Wednesday, February 3, 2016

T-Ranketology

As promised, I've been playing around quite a bit with various T-Rank projects this year. Lately, I've added a couple of fun new rankings, including the Q-Rank (based on performance in Tournament Quality Tests), the H-Rank (ratings in last 10 games only) and the E-Rank, a slightly modified "Elo" rating system. I've also put together a program that uses T-Rank to forecast what teams' RPI will be at the end of the regular season.

My marketing people think I'm diluting the T-Rank brand with all these other rankings, but I have to give the people what they want, even if they don't know that it's what they want, yet.

Speaking of which, the latest: T-Ranketology. This is my NCAA bracket projections system. The inputs, in equal measure:

1) Current T-Rank
2) Projected RPI Rank at the end of the regular season
3) Projected Elo Rank at the end of the regular season

Both RPI and Elo are "resume" ranks, and T-Rank is a measure of team strength. The key to understand here is that the RPI and Elo inputs are projections -- that is, I run simulations of the rest of the season to get an average final RPI and Elo. (The T-Rank, however, is treated as a static determinant of team quality.)

So T-Ranketology is not a prediction based on "if the season ended today." It's a projection based on how the season will likely play out, if the current T-Rank is a correct measure of team strength.

One thing to make clear: my projection of the rest of the season does not include conference tournaments. I ain't got time for that.

Finally, I use current T-Rank to pick the autobids. That is, the highest team in the T-Rank is gifted the autobid, on the assumption (not always correct, given seeding effects) that the best team is most likely to win the conference tournament.

As always, this is all in good fun. If you see anything goofy, let me know.

Wednesday, December 30, 2015

Wisconsin Sports Person of the Year -- 2015

I'm pleased to announce the winner of the 2015 Bart Torvik Wisconsin Sports Person of the Year Award.

It is  ...
...
...
...
...

Monday, December 21, 2015

Does Northwestern have a chance?

Northwestern is off to a promising start, sitting at 10-1, with its lone loss a reasonably competitive game against North Carolina on a neutral court. They finish out the non-conference season with home games against Sacred Heart and Loyola Maryland, so they have a good chance of heading into the Big Ten season at 12-1. They sit at 55th in the T-Rank, 51st in Kenpom, and have been climbing in both.

The bad news is that their schedule has been really weak. They have just one win against a T-Rank top-100 team, a road win over Virginia Tech in the B1G/ACC challenge. That's a decent win, but not one that is going to knock the committee's socks off. Virginia Tech will probably be battling to finish top 10 in the ACC and projects to finish 5-13.

They also won their other true road game, at DePaul (T-Rank #188) in overtime. That's not a good win by any means, but it's the kind of game that even good teams can lose. It was important for Northwestern to avoid that bad loss.

Their only other remotely respectable non-con wins were a home game against Columbia (T-Rank #128, which they also won in overtime, and a close neutral-court win over Missouri (T-Rank #172). Again, nothing to write home about.

Because of this weak schedule, Northwestern is only 74th in the RPI. Worse, RPIForecast.com projects them to finish the regular season at 81st, and that with a projected conference record of 10-8. Once again, the reason for this is that Northwestern is blessed/cursed with a soft Big Ten schedule -- based on T-Rank, only Michigan St. has an easier slate (and that's because the Spartans don't have to play themselves).

The non-conference part of the schedule is by design: Chris Collins purposely set up the schedule so Northwestern could bank enough non-conference wins to give themselves a good chance of getting into the post-season. He's basically admitted as much:
"To me," coach Chris Collins told the Tribune, "this was the best (scheduling) philosophy for this team." 
Some background: Collins' first team in 2013-14 doomed itself with a nonconference slate featuring Stanford, Illinois State, Missouri, UCLA, N.C. State and DePaul. Even after pulling off road victories against Indiana, Wisconsin and Minnesota, the Wildcats finished 14-19. No postseason. 
They went 9-4 last season in nonconference play and finished 15-17 — about three victories short of an NIT berth. 
The NIT is not the goal, of course, but it would be a step forward. And Collins wants his players to have a winning season, at minimum. 
"We're still a young program, in Year 3," he said. "We're in a great place with all the stuff behind the scenes: culture, attitude, practice habits, strength and conditioning. The next step is you learn how to win. A place like here is not a quick fix. It's development."
So this year was essentially "NIT or bust" for the Wildcats. That's a perfectly rational and defensible strategy for this program.

What Collins likely did not anticipate, however, was that the Big Ten schedule would have so many wins there for the taking. At least two perennial powers (Wisconsin and Ohio State) look very beatable this year, and several other middling teams (Minnesota, Illinois, Nebraska) look downright bad. That's why the computers (T-Rank, Sagarin, Kenpom) think Northwestern will probably win 9 or 10 B1G games, even though they're ranked around 50th in all of them.

This is not last year's Penn State, which lucked its way to a 12-1 record facing a slew of terrible teams. So despite the gaudy record, Penn State was 100th in the T-Rank and projected to win just six Big Ten games.

For the most part, Northwestern has looked like a solid high-major in its games against bad teams -- blowing out its sub-200 competition. (A seven-point win over Fairfield is the exception there.) And it has no bad losses. Thus, the respectable T-Rank / Kenpom / Sagarin ratings.

So, does Northwestern have a chance to make its first NCAA tournament? Yes. Like the Badgers, they'll have a good chance if they get at least 11 Big Ten wins. T-Rank gives them about a 16% chance of going 11-7 (or better) in the regular season, and a 17% of going 10-8. If they go 10-8, they'll certainly have to avoid a bad loss in the B1G tourney's first round. Let's say they've got only a ~50% chance of winning that game, leaving about 9% chance to get to 11 total B1G wins in that scenario.

Add it up, and it's about a 25% chance of getting to 11 wins. That seems about right. It's too bad, because with a decent non-conference schedule they would probably be a lock with 10 wins, which would get them close to 50/50.

It will make for an intriguing season for Northwestern fans, though. They're in it -- they just need a little bit of luck. Their first six B1G game include winnable road games at Nebraska and Minnesota, winnable home games against Ohio State, Wisconsin, and Penn St., and a big win opportunity at home versus Maryland. Probably need four wins to stay viable. Stay tuned!


Wednesday, December 16, 2015

Miss you already Bo

Last night I got home from the game in time to watch the press conference. It was a disappointment. I had bought into the idea that Bo would stay at UW forever and somehow a half robot Bo would still be coaching after I passed away. I don't want to think about UW without Bo.

I went to bed eventually. My girlfriend said I woke her up because I was talking in my sleep. I was yelling "WOW, awesome" and "OH Yeah, just like that!". On a normal day I would assume that meant some kind of sex dream, but last night I bet I was dreaming about the Final 4 win against Kentucky.



Monday, December 14, 2015

Thoughts on Badgers' schedule and tourney hopes

After two more home losses to Milwaukee and Marquette, the Badgers are now 6-5. They've fallen to 60th in the T-Rank and 42nd in Kenpom (where they are still somewhat buoyed by their implausibly high preseason rating of 9th in the country).

T-Rank currently projects the Badgers to finish the season 15-16 and go 7-11 in the Big Ten.

Kenpom currently projects 17-14 and 9-9.

I'm hopeful that the Kenpom projection is more realistic. Although UW is now playing with a very tight rotation the features four juniors, redshirt freshman Happ, and spot duty from three freshman (Thomas, Iverson, and occasionally Illikainen), it's still reasonable to expect that the four freshman in the rotation could significantly improve as the season goes on. I do think we've seen the worst of this team, and I hope we haven't seen the best of it.

Even so, the Badgers will face a tough test in the Big Ten schedule, as almost all their double-plays are against the top teams in the conference: Michigan St., Maryland, Purdue, Indiana, and Illinois. This is why T-Rank projects them to have the second-toughest conference schedule.

But their single-plays are looking better. In the single-play games, you want to play the tougher teams at home, and the easier teams on the road. Overall, I think the Badgers got a bit lucky on the locations of their single play games:

Rutgers, home: would prefer this on the road. Minus.
Northwestern, away: tough to tell, but for now this looks like a preferred home game. Minus.
Penn St., away: a very winnable road game. Plus.
Ohio St., home: a game they should win, but would be tough on the road. Plus.
Nebraska, home: Nebraska looks tough at home, bad on the road. Plus.
Iowa, away: Minus
Michigan, home: Plus.
Minnesota, away: given how terrible MN has looked: Plus.

So, +2 overall. Only the Minnesota and Northwestern games are really questionable calls, so maybe call them both pushes and they're still +2.

In any event, only one of those games would be a shocking win (at Iowa). Anything less than 5-3 would be disappointing, and 6-2 seems doable.

In a normal year, you'd look at that and say the Badgers can count on splitting the double plays, so 10 or 11 wins and -- given that the Badgers will have a top-10ish RPI SOS -- an at-large berth are likely. This year I'm not so sure. This year home wins over Purdue, Michigan St., or Maryland seem like steals. So it's going to be interesting.

How interesting? I think it's likely that the Badgers' season will come down to their first game in the Big Ten tournament. Let's take a mildly optimistic but still reasonable scenario and run it through the RPI Wizard at rpiforecast.com:

Wins: Rutgers, @NW, @PSU, IU, @IL, OSU, Neb., IL, Mich., @MN
Losses: Purdue x 2, Michigan St. x 2, Maryland x 2, @IU, @Iowa
Total: 10-8
BTT: Beat Northwestern, lose to Michigan St.
Result: RPI 41, SOS 5

That scenario would put the Badgers right on the bubble. On the downside, they'd have no really good wins in conference -- only IU and Michigan would probably be tournament teams, possibly Northwestern (though NW probably has to beat Wisconsin to be in contention). And they'd be carrying the baggage of at least one (Western), probably two (Milwaukee), and maybe three (Marquette) "bad losses." Still, a > .500 conference record, a decent RPI, an elite SOS, and two decent non-con wins would probably be enough to at least get the Badgers into the field of 68 (if not the field of 64) -- though they'd probably be in better position if they traded the win at Northwestern for a more "marquee" win over Michigan St., Purdue, or Maryland.

Now run those same numbers, but with a BTT loss to Northwestern:

Result: RPI 55, SOS 9

That RPI would probably be insurmountable, unfortunately.

Of course, the chances of either of these particular scenarios playing out exactly this way is vanishingly small. It's more likely the Badgers will lose a couple here, gain a couple there. But since this is probably worst resume the Badgers could have with 10 wins, it's worth considering. And even in alternate 10-win scenarios, that first BTT game is likely to be a must-win or close to it.

One way or another, I think the Badgers will have to scratch out 11 wins against Big Ten teams to get into the tournament. I think they can do it. I think they will do it. But there's really no margin for error anymore.

Saturday, December 12, 2015

Introducing FUN

Later today at T-Rank I'll be adding a team stat I call FUN, which stands for

Fortune Unexplained by Numbers

or

Failure Unexplained by Numbers

This is an equivalent stat to what Ken Pomeroy calls "luck." Here's how it is calculated: T-Rank goes back in time and calculates an expected winning percentage for each team schedule so far, based on their T-Rank profile. Then it compares it to their actual winning percentage. The difference (expected minus actual) is their FUN.

I'm avoiding the word "luck" here for a reason. If I call it luck, that implies that the T-Rank is a rather absolute reflection of a team's quality, and that any deviation from T-Rank's expected winning percentage is simply random variance, or luck.

Undoubtedly, that is a big part of the story. But it's probably not the whole story. 

Let's look at a team like Maryland, which last year was very "lucky" in that it's actual record was much better than it's Kenpom / T-Rank expected record. The reason for this difference is pretty easy to suss out: Maryland went 12-1 in games decided by 6 or fewer points. That's kind of ridiculous. Most of the time, teams will go about .500 in close games -- even very good teams.

But maybe there's something about Maryland that makes them particularly good in close games. Whatever this something might be, there's no chance Kenpom or T-Rank will reflect it. It could be random variance, just plain luck. But it could be something else.

And with Maryland, there's actually a pretty obvious hypothesis about what that something might be: an all-American point guard (Melo Trimble) who can get to the line at will, and hits 90% of his free throws. This is something that is particularly useful in close games. And it turns out that there's good evidence that Trimble does indeed get a disproportionate share of his free throws in crunch time of close games:
So this is the kind of thing that might be something other than random variance that causes a team to outperform T-Rank expectations. 

The other reason I'm calling it FUN is that variation from reasonable expectations is one of the things that makes college basketball so interesting and, well, fun. After all, college basketball is where the unexpected becomes ordinary. Or so I've been told.

[T-Rank says: FUN is luck, nothing more.]

Thursday, December 3, 2015

The last 5, and the next 5...

You may recall that I was pretty worried about the five-game stretch that just ended with the Badgers' thrilling overtime win at Syracuse. My conclusion was pretty pessimistic:
Long story short, 2-3 over the next five games would probably be the best we can reasonably hope for, with 1-4 sitting out there as a disturbingly plausible scenario.
Even in hindsight, I don't think that was unduly alarmist. The Badgers' two losses -- to Georgetown and Oklahoma -- were not all that closely contested. Both were double-digit losses that the Badgers had no chance to win in the second half.

The Badgers' two big wins -- over VCU and Syracuse -- on the other hand, were extremely tight, touch-and-go affairs that went down to the wire. It took a great drive by Bronson to beat VCU at the buzzer, and of course it took an extra period for the Badgers to beat Syracuse.

In other words, even knowing what I know now, if we played those five games again I would expect the Badgers to go 2-3, or maybe even 1-4. So: hooray!

Luckily, we don't have to play those five games again.** Instead, the Badgers now start a seven-game home-stand, and it starts with these five games:



As you can see, T-Rank has the Badgers significantly favored to win each of the five, with the Temple and Marquette games being the only serious stumbling blocks. If you add up the win percentages, it comes to a projected record of 4.1 - 0.9, so there's certainly a legitimate likelihood of dropping one. But this will be by far the softest five-game stretch the Badgers will have for the rest of the season, and I'm looking for them to make some hay and solidify their tournament resumé.

Prediction: 5-0

That will set up a potentially monster matchup against Purdue on December 29th in the Big Ten opener. Purdue has been absolutely dominant so far this year, and they have a real chance of coming into that game undefeated -- ranked in the top 5 or even No. 1. We'll see how things go, but I'm looking to that game as a coming out party for the Baby Badgers.

**Looking back at the Badgers' non-conference schedules since Bo's first year, I can't find a five-game stretch that is even remotely as difficult as the last 5, which included four games against likely tournament teams, all of them away from the Kohl Center. Going 2-2 in those four games was a tremendous achievement for such a young team.