Data behind what leads to a coach's breakout season
|Aug 19||Public post|| 5|
The week in review
It was a busy week for X’s and O’s over on Twitter. We featured video on:
Two different parts of the Carolina Break
Two different zone sets
One Duke horns set with different dribble handoff options
Also, the newsletter got some love this week!
First, Jeff Greer used the July 14th edition of the Hoop Vision Weekly (“Quantifying Defensive Scheme”) for a great article on Louisville’s pack-line defense.
Next, CJ Moore used last week’s edition (“Post Pins & Bill Self Offense”) for another great article on Kansas’s hi-lo motion offense.
Go check those out, and thank you for the support!
In today’s edition, we take a data oriented look at what goes right in a coach’s “best season”.
What are the ingredients to a “banner” season?
The middle of August is fairly uneventful in the world of college basketball. A couple teams are usually losing to Carleton in Canada. A couple other teams are usually playing in a 100 degree gym somewhere in Europe.
But other than that, August is a time for optimism about your team. Surely this will be the year all of the pieces come together.
Inspired by this time of year, we decided to take a statistical look at the teams that actually did put all of the pieces together.
First, we identified the 353 head coaches who have participated in at least five D1 seasons in this decade. Then, we identified each of those coach’s best individual season of the decade - by KenPom’s adjusted efficiency margin.
With each coach’s best season identified, now we can take a look at what exactly goes right for those special teams - relative to the coach’s overall decade average.
To move this from the abstract to a real example - consider Tony Bennett. He has coached all 10 seasons of this decade at Virginia. In 2019, he produced his best adjusted efficiency team of his career.
Was there something dramatically different stylistically between 2019 Virginia and the previous seasons? Or was this simply the best Virginia team at executing the Tony Bennett style?
Ask these questions on a more macro level for all 353 coaches and we have some interesting data to investigate.
Good teams (to no surprise) improve their four factors
The four factors (shooting, turnovers, rebounding, free throws) are really just explainers of efficiency. It would be essentially impossible to improve offensive efficiency without also improving at least one of the four factors.
So it should be no surprise that during a coach’s best efficiency margin season, the four factors also moved in the right direction. Here are how they changed offensively for the 353 coaches.
How to read: The average eFG% in each of the coach’s best season was 51.5%. Those same coaches averaged an eFG% of just 49.4% in all other seasons.
As we would expect, all four moved in the right direction (a decrease in turnover rate is actually the right direction for an offense). In terms of raw numbers, 298 of the 353 teams shot better (eFG%) during their best season relative to their coach’s average.
For posterity’s sake, here are the changes in defensive four factors.
Again, all four statistics moved in the right direction for the defense.
Now that we understand the methodology, let’s get to the interesting stuff. Obviously the four factors improve during good seasons, but what about statistics that aren’t intrinsically good or bad?
We will take a look at possession length, three-point volume, blocks, height, bench minutes, and experience in the sections below.
Everybody wants to play fast. Should they?
We confirmed via multiple sources earlier this offseason that 2019-20 season will be the fastest in history.
But do the best teams play above the normal tempo of their coach? Again, we can use the same methodology as before to investigate.
During a coach’s best season, his team’s offensive possessions are about 0.2 seconds faster than normal. For perspective, the fastest team in the country last season clocked in at 14.1 seconds (FIU) and the slowest team in the country clocked in at 21.5 (Siena).
The small 0.2 second increase in tempo aligns with intuition and is likely a function of just overall offensive strength - as opposed to an increased emphasis on tempo.
Let’s say a team with good offensive players and a team with bad offensive players are running the same exact scheme with the same tempo-related philosophy. Even without an emphasis on tempo, the team with better players is likely to find acceptable shot opportunities quicker.
While ultimately these things should be evaluated on a case-by-case basis, it’s safe to say that a decision to play faster isn’t a magic bullet to success.
Do the best teams shoot more threes?
I’ve been on record several times in this newsletter with a stance on three-point shooting. Yes, three-point shooting and floor spacing is a good thing for an offense. Not only is the extra point very important (#math), but strong floor spacing helps unlock the rest of the offense.
The thing that doesn’t get discussed enough at the college level is the overall player pool of three-point shooters. Simply put, it’s not an easy task to find efficient three-point shooters that aren’t below average at nearly everything else.
That being said, here’s another opportunity to look at three-point shooting and team success. Do the best teams shoot more threes?
The answer here is no. About 35% of shots came from threes in normal seasons. In the best seasons, three-point volume stayed pretty much the same.
But it’s the stat to the right, block percentage, that really stands out. In the best seasons, block percentage rose from 9.2% to 10%. That increase is in line with some of the four factors we saw earlier.
It’s not that three-point shooting is bad or even neutral. It’s that teams with the ability to take a high volume of threes are often compromising things like rim protection, rebounding, and just general athleticism.
Remember, correlation does NOT imply causation. But if your team has an impact rim protector (or two) on the 2019-20 roster, we won’t hold you back from getting excited.
Score one for senior leadership
To finish, let’s take a look at three more team statistics that aren’t necessarily intrinsically good or bad: Average height, bench minutes, and experience.
(Stats like these all come from the invaluable kenpom.com).
The best teams, on average, had a shorter bench
The best teams, on average, were older
For those unfamiliar with Ken’s experience metric - the units are years of experience. In other words, the best teams were had an average of 0.22 more years of experience than normal teams.
A coach’s best team tends to be older with a shorter bench. Again, that doesn’t mean a team like Kentucky can’t win with a bunch of talented freshman (of course they can). But for the average coach, the best teams have skewed older.
Were you forwarded this email by a friend, colleague or coach? If you enjoyed it and would like to receive original research, insider access, and strategic analysis of college basketball on a regular basis, please tap/click the button below to subscribe to Hoop Vision Weekly.