As most Wake fans know by now, the loss to Virginia on Tuesday night was heartbreaking. What most probably didn't know is that the game was "statistically" safe by the best metric known to assess when a game is over before the final whistle.
Wake Forest was up by 7 points with 20 seconds left with the ball against the Hoos. What transpired from there we all know. Virginia came back and won 72-71 on the unlikeliest of buzzer beaters by Darius Thompson to end the game.
I'm not here to talk about missed free throws, bad defense, bizarre threes, or anything that is actually basketball-related. I am here to discuss the unlikelihood of that comeback.
I got into this a bit in the comment sections yesterday, but the lead, by nearly all intents and purposes was "safe" from a statistical vantage point.
I take the following information from our brother-site Streaking The Lawn, who ironically enough, lost a "statistically safe" game in the ACC Tournament to Miami in 2011. They were up by 10 with 42 seconds left and eventually fell in overtime to the Hurricanes 79-272.
So if there is any poetic justice, it is that UVA fans have in fact felt what Wake fans felt on Tuesday. To add insult to injury, their year was ended on it.
Alas here is the important formula that is used by Bill James:
Applying this to Wake below:
So, how rare is it for a team to lose once the game is considered statistically safe? In 2008 Statsheet used the Bill James Formula to test close to 10,000 game results of college basketball.
How many failed????
Exactly 1 failed, when UNC came back from 8 points down with 17 seconds left to stun Duke in 1974. That was before the three-point line, and was, and still is, a shocking result.
UVA-Miami was the second one to break the test in 2011.
Canisius was down 11 earlier this year with 1:02 left to UL-Monroe and wound up winning (by 12 no less) in triple-overtime. That also broke the Bill James test.
I have listed the only 4 occurrences that I could find this actually happening since 1974. To dive a little more into how unlikely this result is I will throw out some more "fun" observations.
I will yield that I perhaps missed a few, and say that it has actually happened between 5-10 times (let's call it 7 times). Assuming that there have been an average of over 150 D-1 basketball teams playing 30 games a year for the past 50 years. That is a sample-size of over 225,000 games played spanning back to 1966 (those numbers may be slightly off, but this is a practice in futility, so let it be).
Since I am yielding seven games have actually broken the model, that is 7 out of 225,000, which works out to approximately .0000311 percent of games played. What a treat that we got to witness this! That also works out to 1 out of every 32,143 games played will we see a game that breaks the Bill James rule.
Even more bizarre is that two of these games took place in the span of a month. When the model should yield one game every seven years, that is truly interesting to look at from a statistics standpoint.
If we were to see how long (in theory) it would take Wake Forest to play in one of these games given the current rate of games played, it would be approximately once every 1,071 years if we assume 30 games a year for our beloved Demon Deacons.
So to all of the Wake fans that I have seen say "this happens from time to time if you are a Wake fan", it truly, and utterly does not. At least not to this magnitude. Games like this don't happen to pretty any team, any where, and at any time.
There are several interesting things about this though:
- Virginia has played in two of the four games, winning one and losing one.
- There have been two such games in the past 35 days after only two such games in the past 42 years.
- WE GOT TO WITNESS SUCH A HISTORICALLY STATISTICAL EVENT, WHICH IS SO EXCITING!!!! #WeAreAllWitnesses