Categories

Friday, February 12, 2021

It's About As Difficult to Overcome an 8-0 Deficit in NCAAB (3 scores) as a 7-0 Deficit in NCAAF (1 score)

Starting slowly (and going down by a large margin early) greatly reduces your chances of winning the game - the question is by how much. Having the rest of the game to claw back is generally overvalued - for example, an analysis from Harvard illustrated how in the NBA, winning the first quarter is the most impactful on going on to win the game

The concept of momentum is also greatly overvalued - something I found to be true in both NCAAB and NCAAF. So if momentum isn't real, but getting out to an immediate lead increases your chances of winning the game, then it follows that just banking those points is enough to push your win probability higher, substantially. In other words, your opponent now has to outscore you by the amount they're already losing over the rest of the game.

But where do these deficits intersect for NCAAB and NCAAF? For example, is it harder to come back from a 6-0 deficit in NCAAB or a 14-0 deficit in NCAAF (2 scores in each case)?

I took the data from the above momentum exercises (2016-17 for NCAAB, 5,461 games; 2015 for NCAAF, 687 games) to find out how often a team won when starting the game on a certain run. I then limited the exercise to scores that occurred in > 1% of games.

NCAAF% GamesHeld LeadNCAAB% GamesHeld Lead
325.4%52.3%14.1%53.1%
769.3%74.0%268.8%56.6%
1015.0%77.9%329.2%57.4%
1423.2%91.9%418.9%62.3%
178.0%85.5%514.0%68.4%
203.6%92.0%68.3%70.4%
2110.2%97.1%83.9%73.5%
243.9%96.3%101.7%80.9%

Due to the higher volume of possessions per game in college basketball, it is much easier to overcome a larger deficit (i.e. being down 10-0 is being down 4 scores, which would be equivalent to 28-0 in football). NCAAF's slope is much steeper and gets closer to 100% much faster, whereas even a double digit lead only gets you slightly above an 80% win percentage in NCAAB.



An interesting quirk (not shown above): only 5 NCAAF games (0.7%) started out 2-0 on a safety, and the leading team went on to win 4 of these (80%).

But where do these two meet?

NCAAF, Initial Lead
Win % Delta37101417202124
NCAAB,
Initial
Lead
10.9%-20.8%-24.8%-38.8%-32.3%-38.9%-44.0%-43.2%
24.3%-17.4%-21.3%-35.3%-28.8%-35.4%-40.5%-39.7%
35.1%-16.6%-20.5%-34.5%-28.1%-34.6%-39.8%-38.9%
410.0%-11.7%-15.6%-29.6%-23.1%-29.7%-34.8%-34.0%
516.1%-5.6%-9.5%-23.5%-17.0%-23.6%-28.7%-27.9%
618.1%-3.6%-7.5%-21.6%-15.1%-21.7%-26.8%-26.0%
821.2%-0.5%-4.4%-18.4%-12.0%-18.5%-23.7%-22.8%
1028.6%6.9%3.0%-11.1%-4.6%-11.2%-16.3%-15.5%

The closet win percentage that matches in each sport is around 74%, which is equivalent to an 8-0 lead in NCAAB (3 scores) and a 7-0 lead in NCAAF (1 score).

Outside of a 3-0 lead in college football, almost every other lead is safer on the gridiron (7-0 and above) than any counterpart in college basketball.

Saturday, February 6, 2021

Carolina's Hardest Game Each Year Isn't Always Duke on the Road

Conventional wisdom (via a poll, sample size = 2) would think that the annual Duke game at Cameron Indoor is the single hardest game of the season for Carolina, with "hardest" defined as "least likely to win going into it". The easiest proxy for this is the closing spread going in to each game, looking at the largest margin(s) in which UNC was the underdog (which is equivalent to their lowest chance of winning). So I grabbed this data from Covers as far back as it goes (14 seasons).

That away game at Cameron was the hardest in 7 of the past 14 seasons, 50% of the time (I personally would have expected this to occur more often). The single lowest point was a spread of -15, at Duke in the 2009-10 season, which is roughly equivalent to a 7% chance of North Carolina winning. To the other extreme, Carolina's lowest win chances actually came as favorites in 2006-07 (60%) and 2008-09 (58%) - meaning they were favored to win every single game.


On average, the away game at Duke ranks as the 1.7th hardest on the schedule, and the home game in the rivalry ranks 6.9th.

Conference / non-conference splits shake out as you might expect, with the following average difficulty ranks:

  • Conference Games
    • Home: 24.6
    • Away: 13.5
    • Neutral: 14.3
  • Non-Conference Games
    • Home: 16.6
    • Away: 8.1