How to do a 96-team tourney the right way (88: better, 112: do-able)

As someone who runs tournaments and mints brackets for a living, here’s how I would’ve structured the new 96-team format that the NCAA just announced. First of all, since I’m an NBA guy, I’m all for the expansion because it allows NBA’ers to better evaluate draft prospects under one roof – but that’s for another blogpost.

NCAA vice president Greg Shaheen yesterday suggested the following changes to the existing 65-team format…

1. Eliminate the 65th- vs 64th-seeded play-in game, which was originally played on Tuesday, two days after the last conference tournament championships on Sunday.
2. The top eight seeds in each of the four regionals get a first-round bye. Therefore, there are sixteen teams competing for the bottom eight spots in each regional. It follows that there will now be 24 teams in each regional. 24 x 4 = 96.
3. The new play-in games amongst the #9 thru #24 seeds will occur on Thursday and Friday, effectively pushing everything back one round, which leaves an extra two days of games to reach the Sweet Sixteen (one day for half the field). This extra round causes an overflow to be held the Tuesday and Wednesday immediately following the first weekend. That means the Sweet Sixteen will still continue as normally scheduled, still occurring on the following Thursday, Friday, Saturday, and Sunday. There is controversy surrounding the fact that a team that happens to traverse the bracket will have to miss almost an entire week’s worth of classes.

The issue I have is that, if you click on this NCAA-proposed 24-team regional bracket above (remember, there’s four of these regional brackets!), it becomes incredibly confusing for fans, completely altering the established fun of bracket-filling. Yep, now you have to fill out an additional thirty-two bracket pairings than before, or eight more bracket pairings per regional (8 x 4 = 32)! And the strategy does not do that much to eliminate the blowouts that we so often see amongst the #1 vs #16 and #2 vs #15 matchups.

IMPROVE, NOT INTRUDE

In the following descriptions, I’m going to loosely use the terms “NIT” and “play-in”, so that it’s a bit easier to follow. My idea centers around the notion that the NIT is the play-in for the NCAA, but officially the whole thing would be called the NCAAs and they’d probably eliminate the “NIT” moniker.

There is a better way that is less confusing and does a better job of eliminating the throw-away low-seeded blowouts. I’d actually even go further to say that an 88-team format is the absolute best, but I can understand the NCAA not wanting to eliminate 9 teams. 65 in the NCAA plus 32 in the NIT, in the present system, minus my 88-team suggestion means there would be 9 fewer teams in the NCAA+NIT. That’d be 9 more teams for the upstart/obscure CBI and CIT tournaments – not a good competitive business move by the NCAA on that front.

If it were up to me, the NIT would become the play-in for the last few seeds in the NCAA. After all, in the present system the NIT already starts on the Tuesday after conference tournament championship weekend. The underlying assumption for doing this is that East Tennessee State probably would have still lost to Dayton or UNC, the NIT finalists, and that most casual observers would rather see Kentucky play against a more recognizable name, such as a Dayton or a UNC, in the first round of a 64-team field, not an obscure team that happened to win a conference tournament for the automatic bid. In short, you’d get less of a fan revolt by explaining that you’re merely attaching the NIT field to the existing NCAA bracket, effectively letting at-large teams go to war over just the last 2 seeds per regional.

BEST ALTERNATIVE: 88 TEAMS

To the casual observer, the notion of the NIT champion being the 66th-best team this season would still hold. With my 88-team system, it’s less of intrusion into the 64-team field and more of an improvement of the bottom feeders. We’re not really rocking your world all that much, is what the NCAA should or would say.

So you start with 32 NIT-ish teams on Tuesday at the same 16 venues that make up the current 64-team field (4 teams per venue in the first round). By Wednesday morning, you’re left with 16 teams. On Wednesday night, the 16 teams play again to whittle the field down to 8.

The remaining 8 NIT-ish teams are given the #15 and #16 seeds in each of the four regionals. For example, it’d be #16 UNC vs #1 Kentucky and #16 Dayton vs #1 Duke. There’d be a much, much better chance of a #16 overthrowing a #1 than before. And a #2 beating a #15 would be far more likely than it is today. Quite simply, there would be fewer thirty-point blowouts in the #1 and #2 first-round games. From an NBA perspective, those are currently throw-away games because you can’t really continue to evaluate John Wall in a blowout. And does anybody still tune into those #1 vs #16 games?

Besides, an 88-team field certainly has a lot more marketing potential than 96. Eighty-eight is just a more aesthetically pleasing number than ninety-six.

96 TEAMS, IF YOU MUST

Alright, well let’s assume that the NCAA absolutely cannot let go of 96 teams, instead of my proposed 88. I just think that they haven’t thought about the mathematics of bracketing when they simply add the NIT’s 32 to the NCAA’s 64 (okay, 65) to get to 96. It’s so back-of-the-napkin right now. If you have 88, then you’re letting 8 teams go to the competition (CBI, CIT). I can understand the business decision, but I still think the elegance of the number “88″ and the fact that you are effectively making the #1 vs #16 and #2 vs #15 games more competitive outweighs the loss of 8 (okay, 9) teams, but that’s just me.

So then let’s expand my 88-team format to 96. It’s clunky, but it’s do-able. You end up with 44 NIT teams vying for the #16, #15, and #14 spots, instead of just the #16 and #15 seeds in the 88-team format. With 3 NCAA bottom seeds per regional up for grabs, that means you have 13 seeds locked and loaded. 13 x 4 = 52 and 96 – 52 = 44. That’s how we arrive at 44 NIT-ish teams.

With 44 NIT teams, that’s 11 per region. Here’s where it gets clunky because eleven is not a “nice” tournament number to work with. Again, you’re vying for the #14, #15, and #16 seeds in each region. As a bracket-maker and someone who determines the seeds, you have to assume that you were fair in your seeding and that all the top seeds will advance. Of course, it never ends up that way, but at least you covered your bases and did your best in fair seeding.

That being said, you must assume the #1 NIT seed in that regional will survive and be the best team from that field. It follows that with 11 NIT teams per regional, the assumed #1 team should have earned the right to play the assumed lowest-seeded team, the #2 against the next-lowest, and so on and so forth.

That results in a first-round NIT bye for the #1 NIT seed in each regional. Again, the sanctity of the top 64 slots is maintained, only altered with an appendage that determines the bottom 3 seeds in each regional, or the bottom 12 seeds overall. That means 52 seeds are as good as they’ve always been.

The play-in brackets are somewhat imbalanced if you look at the graphic (click on the image), which creates some scheduling difficulties down the line. All the first-round play-in games obviously will start again on Tuesday immediately following conference tournament Sunday championships. But when we get done with the play-ins, we’ll have three spots, not just two, which causes problems with the Thursday-Saturday and Friday-Sunday setups. I don’t have a solution for this at the moment, but an easy out would be just to extend to Monday and, if need be, Tuesday at the same venue.

By the next round, things go back to normal (Sweet Sixteen on Thursday, Friday, Saturday, Sunday of the following week). However, this is a Tuesday-to-Tuesday-then-Thursday scenario and casts a slightly wider net than the NCAA’s Thursday-Tuesday-Thursday strategy. Again, I’m sure there’s a way to get the clunky 11-team play-in bracket done by Monday or even Sunday, if they put their minds to it.

This also appeases the fans. They can, like in years past, ignore the NIT-ish play-in games and just concentrate on the remaining 64 teams, which was the same bracket on that Thursday, as before.

BELIEVE IT OR NOT, 112 TEAMS IS DO-ABLE

Now, let’s say the NCAA wants this whole endeavor to end up with a knockout punch of its competition, the CBI and CIT. Well, due to the format of the brackets, this is very do-able if they expand to 112 teams. I kid you not. It’s certainly much easier logistically than 96 teams, as you’ll soon find out.

In a 112-team field, we are now on a mission to replace the last 4 seeds in each regional — #13, #14, #15, and #16 — by putting those spots up for grabs with teams that would’ve been relegated to the NIT, CBI, and CIT in years past. When you lock and load the top 12 teams in each regional, that means you have 48 teams out of the original 64 field already sealed.

You could then have four more teams vie for each of the four bottom spots in each regional, or 16 teams in a play-in. 16 + 12 = 28 teams per regional, which is 112 teams for four regionals. It’s a rather straightforward and elegant solution. Each of the four bottom seeds, #13 thru #16, is determined by a mini-final-four bracket that whittles 4 teams down to 1, for each of those spots. On Tuesday, you have the four teams play two games. On Thursday, the winners play each other. By Thursday late night and even staying in the same venue (this year there were 16 venues, 4 teams or 2 games per venue, in the field of 64), you’ll have your #13, #14, #15, and #16 seeds determined.

Then on Saturday, that #16 seed plays the #1 seed. To reach the Sweet Sixteen, we need to extend the regional out until Monday. This is a Tuesday-Thursday-Saturday-Monday strategy, all at one venue. It follows that the other side of the regional, for logistics and TV purposes, can go with Wednesday-Friday-Sunday-Tuesday.

That doesn’t leave a whole lot of wiggle room for the concern over athletes being out of class, but if you think about it, if a low seed actually makes it through that whole week beginning on Tuesday, with the Sweet Sixteen looming as early as Thursday after that (two days of rest, in that you might as well head straight to the next venue without flying back home), it’s probably a once-in-a-lifetime Northern Iowa type of experience. The mid-major will take it.

On the other side of the coin, if you’re a #1 seed, you don’t play until Saturday, then probably Monday barring an upset. If you’re a high-seeded team, you actually stay in school longer that first week!

WHY FOCUSING ON PLAY-INS IS THE KEY

Basketball is basketball. Not only in all of my leagues and tournaments, but in leagues all over the world and at all levels, you have elite teams, very good teams, and the middle of the pack. You also have the cellar dwellers, but those are already weeded out here in the NCAA Tournament. It’s a standard bell curve. The difficulty is figuring out the middle-of-the-pack teams. And with the NIT and low-seeded NCAA field as we know them today, we are talking about the middle of the pack.

By focusing on just (replacing) the bottom seeds of the traditional 64-team format, you’re simply letting nature take its course. Better, you’re eliminating a difficult task of the Tournament Committee. You’re admitting that seeds #13 thru #16 might as well be determined by actual competition. And you’re not just cutting off regular season champions who happen to lose their conference tourneys.

The evidence is already there. #13 and lower seeds upsetting #4 and higher seeds just does not happen all that much. Worse, they are blowouts, more often than not.

I really like the expansion approach because of this bell curve, but I think the NCAA’s proposed solution is flawed and will only draw the ire of fans who can’t fathom how they’ll fill out their brackets now. And isn’t that the life-blood of this whole thing? That’s why the ratings are so high. The casual fan knows that a #5 seed is going to fall, so that’s why they tune in.

In my rec leagues, I’ve done 22-team quasi-double-elimination playoff formats in the past, so when analyzing a one-and-done bracket, that stuff is a piece of cake. Here’s an example of a 13-team quasi-double-elimination bracket that I did earlier this year (click to expand)…

You might also like:

• Perpetrators in dreamleague? Never!
• Jeremy Lin heads to “The Heart of High Country”
• Through NBA Eyes: 3 reasons why NCAA Tourney expansion is good
• It’s all on Portsmouth for Jeremy Lin, but…
• Hoops Underworld: NCAA vs NBA
• What Jeremy Lin’s up against at Portsmouth
• CousyAward: Who are these guys?
• Meet Mr. Crunch-Time: Jeremy Lin