THAT'S WHY THEY PLAY THE GAMES.

Five years ago, neurophysiologist Gregg Nigl cobbled together a bracket, then went on a ski trip.

Nigl always filled out a few men's brackets, deploying the same basic philosophy that so many people use -- a little strategy, a little eye test, a little wishful thinking.

For his 2019 brackets, he went heavy on his favorite team (Michigan) and his favorite conference (the Big Ten, except for Ohio State). He had always liked Gonzaga and hoped the school would win a title sometime, so he rode the No. 1-seeded Zags to go all the way. He even sprinkled in the time-honored tradition of picking a random school he had a loose tie to; in this case, he knew someone who lived near the UC-Irvine campus, so he went with the 13-seed Anteaters to upset Big 12 champ Kansas State in the first round.

We're not talking about purely random events, and people filling out brackets can rely on their information, their biases, who has the cutest cheer team, and yet, getting all 67 games right isn't easy.

The possibility of getting every game right is often reported as 1 in 9.2 quintillion. But that figure is slightly hyperbolic. It's computed by assuming all 63 games are coin flips, when, in reality, quite a few NCAA games have much higher percentages in favor of one team. The actual odds of a knowledgeable person picking a perfect men's bracket are closer to 1 in 120 billion.

And guess what? It's not getting any easier. Despite advances in analytics, accessibility of watching games and more experts than ever, the variance of men's college hoops has never been higher. The rise of the 3-point shot has played a huge part, for sure, but so has the transfer portal and NIL. A No. 1 seed hadn't ever lost in the first round of the tournament, then it happened in 2018 and 2023. A 15-seed has won in the first round three straight years, a first in NCAA tournament history. And last year's men's tournament featured a Final Four of no 1, 2 or 3 seeds, which hadn't happened before.

San Francisco State professor Paul Beckman did a study of every men's and women's bracket since 2000, primarily relying upon seeding of teams that made it to the Sweet 16. He found the women's tournament to be slightly more predictable (the odds of picking a perfect women's bracket are still something like 1 in 100 billion) and that the men's tournament has been wildly variable, even from year to year. For example, from 2018-20, the men's tournament Sweet 16 team total seeds went from 85 to 49 to 94. Even the predictability of the tournament is unpredictable, let alone picking individual games.

I'm still not sure how the pools handle those first four games, and with those pairings sometimes involving middling seeds, the rules for treating those teams as wildcards must be arcane.

Filling out a winning entry in your office pool, though?  It's pure backwards induction.

Illinois professor Sheldon Jacobson has spent 20 years studying the NCAA tournament, and he has come to several interesting conclusions. On his website, Bracket Odds, he suggests working inside-out when you fill out a bracket, sorting out the Final Four or Elite 8 and going backward because most pools will be won by the people who rack up the biggest points from the end of the tournament rather than the beginning. "You can do terrible the first two days of the tournament and still win if you hit the last few games," he says.

Jacobson has found that the single best factor is also the most obvious: seeding. As wild as early-round games are, and as fun as it can be to see a George Mason or Saint Peter's come out of nowhere to make a run, the best strategy over time is to fight the urge to nail the 15-seed who'll win one game and simply load up on the highest-seeded teams.

You can't argue with The Science, can you? Mr Nigl became guest of the NCAA to watch Michigan play in the round of sixteen because his bracket was perfect through the first two rounds, 48 games (again, there's that convention for those first four wildcards?)  He finished with 53 of the 63 games correct in his pool, but because of the way the pools reward getting the finalists and the winner right (reason ahead, then apply backward induction) he might not have won that pool.  To get the champion right requires four correct picks, to get the Final Four correct, another three correct picks, and his entry had five correct picks in the final rounds.  It's about getting the incentives right: to get the round of eight right requires 25 correct picks.  On the other hand, he got some quality time with his son, which is a pretty good outcome for a bracket he filed whilst groggy on cold medicine.

If anything, the way the brackets break suggests there is parity in the college game.

Jacobson thinks the Selection Committee does an overall bad job with seeds. He says the data shows that the top and bottom seeds -- 1s and 2s, 15s and 16s -- are usually spot on and that the next tier (3-4, 13-14) are very close every year. But he thinks the committee routinely whiffs on 5-8, usually underseeding non-power conferences. His favorite value pick is the 11 seeds, who have gone 19-17 overall since 2014. The 11s that advance are then 26-32 in the second round against either the 3 or 14, creating a surprisingly frequent road into being an 11 seed in the Sweet 16. "I'd rather be an 11 than an 8," he says. "I think the path is better."

Conventional wisdom used to be to go against the boffins with the twelve against the five, but with parity among the teams, the notional eleven is probably closer in overall ability to the six; whilst those sevens and eights often go to the also-rans in the so-called power conferences.  Sometimes that gets a Wisconsin to the Final Four, but then, Arizona eliminated in the second round used to be a truism when you fill out your bracket.

On the other hand, when the backwards induction starts from the wrong premise, your entry is going to be an also-ran, quickly.  "Only about 5% of brackets correctly picked Oakland's upset win over Kentucky on Thursday night."  Anybody who had Kentucky in the round of eight has to hope for the other seven to come through to have any chance at the big prize.  That's the seven-ten split, isn't that a bowling challenge?

Last night, it was better to be the eleven.

The biggest drop-off of the day came after No. 11 Duquesne pulled off a huge 71-67 upset win over No. 6 BYU. The win was the first in the tournament by the Dukes since 1969. A little more than 75% of brackets had BYU advancing out of the first round, so that game single-handedly knocked out most users. Duquesne will take on No. 3 Illinois next in the second round after the Illini flew past Morehead State.

As today's games tip off, all of 116 brackets are still perfect, with the five-twelves about to take place.  Northwestern ousting Florida Atlantic in the early Friday game, a modest nine-ousting-eight,  further reduced the count of still perfect brackets.  In the women's tournament, one eleven (Middle Tennessee) ousting a six (Louisville) wrecked almost all the brackets.  That's what equity parity looks like.

SECOND SECTION.  As of the conclusion of the late afternoon Friday games, there are no perfect brackets left in the men's tournament.

No. 9 Northwestern's win over No. 8 Florida Atlantic whittled the field down to 60 [perfect brackets] in the opening game on Friday. The rest of the early slate went chalk. No. 3 Baylor made easy work of Colgate. No. 2 Marquette fended off an upset bid by No. 15 Western Kentucky. And No. 5 San Diego State held on to beat No. 12 UAB.

No. 6 Clemson's win over No. 13 New Mexico and No. 1 UConn's win over No. 16 Stetson Friday afternoon further reduced the perfect-bracket tally to 13.

Then No. 10 Colorado held on and beat No. 7 Florida 102-100 on a last-second shot. The Buffaloes nearly blew a 13-point lead late in the second half, but pulled off the win after a clutch shot from KJ Simpson. Almost 73% of users picked Florida to advance. That knocked out 12 of the 13 remaining perfect brackets.

A few moments later, No. 13 Yale rallied back to pick up just its second NCAA tournament win in school history. The Bulldogs fended off No. 4 Auburn 78-76 in a chaotic finish in Spokane. 90.4% of users picked Auburn to make it through the first round.

That eliminated the final perfect bracket, ending this year's quest before the first round ended.

In the women's tournament, a lot of punters picked against Duke and North Carolina, and when both of those won, pop went a lot of those brackets as well.  A prior of "them that has, gets" still works in betting on that tournament.