Tuesday, March 18, 2014

Complete Final Four Analysis

Opening Thoughts:

I only have a single 1 seed in the Final Four, but I think that's going to be fairly typical this year. I think most people recognize that Virginia and Wichita State are relatively weak 1 seeds, while two of the top tier national title contenders are both 4 seeds. Basically, a whole lot of your friends are also going to have Louisville and Michigan State in the Final Four, so if you're in a big bracket pool you might actually have an edge by taking a team like Wichita State or Creighton to the Final Four just to be different, and to have a better chance of actually winning outright.

I'll talk about a bunch of possible match-ups here, but I can't possible get to everybody. This post will go 20,000 words if I do. So I'm going to stick to the top tier in each region. In my opinion, that's Virginia and Louisville in the East, Arizona, Wisconsin and Creighton in the West, Florida, Kansas and Syracuse in the South, and Wichita State, Louisville, Duke and Michigan in the Region Of Death.

My Picks:

Final Four:
4. Louisville over 2. Wisconsin
1. Florida over 4. Michigan State

Let's start with the first Final Four game, which I have between Louisville and Wisconsin. I think both of these teams have interesting match-up advantages over each other. Wisconsin has an advantage in that the led the nation in fewest turnovers per game again, and Louisville will not be able to get easy points in transition like they prefer. Louisville is also a team that fouls a lot, and Wisconsin is very good at drawing fouls and could see a significant free throw advantage. That said, Louisville has their own advantages. First, Wisconsin has struggled to deal with speedy, talented perimeter players, and it's hard to think of a team described by that better than Louisville. Louisville's defense is also very aggressive and long, and while they might not force turnovers, they will turn Wisconsin into more of a jump shooting team. And while Louisville is vulnerable against teams with multiple bigs that crash the offensive glass, that certainly doesn't describe Wisconsin. If you're picking Wisconsin here, you're counting on them to shoot significantly better. To me, that's too much of a risk. Louisville is the safer pick.

Between Creighton and Arizona, who would cause Louisville more trouble? I think the answer is Arizona. Creighton has struggled this year against very long, athletic defenses. Louisville will close down on those shooters very quickly, and Creighton is even thinner in the front court than Louisville is. Arizona, on the other hand, is a team that plays inside-out. Assuming TJ McConnell and Nick Johnson can get the ball over halfcourt reliably (and they should be able to), Arizona has multiple bigs who are aggressive in and around the paint and can crash the glass (they led the Pac-12 in OR%). Also, since Louisville's offense is very dependent on the three (they led the AAC in 3P% and were second in 3PA/FGA), Arizona's ability to defend the three is significantly better than Creighton's. I think that if you have Arizona somehow surviving to the Final Four, they're probably the toughest possible West region opponent for Louisville, though you can also argue for Wisconsin. Louisville seems like a nightmare match-up for Creighton, so don't put Creighton in your title game.

Let's say Louisville goes down to Wichita State, Duke or Michigan? I guess we should start with Michigan first, since they played Wisconsin twice during the season, with each team winning on the other's home court. The two games basically came down to shooting, as Michigan hit 54% behind the arc in the win and 38% in the loss. Wisconsin has the ability to limit Stauskas and force other players to score, and Michigan doesn't have anybody in their front court who can challenge Wisconsin's limited front court. Caris LeVert went for 20+ in both Wisconsin games because of the efforts to stop Stauskas, but are you really going to pick Michigan to win a title game because Caris LeVert carried them there? I think Wisconsin is the safer pick.

Of Wichita State and Duke, who poses the bigger threat to Wisconsin? I think it's Duke, because of how explosive their offense is. They have so many guards that can score so many different ways that I just don't see how Wisconsin's relatively mediocre perimeter defense can handle them. The one caveat with Duke is that Wisconsin's bigs would have a shot to get Jabari Parker in foul trouble and totally turn the game around. Wichita State, as I've said before, is an incredibly solid team that isn't really great at any one thing. The fact that they are solid preventing turnovers and collecting defensive rebounds, for example, isn't really useful against a Wisconsin team that doesn't care. The issue for Wisconsin would be that if the game was called tight, Wichita State has significantly more depth. Both of these teams did a good job of drawing more fouls than they committed this season, so depending on the reffing it's possible that we would see a lot of foul trouble.

In the other semifinal, Florida and Michigan State would be a battle between a perimeter-oriented team and a team that works inside out. Florida takes a lot of jump shots, but they tend to take them in and around the paint, and defensively they are good at preventing easy shots as well. They led the SEC in both offensive and defensive 2P%. Michigan State, on the other hand, is pretty reliant on hitting threes. Even big guys like Adreian Payne and Kenny Kaminski can step outside and hit threes at over a 40% clip. Florida's defense has actually been fairly mediocre this season at defending the three. The concern for Michigan State is ball handling, which has occasionally been pretty choppy this season (they finished 9th in the Big Ten in offensive turnover rate), against an aggressive Florida defense that can force a lot of turnovers and scores well in transition. Also, I think we should be just a little bit wary of declaring "Michigan State is BACK". I certainly believe it, and I think most other people do, but it's also possible that Michigan State just got hot in the Big Ten tournament and will regress to the mean in the NCAA Tournament. Certainly if we look at these two teams over the course of the season, Florida was better. And don't forget, Florida themselves was very shorthanded in non-conference play (though obviously they haven't been hit by injuries in recent months as badly as Michigan State). Michigan State's defense has been mediocre at times, and they do commit a lot of fouls. Florida has been the better team over the course of the season, and I make them the favorite here.

If you have Virginia instead of Michigan State getting to the Final Four, that would definitely be one of the best defensive match-ups we've ever seen in a Final Four game. Those two teams are both in the mix for the best defense in the nation. They are different in style, though. Florida's defense is more aggressive, generating many more turnovers. Virginia's defense is more about preventing easy shots and preventing second chance opportunities - they squeeze the life out of a game without making the same spectacular individual defensive plays you see from other teams. Both Florida and Virginia are pretty poor at the free throw line, so that kind of washes out. If there's a difference, it is that Florida is significantly better at scoring inside the arc. Virginia was 2nd in the ACC in 3P% but only 8th in 2P%, and against the swarming Florida defense will likely be pushed even further from the hoop. If those threes aren't falling, Virginia will not score like they're used to. I'd give Florida the slight edge.

What of Kansas and Syracuse? Well, for one thing, the odds of Joel Embiid being fully healthy will increase as the Tournament goes along, so they are a team that should get better the further this competition goes. Andrew Wiggins would be a really difficult match-up, and Michigan State doesn't have the type of imposing 7-footer that has given Embiid problems this year. Also, Michigan State doesn't apply the type of perimeter pressure that would give Naadir Tharpe the kind of troubles he's had when Kansas has lost this season. As for Syracuse, I do think that Michigan State is not ideally suited to play a zone defense. They tend to turn the ball over a lot (Syracuse led the ACC in defensive turnover rate) and the fact that their bigs can come outside to shoot threes doesn't really help them much - they won't be stretching a zone defense the way they stretch a man defense.

Championship Game:
4. Louisville over 1. Florida

I didn't originally intend to have Florida against Louisville in the title game, but it's an interesting quirk. For those who remember last season well, these were the two teams that the advanced metrics loved all season long, and that the national media called "overrated" due to bad luck in close games. In the end, Louisville went on to win the National Title while Florida got upset in the NCAA Tournament, leading the media to ignore the former and use the latter as proof that "advanced metrics can't measure heart", or some such crap. Florida wasn't "clutch" because of their 0-6 record in games decided by single digits. This year, they're 9-2 in games decided by six points or less. But overall, by advanced metrics, they're nearly identical.

Last year's Florida team, even after the early loss in the Tournament was rated 2nd in Pomeroy, while this year's team is 3rd. This year's team outscored SEC opponents by 0.19 PPP en route to their 18-0 record, while last year's 14-4 team outscored SEC opponents by... 0.19 PPP. The teams are the same. And we know that results in close games are almost entirely luck. So how does the traditional media explain Florida's big swing in luck with nearly identical peripheral numbers? Those that follow me on twitter have seen my snark every time Jimmy Dykes goes into one of his speeches on the topic. The three reasons he always ends up giving? They're so good defensively, they're so experienced, and they "know how to win". Well, here's the thing. Last year's team was better defensively and more experienced, and "know how to win" doesn't mean anything. So it's crap. This year's team is the same as last year's team. But this title game, if it comes to pass, will be their chance to win a title that got away from them last season.

Anyway, rant over, how does this year's version of Florida stack up with this year's version of Louisville? Both are great defenses, of course, and aggressive defenses that will force a lot of turnovers, though both are vulnerable to strong offensive rebounding attacks. Offensively, Louisville is a bit better than Florida, but the Gators are no slouch (best in the SEC and, according to Pomeroy, 17th best in the nation). Both teams shoot a lot of threes, but both are also very good at scoring from 5-to-15 feet. Both are also poor free throw shooting teams, though Florida is better at avoiding fouls in the first place. I think Louisville has the backcourt advantage, and will have the best player in the game, with Russ Smith.

I've talked before about how Florida is vulnerable to turnovers from time to time. This year when turned over on more than 20% of their possessions against Pomeroy Top 100 teams, they have failed to win a single game by double digits (and it also included their only non-Wisconsin loss). That includes a 4 point win over Ole Miss, their SEC tourney struggles against Tennessee and Kentucky, and their 2 point win over Memphis. If you watch to stretch beyond the Pomeroy Top 100 you have Florida's rough 7 point win over Auburn, where they turned the ball over on 22% of possessions. In comparison, their blowout win of Tennessee, their regular season handlings of Kentucky and their win over Kansas all featured turnover rates below 15%. Naadir Tharpe and the other Kansas ball handlers can be turned over, and Louisville is one of the most well-equipped teams to do that. And for that reason, I'm picking the repeat.

In years past I've tried to break down a few other national title match-ups here, but there end up being just far too many iterations. I always end up talking about different match-ups in the comments. So I think I'll leave it there. If you really want to talk other match-ups we can, but I feel like Florida and Louisville are the two best teams in the nation (and arguably were the two best teams in the nation last season as well), and it would be fitting if they end up playing for the national title in Arlington.


GriffinCarroll said...

Can you touch upon a possible Wichita St/Michigan St Final? and wichita/florida.....you don't have to go into detail just wanna know how wichita stacks up

Ryan said...

Hi Jeff,

Just have to say that ever since finding your blog a couple of years ago, one of my favorite March pastimes is to compare my bracket to yours and read your data-driven arguments. Thanks for the great read!

What are your thoughts on a Florida/Arizona Championship game? Thanks!

Jeff said...

Thanks. Not sure Arizona would have a good shot against Florida. They're not going to force a lot of turnovers, and they don't have the depth that Florida does. And Arizona is a poor outside shooting team, where outside shooting is really the best way to try to score against Florida if you can't get in transition or easy second chance points.

Anonymous said...

How about a final score prediction for Louisville/Flordia matchup?

Jeff said...

How about a homework assignment? Use the Pomeroy ratings and tell me what his projected score for Louisville/Florida is.

The formula is that projected pace is PaceA*PaceB/AvgPace, and then Team A's score is OffA*DefB*Pace/AvgOff.

Steve said...

Hey Jeff. I've spent the last 2 nights trying to analyze this data and come up with a winning formula. Working on a bracket for my wife as well...in our same pool. Makes things a little challenging since I want to make sure that if my core teams get bumped, then at least she can still have a chance. Anyways, one question for you: Do you really think this Louisville team is good enough to repeat? Are they that much (along with Florida) better than the competition that this doesn't concern you? Obviously, you have to pick a winner. Seems like the odds of going back-to-back would be slim though, no matter the rankings and what they look like going into the tourney. Thanks!

Jeff said...

I think Louisville and Florida are the two best teams. But obviously there's a pretty high likelihood that one or both will be upset by somebody else. That's just the nature of a one-and-done tournament.

I wouldn't worry about the "teams rarely repeat" thing. Because the thing is, that doesn't really apply to Louisville. The primary reason teams rarely repeat is because teams that win titles often aren't one of the best teams their year anyway (a bit lucky) and also because they tend to lose their best players. Teams like the two UNC title squads were great teams, but they were all packed with seniors. This Louisville team brought back its best player and was legitimately one of the two or three best teams last season (if not the best team last season).

Musical Accents said...

Out of 26 teams in my extended family pool, I am the only one who has Louisville winning it all! My wifes bracket, with Florida as the winner, has 4 others. Fortunately, the other side has different teams, except 1. Thanks again for the help this year!

Jeff said...

Cool. Always good to have a very different bracket from everybody else when it's more than, like, 10 people on your pool.