through into the results. 

I think I'm going to get Jerry on CS and have a little educational seminar for everyone.  Palm had this to say last year about the RPI vs other metrics.  "When asked if it’s time to replace the RPI with other metrics, Palm said: “So what are their formulas? Really, what are the exact formulas? They are secret. So we don’t have any idea how these teams are being judged other than that we take their word for what is going into it. At least the RPI can be explained.”

Again, the RPI is flawed, every system is flawed.  The difference is, the RPI is the NCAAs flawed baby and not someone else's.  That's a big reason why it is used the way it is, plus the comments of Palm above.  It's the reason why the BCS had some of their computer formulas thrown out over the years, because the creators of them refused to show exactly how the sausage was made and there was too many conspiracy goofballs out there.  I happen to love Massey's stuff, Pomeroy, Sagarin, etc...but each is different to the point I can find one team that is 20th in one of those ratings and as high as 65 (same team) in another.  So which one do you go with?

For a general overview...http://www.collegerpi.com/rpifaq.html   but I've worked with Jerry enough in the last 10 years I think we'll try and do a little Q&A

[Jay Bee]

But you are ignoring that it does, indirectly, consider where the games are played because a team earns more points for a road win vs a home win.  It impacts, indirectly, .25 of the score where a home win is worth .6 and a road win worth 1.4 (inverse for losses).  That flows through into the results. 

NO, the RPI's Strength of Schedule doesn't at all consider where the games are played.

You again are completely incorrect.

You don't need the Purdue guy, you need to put pen to paper and try to understand things yourself.

[Jay Bee]

I happen to love Massey's stuff, Pomeroy, Sagarin, etc...but each is different to the point I can find one team that is 20th in one of those ratings and as high as 65 (same team) in another.  So which one do you go with?

...and back to another one of my points - even many of the 'experts' do not understand the sensitivities of RPI calculations. Go look around at RPI per a number of sources - they'll all be a bit different. The one that matters is the NCAA's.

I can tell you where certain RPI-calculators have things wrong - some are easier to decipher than others. But I assure you MANY have horrible errors in their calculations (including ESPN, bbstate, etc).

Brutal. Fortunately for them the masses (i.e., people like you) open up and eat whatever is being fed.

edit: changed easy to easier

[Lennys Tap]

Chicos is an email buddy of Jerry, but has been tight with his sister Rosie for more than 30 years.

[ChicosBailBonds]

No - you don't understand the rules and the calculations. Having UMBC or NJIT on this year's schedule instead of some others would have helped the current SOS and RPI. It's a very specific analysis that must be performed. Your high-level commentary on MU's scheduling is fine, but you're way off the mark when you try to speak specifics.

I understand them just fine...it all depends for WHOM you were to switch out those games.  So me saying not having NJIT or Grambling on the schedule would be beneficial to us is not inaccurate at all.  It's all about what the average benefit is going to be based on the pool of available cupcake teams for you to schedule.  You are making your conclusion based on current data and actual games played to say swap out NJIT with team X and their is improvement.  Well duh, of course there is if you know what the results are and you can go backward to cherry pick what teams have an acretive impact and which ones are dilutive. Wouldn't that be nice if we could all do that....I'll just go out and pick today's lottery numbers having them already in my pocket. 

No one when scheduling knows what the results are going to be, but that is where common sense has to kick in.  More often than not, scheduling NJIT, Grambling, etc is going to HURT you, not BENEFIT you in the RPI.  Yes, you can go through all 351 teams and find the 18 scenarios or whatever it is where this isn't the case.  This is where you play the odds, and the odds say you don't schedule those teams because more often than not you are going to get bit in the arse.  Or a better way to say it, of the "cupcake variety teams" that are out there, in how many cases would not scheduling Grambling benefit you vs how many would you be benefited by? 


But since anything I say you will rebut regardless, I'll have Jerry Palm available and you can duke it out with him...that's if I can get him in the next few weeks.  Might have to wait until after the NCAA tournament due to his current schedule.  He's a good dude, we've done some work in the past on the television side.

[ChicosBailBonds]

Chicos is an email buddy of Jerry, but has been tight with his sister Rosie for more than 30 years.

LOL.  I laughed.   Actually the relationship goes on the television side when I used to run Mega March Madness and predictive game we were putting together, extends to CBS, etc. 

More than an email buddy, but if you want to get your digs in, that's fine. 

[Jay Bee]

But since anything I say you will rebut regardless, I'll have Jerry Palm available and you can duke it out with him...that's if I can get him in the next few weeks.  Might have to wait until after the NCAA tournament due to his current schedule.  He's a good dude, we've done some work in the past on the television side.

What have I said that Palm would not agree with? Zero (or else he'd be wrong).

I have no questions for him - you're the one who is confused.

Again, "strength of schedule" per the RPI ignores where games are played. Completely ignores it. Go ask your bud that and if he doesn't say, "yes - Jay Bee is correct and you are wrong" then let's talk.

[willie warrior]

LOL.  I laughed.   Actually the relationship goes on the television side when I used to run Mega March Madness and predictive game we were putting together, extends to CBS, etc. 

More than an email buddy, but if you want to get your digs in, that's fine. 

That may be true Lenny, but we hear that you are tight with Jerry's twin brothers: Sweaty and Greasy.

[Texas Western]

I think the cupcake portion of the schedule would be better served with teams from the MAC, Horizon and Missouri Valley. The  mid to lower half of those leagues have much better RPI's. Many of these teams might also attract a few more fans do to their local nature such as Loyola, Northern Illinois who could run some student fan buses up. Also they would be more traditional such as Detroit, who we had good games with in the glory days.

[Avenue Commons]

It's the Butler L and the Nova L that could doom us. Pick up one if those, especially Nova, and a lot of these concerns drain away like so many Miller beers after a MU game.

[chapman]

I think the cupcake portion of the schedule would be better served with teams from the MAC, Horizon and Missouri Valley. The  mid to lower half of those leagues have much better RPI's. Many of these teams might also attract a few more fans do to their local nature such as Loyola, Northern Illinois who could run some student fan buses up. Also they would be more traditional such as Detroit, who we had good games with in the glory days.

Agree.  Won't get any of the MVC or half of the MAC teams to take straight buys, but the Horizon in particular always has a lot of teams that are good on the RPI - nine teams in our backyard, a good number take a few buy games, and yet we scheduled none.  I realize Nova gets a nice arrangement with the Philly Five, but their other non-conference games are all in their area and almost all decent to their RPI: Delaware, Towson, Rider, Mount St. Mary's, Lafayette.  I'd hope we thought of a similar strategy.  With attendance strong, more tournament shares than ever and more television money than ever, it would be worth the investment to bring in teams from the Horizon rather than the SWAC and others that are the cheapest of the cheap.

[Jay Bee]

I think the cupcake portion of the schedule would be better served with teams from the MAC, Horizon and Missouri Valley. The  mid to lower half of those leagues have much better RPI's. Many of these teams might also attract a few more fans do to their local nature such as Loyola, Northern Illinois who could run some student fan buses up. Also they would be more traditional such as Detroit, who we had good games with in the glory days.

Their RPI's don't necessarily matter, though. You want them to have good records. Bad teams that are "good" in their conference (i.e., win a lot of conference games) can help our RPI as compared to teams that struggle in "better" MM conferences.