Defying the odds: The staggering math behind Texas A&M's miracle comeback against UNI

We've all had 24 hours to deconstruct the greatest comeback in College Basketball history, which has led to a few fascinating posts surrounding the statistical oddities of last night's game.

****************************************************

The Depths

First, we start with a win probability calculator that had the Aggies at a 0.01% chance of winning before Alex Caruso missed his three pointer in the last minute of regulation.

The formula then attempted to kick the probability down that final .01%, but the graph wasn't having it.

Here it is. 99.99%. This is an all-time sports collapse. Bar none. pic.twitter.com/A3f1iUlkXW

— numberFire (@numberFire) March 21, 2016

Which, to me, meant that the graph had become a sentient being with one of two personalities:

1. A graph that angrily shouts things like WE AREN'T GOING TO SAY 100%. DO YOU HEAR ME? WE AREN'T. PROBABILITY DOES NOT DEAL IN ABSOLUTES.
2. A graph that didn't know what to do with this information. That, at its core, had already told the user who was going to win the game and did not know how to display that information any more clearly.

Either way, the message was quite direct. This basketball game has a winner.

The Analysis

We found a good piece at @FiveThirtyEight that laid out the win probability at various stages of the comeback. It's a great read that does a really good job placing this game in a historical context.

Via @FiveThirtyEight a look at @12thMan (Texas A&M) comeback from a win probability perspective pic.twitter.com/o1TB9Iv8kA

— ESPN Stats & Info (@ESPNStatsInfo) March 21, 2016

Texas A&M pulled off a 1-in-3,000 comeback: https://t.co/wTt9WoXQPO pic.twitter.com/4evFUu7dLg

— FiveThirtyEight (@FiveThirtyEight) March 21, 2016

We struggle with the "1 in 3000" line, however. They may have missed a couple zeroes.

Wasn't this the greatest comeback of all time? And doesn't a typical college basketball season contain 300+ D1 teams playing 30 games each?

The very nature of those two statements seems to imply that this comeback had a much, much lower likelihood... but we still recommend the read. Great stuff.

Destroying A Legend

Now we turn our attention to Bill James, who is widely recognized as a foremost authority in advanced metrics. James' early work was the rock for the Moneyball movement in professional baseball, but he has since developed a formula that could determine a true "safe" lead in college basketball.

We broke it.

We broke his formula.

One of the most respected minds in the field spent his own time and energy to develop a formula for this specific purpose, and we cast it aside with the grace and elegance of a toddler destroying a ten million dollar vase.

Per The Bill James Lead Calculator, Northern Iowa's lead was 100% safe with 35 seconds left. Now in OT against A&M pic.twitter.com/TUoKjhmvnY

— Dan Feldman (@DanFeldmanNBA) March 21, 2016

WE CARE NOT FOR YOUR OPINION, EXTREMELY RESPECTED SCIENCE MAN!

Texas A&M's comeback was so epic that it broke Bill James' formula for determining when a lead is unassailable: https://t.co/9l3p5M3Iov

— Rany Jazayerli (@jazayerli) March 21, 2016

The Overall Picture

Finally, we take a look at the full picture beyond that miracle comeback. Just look at those swings in the first overtime!

There she is: Texas A&M’s win probability graph, via @kenpomeroy. A sad descent, then the jolt of madness kicks. pic.twitter.com/EGpdWK87aX

— Matt Norlander (@MattNorlander) March 21, 2016

We had our backs against the wall again, and we made the plays we needed.

****************************************************

We don't think we'll ever get tired of talking about this.

BTHO OU