Starting the year strong (Hoop Vision Weekly: 1/1/21)
After eight games, Baylor is shooting 46% from three. What does that tell us about the rest of the season?
Welcome back to the Hoop Vision Weekly and Happy New Year!
First off, we want to give a big thank you for all of your support of Hoop Vision in 2020. In a year of many ups and downs (well… mostly downs) and uncertainty, we were fortunate to still take steps forward in the process of turning HV into a sustainable business.
We couldn’t have done that without our readers’ support, even during times when there was no basketball being played. Thank you!
Due to the holidays, it’s been 14 days since we sent out the Weekly, but we’ve still been busy with other content.
We published two new voiceover videos on our YouTube channel:
We also sent out our weekly edition of The Starting Five newsletter to Hoop Vision PLUS subscribers. Including a nice little rant — inspired by this Richard Pitino tweet — on why counting passes isn’t a particular insightful statistic.
Join Hoop Vision PLUS ($10/month or $100/year) and receive The Starting Five newsletter straight to your inbox every Monday during the college basketball season.
In today’s edition of the HV Weekly:
Baylor is shooting 46% from three in their first eight games. What does that tell us about the rest of the season?
Indiana’s unique ball screen coverage — and a general framework for evaluating defensive coverages
Links from around the internet
Baylor’s hot three-point shooting
Due to COVID cancellations, Baylor has still only played one top 100 kenpom opponent this season — a 13-point win over Illinois. Scott Drew’s team has won their other seven games by an average of 35.6 points.
But even though the schedule has been weak, Baylor’s first eight games have been eyebrow-raising. Last year’s Baylor — a team that was headed for a one-seed in the NCAA tournament — succeeded with elite defense and offensive rebounding. This year’s version remains strong in those areas, but also currently leads the entire country in three-point percentage.
Of course, Baylor’s offensive statistics are due for some regression in Big 12 play. They won’t shoot 46% from three for the rest of the season.
But even if Baylor’s true three-point percentage is in the 38-40% range (they shot 35% last season), there’s reason to believe — given the track record of their defense and rebounding — that the Bears are right in line with Gonzaga for top team in the country. So the question here is…
What do eight games really tell us about three-point percentage?
Let’s start with a couple recent examples.
After eight games last season, South Dakota led the country in 3P%. The Coyotes were shooting 47.4% from behind the arc. In their final 24 games over the rest of the season, South Dakota shot 36.7% from three.
The season before that had an even more extreme example. After eight games in 2018-19, Indiana State led the country in 3P%. They Sycamores shot 48.2% in their first eight games, but below the NCAA average in their final 23 games at 32.7%.
The table below contains the 10 teams in the kenpom era (2002-2020) with the hottest three-point shooting results to start a season.
Just three of the 10 teams remained over 40% for the rest of the season. The average rest of season three-point percentage was 36%.
At face value, that’s not the best news for Baylor.
However, the three teams that finished the rest of the season under 33% all have something in common: They were low volume (3PA/FGA) shooting teams.
That makes sense for a couple reasons…
Small sample size. Eight games is a small sample size to begin with. But for teams that don’t take many threes on a per game basis, it’s an even smaller sample.
Low volume adds valuable information. In theory, an offense shooting well over 40% from behind the arc should be looking to take a lot of threes. If not — particularly in today’s game — it’s probably an indicator that the offense doesn’t actually feature very strong shooters. Indiana State’s early season success in 2018-19 is the best example of this.
(By the way, 39% of Baylor’s shot attempts have been from three this season.)
If we change the criteria of the previous “10 hottest offenses” list to “10 hottest high-volume offenses”, the rest of season average increases from 36.0% to 37.6%.
The graph below contains the correlation between the first eight games and the rest of the season for every team in the kenpom era with a 3PA/FGA over 35%.
I’ll set the over/under for Baylor’s rest of season 3P% at 37.5%.
Indiana’s unorthodox coverage
Earlier this week, Zak Boisvert uploaded a video on Texas’s corner step-ups.
On ball reversals, Texas likes to set ball screens towards the baseline instead of the middle. But something else caught my eye in Zak’s video (and immediately had me rewatching a game from a month ago): Indiana’s ball screen coverage.
The Hoosiers were “weaking” Texas’s step-ups — sending the Longhorns to their left hands. Normally, weak coverage is reserved for screens taking place in the middle of the floor — rarely for ball screens occurring on the wings.
(Carleton’s had a lot of success breaking the rules though.)
Here’s the best example of Indiana’s “weak” coverage, with #1 Al Durham taking away the right-hand baseline drive.
Archie Miller couldn’t ask for a better result than that. #3 Courtney Ramey tried forcing the issue to the baseline, settling for a very difficult jumper.
The coverage looks great when executed successfully, but there are two potential problems here:
Difficulty of execution. Because of the location, angle, and pace of those step-up screens, it can be very hard for the defender guarding the screener to get into proper positioning.
Necessity of execution. Obviously every defense wants to execute every single coverage, but mistakes happen. An issues with weaking step-up screens is that a failed execution is particularly punitive to the defense. It’s hard to get away with a failed coverage.
This next clip illustrates these two points.
Unlike the first example, Texas has better pace getting into the step-up screen. As a result, #21 Jerome Hunter is forced to go immediately from closing out on the corner into defending the ball screen. That’s a difficult path for Hunter to get all the way into the middle of the court.
Here’s a screenshot from the clip of Hunter on the wrong side of the screen.
Durham is weaking the ball middle, but Hunter isn’t in position to give help — leading to the slot three-point attempt that pack line defenses tend to allow on middle drives.
When evaluating a defensive coverage, it’s important to also consider what happens when the coverage doesn’t work. The three general variables associated with a coverage are:
Probability of successful execution
Points per possession allowed given successful execution
Points per possession allowed given failed execution
A high risk/high reward coverage (like blitzing) will be very good at #2, but vulnerable to #1 and #3. A lower risk coverage (like dropping), will be the opposite.
Other Links
Oscar Tshiebwe is leaving West Virginia. An unfortunate loss for the bully ball experiment
Adrià Arbués hosted an interesting “Beyond the Four Factors” talk with three prominent member of the NBA analytics community
Thanks for reading the Hoop Vision Weekly!
Become an HV+ subscriber today ($10/month or $100/year) to receive more Hoop Vision content — including The Starting Five newsletter — throughout the college basketball season.
Your support helps us keep Hoop Vision independent and ad-free!