BrackAnalysis: Gonzaga

You may wondering: BrackAnalysis? Yet another branding change? NO! BubbAnalysis focuses on bubble teams. BrackAnalysis takes a wider view. Thus, today's focus: Gonzaga's case for a one seed.

As of yesterday, the Zags were widely considered a lock, maybe THE lock, for a one seed. I wasn't sure if the lock consensus was really based on "they're a one seed even if they lose" or if it was "they're a one seed because there's no way they're losing to Saint Mary's, who they've beaten by like ten thousand points in two games this year." In any case, my algorithm did not agree with the consensus. It was projecting that a loss to Saint Mary's would knock Gonzaga down to the two line. And that's where they are today.

Yesterday's consensus may well turn out to have been correct but I think it's worth looking into the sources for my algorithm's dissent. One preliminary note: none of this is intended as a subjective opinion on whether Gonzaga is good, or whether it deserves a one seed—only an attempt to analyze Gonzaga's profile the way the Committee might.

The Case Against Gonzaga as a One Seed

Gonzaga's problem is that it has only four Q1 wins and only six Q2 wins, so it is projected at 27th in the resumé metric. In my database (since 2008), the only team to get a one seed with four or fewer Q1** wins was undefeated Wichita St. in 2014. They had three Q1 (top 50) wins and eight Q2 (top 100) wins. But they were also undefeated—33 games over .500 on Selection Sunday—and had those two more Q2s. No team (since 2008) has been awarded a one-seed with ten or fewer Q1+Q2 (top 100) wins. Here are all the one seeds since 2008 with fewer than 15 Q1+Q2 wins:

So here's the case against Gonzaga in a nutshell: their volume of Q1+ Q2 wins has never before been good enough for a one seed, and their volume of Q1 wins was only good enough for an undefeated team. Given the primacy of this kind of analysis in the Committee's work, I think this is a very reasonable argument that Gonzaga may not be awarded a one seed.

The Case For Gonzaga as a One Seed

Nonetheless, there are still good reasons to think that the Committee will bestow a one seed on Gonzaga:

• The eye test. Gonzaga passes it. Mainly because ...
• They beat Duke. Full strength Duke. In Maui, in a game everybody watched. And for much of the game, they dominated. 
• They are really, really good. Gonzaga is second in Kenpom and most other similar ratings. They are certainly worthy of a top seed.
• Lock in effect. The Committee was already in session, as of Monday (I believe). I think one of the first things they do is consider one seeds. So it's possible, even likely, that there was a provisional one line, and Gonzaga was sitting on top of it. Once things like that get started, the analysis become less about looking at the whole resumé, and more about "how much should this one result affect what we already thought?" Cognitive biases being what they are, new results tend not to affect what people think as much as results that went into the original opinion.
• No one else fits in the West. All of the other potential one seeds are well east of the Rockies, most east of the Mississippi. If there's a close call between Gonzaga and an eastern team, Gonzaga may get a logistical bonus. This is especially true if the 4/5 on the s-curve comes down to Zags versus Duke because ...
• They beat Duke. Did I mention that?

These are all good reasons, so it will be interesting to see how this shakes out.

**Prior to last year, Q1 = top 50 and Q2 = top 100, with no adjustment for venue (because that's how the Committee used to do it).

BubbAnalysis: NC State

Welcome to the latest edition of BubbAnalysis, formerly known as Bubble Breakdown, formerly known as Bubble Banter.

Right now I want focus on NC State. In the latest bracketmatrix, NC State is an 11. Yet my algorithm currently projects them on the 9-line, and unlikely to fall out even with a loss. What gives?

It inescapably comes down to non-conference strength of schedule. Because they played so many of the worst teams in DI, their "NCSOS" is dead last—at least based on the primary measure of NCSOS highlighted on the NCAA's Nitty Gritty report and Team Sheets, which is simply opponent's winning percentage (i.e. RPI Factor 2). This primitive measure is not an input to the T-Ranketology algorithm. In fact, my algorithm gives no separate weight to strength of schedule as a standalone input (though it is certainly incorporated into the algorithms inputs).

Most bracket projectors, I gather, do separately consider SOS and especially the NCSOS, and I understand why they do. First, it's a number on the official documents. Second, there have been a few high profile cases where a bad NCSOS was specifically mentioned as a reason that a team was excluded. A couple of Seth Greenberg's Virginia Tech teams come to mind here.

I did not include SOS or NCSOS in the algorithm for a few reasons:

• In prior years, the committee used RPI, which was highly correlated to SOS because it was 75% composed of schedule metrics. Indeed, the current "SOS" numbers on the official NCAA documents are just what used to be called RPI Factor 2, which was 50% of the RPI formula. So it was possible to simply use RPI and get good enough results with the algorithm. For example, the 2010 Virginia Tech team that was snubbed supposedly because of its terrible NCSOS also had an RPI rank of 59—not a disqualifying rank, but definitely in the area where a team is not likely to be selected. (My current algorithm would have had them the last team in the field—close enough!)

• It's a very very noisy signal. There are many examples of teams with terrible schedules getting in, and getting good seeds. Just as an example, last year Michigan had a bad NCSOS and got a three-seed. And there are many examples of teams with very good schedule metrics who nonetheless were left out or down-seeded. So it's hard to figure out how to put this into an algorithm because in the vast majority of cases it clearly has no impact whatsoever (at least independent of its effect on RPI).

• Strength of schedule is accounted for in Kenpom, BPI, Sagarin, KPI, and SOR — the other metrics on the official docs. People seem to resist this, but it's clear that Kenpom has long had some effect on selection decisions, and that was formalized last year with its inclusion, along with those other metrics, on the team sheets, etc. Teams that play terrible schedules will typically be punished in those systems unless they actually perform very well against those opponents.

Now, with the NET this year things are up in the air. If we were still using RPI, there is zero doubt that NC State would be disqualified from consideration for an at-large. (Their RPI rank would be 100+, far beyond the realm of consideration, certainly in my algorithm.) And the reason their RPI would be so bad is because their NCSOS kills it. But their NET is in the low 30s, their power ratings are in the 30s, and they have 8 Q1+Q2 wins. Only one team with that profile has been left out since 2008: last year's USC team, which was widely considered an at-large shoo-in. (By the way, their NCSOS rank was a perfectly respectable 62.)

That leaves the question: will the Committee rely on RPI Factor 2 to downgrade NC State, despite having a profile that otherwise matches dozens of at-larges and just one snub? It might! And if it does, I'll probably tweak my algorithm to incorporate some kind of special punishment for having an extremely bad RPI Factor 2 in non-conference games. The easiest way to do this would be simply to put RPI itself back into the mix somehow.

Bubble Breakdown March 02

Note: After consultation with marketing and legal,  I'm rebranding Bubble Banter as Bubble Breakdown™.

Here's a look at the bubble through the robotic yet strangely beautiful eyes of TourneyCast:

The nice thing about this view, as opposed to the current T-Ranketology bracket, is that this includes simulations of conference tournaments. So a team like UNC Greensboro, which might reasonably considered on the bubble as of today, fades to extreme long-shot as an at-large when you factor in a loss in the conference tournament. It also is the best way to look at bubble teams who are projected auto-bids in the same view as bubble teams who are not.

Let's do this.

Probably safe: Ohio State, Oklahoma, Syracuse, Washington, TCU

These teams generally just need about one more win, at least according to my algorithm. The big mark against several of these teams is poor conference record. Oklahoma and TCU are both currently projected to go 7-11 in Big 12 play, and Oklahoma would be projected to be fine even at 6-12. It's certainly possible that the committee will balk at those figures, as it's pretty much unprecedented for teams with those kinds of records to have tourney credentials otherwise.

Getting sketchy: St. John's, UCF, Clemson, VCU, Belmont, Texas

St. John's has 6 Q1 wins (though FWIW I'm projecting Georgetown will fall below the Q1 cutline eventually) but also has home losses to the likes of DePaul, Providence, and Xavier. Now they have two road games left: at Xavier and at DePaul that T-Rank sees as tossups. If they win at least one of those, they'll probably be okay heading into the Big East tourney in their backyard (note that home-court gives them a 1 in 5 chances of getting the auto-bid). But losing both those games is a real possibility, in which case they'd be squarely on the bubble.

UCF isn't worth discussing too much because their resumé remains largely to be written: three Q1 games still left in the regular season. Sweep those, and they might be wearing home jerseys in their first game of the tourney. Get swept and say sayonara. Anything in between and they'll be on the bubble.

Clemson is a team my algorithm likes more than most bracketographers. They lack the top-line wins (just projected to have one Quad 1) but have a good NET, good power ratings, no bad losses, SOS is fine. Their profile comps are all in, though never by much:

Obviously a win today over UNC will seal the deal. But if they just win the games they're supposed to, my algorithm will likely continue to favor them. I'd generally trust the conventional wisdom over my algorithm, but this could end up being an interesting test.

VCU and Belmont are both mids with a decent enough shot to get in even if they're knocked out in their conference tourney, as at least one of them likely will be (both projected at about 50% to win). All bubble teams should be rooting for them.

Texas is here, as I mentioned last week, because it is projected to be 16-15 at the end of the regular season. My algorithm starts punishing a team once its projected record is less than five games above .500. The punishment is pretty severe at a one game over .500, so it's pretty remarkable that Texas is still being projected on the right side of the cutline. If Texas gets two more wins, they will vault out of the bubble zone, but this is another sort of unprecedented case brewing, so it's hard to really say what would happen if Texas finishes at, say, 16-16 with five Quad 1 wins.

The Genuine Bubble: 

Minnesota needed two wins last week to feel safe, but they got one. They remain on the bubble with two Quad 1-A games remaining. One upset would likely do it for them. If not, they'll need two Big Ten tourney wins to be safe, I think.

Utah St. I still think Utah State needs to beat Nevada once, and that's basically where this 50% chance of an at-large comes from: they've got a tossup against the Wolfpack tonight. Win it, and they're in. (By the way, my algorithm's early projection of Utah State as an at-large contender, when no one else was even considering them, is why it is a cool thing.)

Alabama, like UCF, has three big Q1 games left in the regular season, so their profile is incomplete. All three are projected by T-Rank as tossups. Tune in next week, but in all likelihood they'll be on the bubble heading into their conference tournament.

Why does everyone hate Saint Mary's? This has vexed me for a while, and Ken Pomeroy mentioned it on a recent podcast. Even people who tend to root for mid majors root against Saint Mary's for some reason. I don't get it at all. People hate Saint Mary's resumé because they ain't beat nobody. I like Saint Mary's because they are, objectively, a good basketball team. They also have a great player in Jordan Ford. If they beat Gonzaga tonight, their profile will look good enough. If they don't, it won't.

Seton Hall is a team that projects in the field as of today, but they have three tough games upcoming (at Georgetown, Marquette, Villanova), and could very easily lose their first game in the conference tourney. They can definitely win their way into the field, but T-Rank foresees losses.

Furman and Lipscomb are alive. Technically. Better not to chance it though.

Temple is a team that I think many have been overrating, in terms of tourney chances. They have just one Q1 win, and only one more regular season opportunity: a Bracket Buster game against UCF on the last day of the regular season. The loser of that game could be in trouble, especially if it's Temple. That said, Temple's profile comps are 9/10 in the tourney:

Lastly, Arizona State. Bracket Matrix contributors refuse to budge on this team, despite their 28-point loss to Oregon the other night. Yes, they have a few Q1 wins (though many are giving them credit for a win over Washington that doesn't currently qualify as Q1). But they have, in the Committee's parlance, a fuckton of bad losses. And they are bad (67th at Kenpom). And their NET sucks 68th. And they'll be underdogs in their last two games, Oregon State and Arizona, both on the road. Only one team in my database has made the tourney with a sub-60 NET/RPI and a sub-60 Kenpom/Power ranking: Boston College in 2009. (That team had wins over Duke and at (eventual champ) North Carolina. Put it all together, and the profile comps are not pretty:

Now, to be fair, the comps for last year's teams were just as bleak, and the Committee decided to let that team in. As far as I can tell, that selection was based exclusively on the win at Kansas. Will a win over Kansas at home be enough this year? Maybe, but it shouldn't be—unless ASU wins a couple more games.

Bye for now!

Bubble Banter Feb 24

Introducing a new feature: Bubble Banter.™*

Here's a look at the ever-shifting projected pre-conference-tourney bubble, courtesy of the all-seeing but somewhat vision-impaired T-Ranketology Algorithm:

First, some overall thoughts.

The bubble is thin. I don't think anyone beyond Butler here is a real threat to get an at large, and the teams below Temple (Furman, San Fran, and Butler) are pretty questionable too. If you look at the "Bid%" column, which is the ultimately score that I sort this by, historically only one team since 2008 has made the tourney with less than a score of 10 (Iona in 2012) and only four others have made with less than a 20.

But the bubble will tighten up during conference tourney season. There are number of conferences that are fairly likely to produce "bid thieves" this year. Here are current auto-bids toward the bottom of at-large portion of the bracket:

The top four (maybe five) teams there are all probably in the dance no matter what, and each has a pretty good shot of not winning their conference tournament. Belmont and Lipscomb would be long shots as an at-large, but stranger things have happened.

All told, at the end of the day there's a decent chance that we get a very well defined bubble—similar to what we had in 2017, where there was near unanimous consensus on who should be in, and only one or two other teams with even a borderline case.

Now for some team-by-team.

TEXAS. This is a team that is not widely considered on the bubble—they're currently the top 9-seed at Bracket Matrix—and the consensus is probably correct. They have four Q1 wins, three of which are "Q1-A." That's why they've got that "19" in the resumé column. So why are they here? Because they are 15-12 and T-Rank is currently projecting them to finish the regular season 17-14. Teams less than 4 games over .500 do not generally make the tournament. Their remaining schedule is very difficult, so they could play well and still finish with a very ugly record. Something to keep an eye on.

CLEMSON. Clemson has the opposite problem, in that they are 1-9 in Q1 games. They've only got one sure opportunity left, a home game versus UNC. A win there would be very helpful. But if they lose that and win the other three (at Pitt, at Notre Dame, Syracuse) they figure to be sweating it out.

MINNESOTA. The Big 10 has six locks, one likely (Ohio St.), and Minnesota on the bubble. Minnesota has a pick em game at Rutgers today and another toss up coming up at Northwestern. I think they'll need to win both of those, or else they'll need to upset Purdue or Maryland.

UCF. Tacko Fall's crew just throttled SMU by 47 today. It will be interesting to see how much this gooses their already respectable NET ranking. This is another team that lacks the committee's prized Q1 wins (0-3), but they've got at least three more opportunities: at Houston, Cincinnati, and at Temple. One win there keeps them in the discussion, two should be enough.

UTAH ST. The Aggies have two huge home games coming up: San Diego St. and Nevada. They're on the bubble now because T-Rank pegs them as a slight home favorite over Nevada. If they lose that game, they'll probably need to win the MWC tourney.

SETON HALL. A prototypical high-major bubble team. They're buoyed by three Q1 ones, including two super wins over Kentucky and at Maryland. But T-Rank projects them to finish 17-13, which means winning just one out of their last three (at Georgetown, Marquette, Villanova). Beating the Hoyas and winning one of those big home games would go a long way.

SAINT MARY'S. This is another team that is not on the consensus radar because they lack Q1 wins. And they are below the actual realistic cut-line (accounting for a couple of bid thieves) even here. But they show up in this projection because they still have a home game against Gonzaga on the calendar. They'll have to win that game or beat Gonzaga in the WCC tourney to get in.

ALABAMA. Another major conference team with a barely-acceptable overall record. T-Rank is projecting them at 18-13, and this is not a team that can likely get in with less than 19 overall wins. They do have home Q1 opportunities coming up against LSU and Auburn, so they can forge their own path with wins there.

ARIZONA STATE. The algorithm is lower than the consensus here (Bracket Matrix has ASU on the 11 line, above five other at-larges). ASU has some terrible losses, and it's always hard to know how those will be factored in, if at all. But what's really driving this projection is that they finish the season with three losable road games (Oregon, Oregon St., Arizona) that won't give them any juice for winning. If they win two of those, they'll probably be in good shape.

TEMPLE. Another team just above the consensus cutline with some tough games ahead: at Memphis, at UConn, and UCF. The Memphis game would be a Q1 win, so it is a crucial high leverage tossup.

FURMAN / SAN FRANCISCO. I don't think these teams have a realistic case any longer as an at-large because their resumés cannot withstand a conference tourney loss and they have no opportunities to improve it. Theoretically USF could beat Gonzaga in the WCC semis and lose to Saint Mary's in the finals, but that still probably wouldn't cut it.

BUTLER / CREIGHTON / GEORGETOWN / NEBRASKA. These are the dregs of the theoretical bubble. They're here because they have the opportunities to improve their resumés with big wins. But they probably won't.

Update: On reflection, probably too harsh on Butler here. Main problem with their profile, per the algorithm, is the projected record of 17-14. That's presuming a 2-2 finish against Providence, at Villanova, Xavier, and at Providence. If they go 3-1, even losing to Villanova, they'd probably still be a live heading into MSG.


*It is extremely likely this will be the only installment of Bubble Banter™ so please cherish it.

Teamcast tip: note the Projected T-Rank variable

One of the newer features over at the T-Rank site is a tool I call "Teamcast," which lets you select the results of a team's games and shows how those results would affect the T-Ranketology forecast.

I'm not gonna lie, I think this is a pretty awesome tool and I play around with it a lot. But of course there are some pretty important caveats to keep in mind:

1) The underlying T-Ranketology algorithm is not a crystal ball. Although it performed really well last year, last year was a pretty predictable year. In retroactive investigations, T-Ranketology does "pretty good" for prior years too. But overall it's safe to say that you should consider any T-Ranketology forecast to have at least a +/- one-seed margin of error.

2) There's reason to believe that this year will be much less predictable than prior years, because the Committee will be considering new and different information (most importantly, the new quality win quadrants). We can only presume how the Committee will use the new information. My guess is that they will use it to make some ... unpredictable ... choices.

That said, the Teamcast is still a fun tool, and at the very least it should be pretty good for looking at relative changes. If it's predicting your team to be an 8-seed, and picking it to win a game moves them up 4 spots to a 7-seed, it's reasonable to say the win was worth about one seed-line—even if it was actually from a 10 to a 9, or a 6 to a 5.

One important thing that gets overlooked, I think, by people playing around with the tool is that the current T-Rank rating plays a role. First, it is obviously how predictions for future games are calculated. Second, it is an independent variable in the T-Ranketology algorithm, and a not insignificant one. (Not because I think the committee is using it, but as a good-enough proxy for other ratings like Kenpom that the committee is using.)

So particularly if you are exploring an extreme scenario, like say, "What would happen if my team won the rest of its games?" you might get a weird result if you do not change the projected T-Rank as well. Because in the scenario where your team wins out, its T-Rank is probably going to improve a lot.

Fortunately, the Teamcast tool lets you change the projected T-Rank (right at the top of the page), which fixes this problem. In fact, simply changing the projected T-Rank and doing nothing else is a pretty good way to use Teamcast earlier in the season. E.g., keep upping the projected T-Rank until your team gets in the field, and that shows how well your team will have to play the rest of the way to likely make the tournament.

ADDED: One way I've attempted to automate this is the "DynamaRank" feature. Basically, this estimates changes to power ratings (including NET) based on the changed results. So this box should be checked for any extreme scenarios where a team loses or wins a lot more games than expected.

If you get any funky results with the Teamcast, or have ideas for improvement, please let me know.

T-Rank: New Stuff for 2017-18 ICYMI MEGAPOST

I've added a bunch of stuff to the T-Rank site this offseason (or at least since the start of last season) and thought it might be a good idea to inventory & explain it somewhere. Somewhere like here.

First, a little back-patting. Three different aspects of the T-Rank site were validated as pretty useful last year: the ratings system itself, the "T-Ranketology" tourney predicting algorithm, and the preseason projections.

The rating system performed well. In terms of predicting games, it was the best full-season predictor, at least according to this analysis. At the very least, T-Rank is in the same league with Kenpom (no surprise, since I copied Kenpom liberally) and Sagarin. This means the adjusted efficiencies that drive the ratings are likely sound as well, and all the fun stuff that I use those adjusted efficiencies to power has at least some relation to reality.

The T-Ranketology algorithm did a great job last year, finishing near the very top of the Bracket Matrix competition. Of course, this was probably lucky, and last year was a bit of a strange year in that most people were able to correctly pick the field. But again, the main takeaway is that T-Ranketology is "good enough" to use for other fun stuff, such as the new Teamcast and Tourneycast features, and feel confident that the results are at least worth taking seriously.

Finally, last year's preseason projections turned out the be the best 1 to 351 projections, at least according to Dan Hanner's analysis. This is somewhat of a bittersweet victory because—as I've explained in an update to this old post—I have actually totally revamped the preseason projections for this year. But I'm confident that the new system is even better, and I'm also quite certain that last year's "victory" was mostly luck. Still, coming out ahead of Hanner and Kenpom by any metric at least shows that the ideas behind the system aren't garbage.

So putting these three facts together, I think it's safe to say that T-Rank's underlying algorithms and ideas are reasonably sound, which makes the tools on the site even more fun to play around with. Speaking of those tools:

Teamcast

This is probably the coolest thing I've added: a tool to play around a team's schedule (picking wins and losses, adding dropping games) to see how it affects their tourney chances and seed, based on the T-Ranketology algorithm.

The last two columns there show how much of an effect a win or loss would have on a team's T-Ranketology score (in isolation from all other games). Teamcast is also retroactive back to 2008, so you can go back and look what would might have happened if a certain game had come out differently.

I'm excited for Teamcast this year to play out various bubble scenarios—especially given the likelihood that the Badgers will be on the bubble.

Daycast

Daycast is similar to Teamcast, except that instead of playing around with a team's schedule, you play around with the games on a given day to see how it affects the tourney. As part of this, I've developed a new thrill quotient, the "Torvik Tourney Thrill Quotient" or "T3Q" which ranks games in their anticipated effect on the tourney. Basically, games between potential bubble teams reign supreme.

TourneyCast™

For TourneyCast, I run 10,000 simulations of the season (including simulations of every conference tournament), run the results of each sim through the T-Ranketology algorithm to get a projected field of 68, and then simulate the NCAA tournament. The output is odds for every team to get to the tournament (whether through an autobid or at-large) and then odds for advancing to various stages of the tourney.

ADDED: one cool feature of the TourneyCast is that if you filter by conference you get the projected number of bids, at-large chances (interesting for mid-majors), final four chances, and championship chances.

Backfill to 2008

Not sure exactly when I did this, but the site is completely backfilled with team and player stats all the way back to 2008. This includes advanced stats boxscores, which certain other sites don't have back past 2011.

The only exception to this is play by play based stats (win probability, game score). I have a complete set of PBP for last year (2017) and 95% coverage for 2015 and 2016. I will probably add more backfill on that in the coming months, but my impression is that available PBP peters out around 2011 anyhow.

This also includes conference-only advanced stats, and all the other various ways I've got to split the player stats, back to 2008. Notable other sites have conference-only advanced stats back to just the 2014 season.

More Player Stats Splits and Filters

As I just mentioned, I've added new ways to split player stats in the pretty awesome Player Finder tool. Now you can filter out the stats so you look at performance against only top 50 (adjusted for venue) opponents. You can even filter by date, to see who the stars of November were. I've also added a "max height" filter so you can see who the best short rebounder is, for example.

Player Histories and Game Logs

Part of the backfill is that I have advanced player stats back to 2008, including game logs. In addition, one of my major projects this summer was to create a linked database of player histories, so that when you click on a player's name, all his seasons come up. You cannot imagine what a supreme pain in the ass this was. The things I had to learn, I shit you not, were things like natural language processing. Woo boy. But at this point the player histories are about 99% linked. If you see any any discrepancies, please let me know.

Charts and graphs! Charts and graphs!

I just learned how to make some awesome interactive charts and graphs, replacing the old jpeg-based graphs it took me a week to learn how to make. Pretty proud of myself. This new skill is on display in three areas:

Win probability charts

I blogged about my fun foray into created a win probability model here. I've now learned how to make the result more useful and interactive.

Old:

New:

Also: Win probability calculator.

Team Trends Charts

Previously, I had team trends charts for just offensive and defensive efficiency, and they were pictures like the old win prob chart. Now I've got em for most team stats, and they are interactive and fabulous. Note: these charts are also available at the bottom of the "Team Results" page.

Player Stats Trends

Not to be done, the player stats department demanded access to the charts and graphs technology.

Strength of Schedule Stats & Page

Every team page has strength of schedule stats, and there's also a comprehensive page for strength of schedule. These are broken down by overall SOS and non-conference SOS, and also broken down by the SOS so far, and projected SOS for the entire season. (Conference SOS calculations are on the conference pages.)

Might as well say a word about my SOS metrics, since they're a bit unusual. The "basic" metric is just the average of the team's opponents' pythagorean expectancy (adjusted for the location of the game).

The "elite" metric is different, and better. This is the percentage of games an "elite" team (approx .9000 Barthag) would be expected lose against a given schedule. This is better because it reflects how little difference there is between playing mediocre and bad teams at home, and really rewards playing games that a good team could realistically lose.

To see what this is better, imagine three potential opponents: one good team (Barthag of .9000), one mediocre team (.5000), and one terrible team (.1000). Under the basic method, playing two mediocre teams is the same as playing a good team and terrible team because (.5 + .5) / 2 = (.9 + .1) / 2. But this really isn't right for most teams that we care about, because an elite team's chances of losing to the mediocre team (at home, at least) aren't really that different than its chances of losing to the terrible team, while its chances of beating the elite team are quite a bit different: (.05 + .01) / 2 < (.50 + .01) / 2.

Coach History and Team History pages.

A concise way to look at team and coach performance back to 2008.

Advanced Analysis of Unbalanced Conf. Schedules

I blogged about the underlying analysis here: I run simulations to see how differently teams would perform against a true round robin schedule (both winning percentage and chance of winning the conference) and log the results. 

RPI Forecast

Here's one I forgot about: a page showing the projected RPIs of every team based on their current T-Rank rating. This page also shows projected records in each of the selection committees newly delineated "quality win" buckets. If you hover over those records, you can see which teams are in each bucket:

T-Ranketology note

Well, it's Selection Sunday and the current version of the T-Ranketology algorithm has the same at-large field as the consensus (of which it is a small part) over at Bracket Matrix.

Mission accomplished.

I say this because the point of T-Ranketology isn't to try to predict the most accurate bracket on Selection Sunday. The point is to project a reasonably plausible bracket earlier in the season, so that we can see where things are reasonably likely to end up if teams keep performing like they have been. That T-Ranketology is able to basically produce the consensus field on Selection Sunday, with most teams seeded within one line of the consensus, shows that it is "good enough" to provide those useful projections earlier in the season.

ADDED:

Should note somewhere, so might as well be here, that I added one tweak to the algorithm on Selection Sunday: a good record bonus. One of the notable things about the bracket the algorithm was producing over the past few weeks was that it was notably down on the three PAC-12 teams. It seemed to me that this was probably a result that those teams just had really great records in their so-so conference. Whatever you want to say about the Pac-12, it's just hard not to be impressed by a team that's 29-4 or 30-4.

But I resisted adding a good record bonus—until Gonzaga fell off the one-line. I thought it was pretty clear that Gonzaga was going to get a one seed. The only drama on the one line was whether Duke or UNC would get the ACC's slot. After winning the ACC tournament, Duke did indeed sneak onto T-Ranketology's one-line—but at the expense of Gonzaga, not UNC.

So I pulled the trigger on a record bonus—really a "few losses bonus:" Teams got one point subtracted from their score for each loss under five. In other words, four-loss teams (like Arizona and UCLA) got one-point subtracted, and teams with one loss (Gonzaga) got four points subtracted. This not only got Gonzaga back on the one line, but it was also just enough to push Arizona onto the two line, which was pretty clearly where it was going to end up.

The result was that T-Ranketology was one of the few brackets to nail both the ones and the twos. I'll take it.