Wednesday, January 25, 2017

Hard Schedules, Easy Schedules, And Florida State

What does it mean to say that a team has a hard or easy remaining schedule? In the simplest sense, we can measure how difficult the average remaining opponent is. But we can ask a different question: What makes a schedule harder or easier to deal with?

Oh, and why am I only projecting a 4 seed for Florida State when they're 18-2 with 9 RPI Top 50 wins and the BracketMatrix has them as the ACC favorite and a 2 seed?


The fact is that on Selection Sunday teams are judged on their resumes, but not in a purely objective way. There are biases driven because of the power of the RPI, and due to the fact that it is hard to take home/road into proper account. I've written many times about how the RPI process can be manipulated, with the most thorough piece being this one.

In short, you are rewarded for "quality wins", meaning RPI Top 50 wins, with special bonuses for wins over ranked teams, and with a win over a Top 10 team being a jewel in your crown. Meanwhile, you are punished for "bad losses", with a sliding scale from non-NCAA Tournament teams, RPI 51+, RPI 101+, et cetera.

The problem with the system is that by not taking home/road into account, home games are massively overvalued. Right now, Maryland sits 49th in the Sagarin PREDICTOR. And so despite being ranked in the human polls, a home game vs Maryland is equivalent in the Sagarin ratings to a road game vs 10-8 Lehigh, who currently sit 117th. Yet a win vs Maryland is a "quality win" and a loss is no sweat, while a win vs Lehigh means nothing at the same time that a loss to Lehigh would be seen as catastrophic.

So, what are you looking for in your schedule? You want a lot of home games versus teams that would make up "quality wins", and you want to avoid road games vs decent teams unlikely to make the NCAA Tournament. As I wrote after last year's brackets were announced, this provides a tremendous penalty and hurdle for mid-majors. But this exists on a smaller scale within major conferences as well. When you're in the ACC you want a lot of shots at home against teams like Louisville and North Carolina, and you sure as hell want to avoid road games in places like the Carrier Dome.

So what about that Florida State schedule remaining? On first glance it seems soft by ACC standards - nothing against Louisville, North Carolina, or Virginia. But instead, 7 of 11 remaining games are on the road, mostly against decent-but-not-likely-tourney teams like Syracuse, Georgia Tech, and Pittsburgh. At home they play mostly teams with no upside like NC State and Boston College. Their two remaining chances for Top 25 wins, Notre Dame and Duke, both come on the road.

In other words, as good as Florida State is, there is nearly a 50% chance that they fail to pick up another particularly impressive win in the regular season, while the odds are high that they lose two or three games against the likes of Georgia Tech and Pittsburgh.

For example, Florida St has beaten Virginia, Duke, Louisville, and Notre Dame in the last two weeks, yet according to Pomeroy they have just a 29% chance of escaping their next three games unscathed against Georgia Tech, Syracuse, and Miami. Under 30 percent despite being favored in every game! If Florida State loses two of three, the media is going to start asking questions about what has "gone wrong" with a team that just beat Virginia and Louisville falling to, say, Georgia Tech and Syracuse. Yet nothing will have changed other than a schedule less conducive to perceived success.

In contrast, check out this piece of Virginia's upcoming schedule.

Here we see teams that would be perceived as massive resume-building victories (Louisville and North Carolina) coming to Charlottesville, where the Cavaliers will likely be favored. And they only have one real "bad loss" land mine in NC State. They have already suffered one of those losses earlier this season (at Pittsburgh, on the road and in overtime, on January 4th).

In other words, Virginia's schedule is built for success, with two great chances for big wins, and few chances to pick up bad losses. Florida State's is the opposite, almost certainly set up to lead to a decline in NCAA Tournament seed. Florida State already had their resume-building stretch, beating Louisville, Duke, and Notre Dame at home, but those games are all behind them. Most likely, they're heading for a decline down the stretch.

And that is why, despite Florida State having a stronger resume than Virginia at the moment, I am projecting a stronger seed on Selection Sunday for the Cavaliers than the Seminoles.


Trevor Kraus said...

This is amazingly detailed and you do this much work on all 68 teams you predict into the field on every one of your BP68 posts? Also, was surprised to see Gonzaga as a 3-seed in your last bracket...hard to see them losing more than two games the rest of the way to Selection Sunday.

Jeff said...

Thanks. I'm not constantly breaking down every team, I just keep an eye on remaining schedules. It's easy at a glance to see whether a team has a helpful or difficult schedule down the stretch.

Chris said...

This is a fantastic breakdown. Curious if you think the ACC could eat itself, and be squeezed out of a number 1 seed come March?

Also, do you see the schedules becoming even more magnified when the ACC moves to 20 league games in 2019-2020?

Miss the days of smaller conferences where you played everyone home and away.

David said...

I agree with your larger premise here. For a tournament hopeful (or a 1-2 seed hopeful, the distinction isn't really important here)road games against teams in the 70-150ish range suck, because they are often tough games, get overlooked by teams/coaches, and can only really hurt you. No argument there.

But there is one point that you consistently make that I'm not prepared to accept. Saying that a game @Lehigh is somehow equivalent to a home-game against Maryland (or a similarly decent but overhyped Power-conference team is, frankly, a bit silly. I understand that you are saying that based on Sagarin or Ken Pom projections...and while it is important to look at those rating systems when evaluating the difficulty of an opponent, there are times where you have to think to yourself "Does what I'm saying actually make sense"? This is one of those times. A team like Maryland is ridiculously athletic, can cause matchup nightmares, etc, that an average to slightly above average mid-major is never going to replicate* whether the game is at home, on the road, or on the surface of Mars. It just doesn't make sense to compare those two, even when Sagarin might tell you otherwise.

I only say that because I really, really enjoy the vast majority of your analysis...and always cringe when you say something like "Well, a road game at Mississippi Valley State is actually tougher than playing Kentucky at home, so this isn't a bad loss". The exaggeration there is obvious and intentional but you know what I mean. It just doesn't work when you mix common sense into the equation.

*Obviously there are mid-major exceptions that really can compare with average to above average major conference teams. Ironically enough Lehigh was one of those not too long agao. But that's not what I'm talking about here.

Trevor Kraus said...

@ David —

You have a good point regarding the quality of the teams. Maryland and Lehigh...yeah, way different. But referees favor home teams so drastically (this has been studied and proven) that they more than make up for that difference. Playing a team like Lehigh, while having most of the calls go against you, is probably about the same difficulty as playing Maryland with most of the calls going in your favor.

Jacob Schiros said...
This comment has been removed by the author.
Jacob Schiros said...

And if FSU goes 1-1 against Duke/ND while winning the remainder of their games? While you do make a good point about FSU not having many big games left with road mines you do leave out the improbable of FSU actually getting to an ACCCG and picking up two more major victories along the way. It is quite ridiculous to still put FSU at a 4/5 seed even with what they are projected to finish at. 25-6 is where they are projected and that would be pretty good for a 3 seed if you add just one win in the ACCT. With the projections that include a 6 loss FSU that would place them somewhere in the 9-12th place finish and add a resume of what they had in the heart of their schedule it is much more likely to place them in a 3 spot and maybe even a 2 seed before a 5th spot which today you have them in. The committee will likely jumble the top 5 acc teams in the 2-4 seeded range with FSU projecting to finish in the top 3 it would be safe to give them a 3 seed. I respect your breakdown here but again you are pushing a 5 seed which is pretty far from the median.