Kentucky Wildcats Newsletter #1
Newsletter #1
Who's Irreplaceable? On and Off the Court Impacts
January 28, 2017
When I started tracking play-by-play data for college basketball a few years ago, I had a goal of measuring all the things that happen in a basketball game and who was on the court. With that data, I hoped to determine how teammates impacted each other's play, and measure how teams played when various players sat out. One of the first things I realized was that there was a lot of noise in the data, and teams didn't always substitute one player at a time. So, if 2 players went to the bench at the same time, and the team all of a sudden played worse, how do you know which player made the bigger difference? I realized I was going to need to isolate the impact of each player, so I started looking at how a team performed when you kept 4 of the 5 players the same, but swapped out that 5th player. I now can do this for every lineupa team uses, and every possible 4 man combination. I look at how a team plays when Player X is in the game, then I look at every 4 man combo that played with Player X, and how those combinations performed when Player X was on the bench and another player took his place. I separate this into 3 stats: points scored per possession, points allowed per possession, and plus/minus per possession (which is points scored minus points allowed). I then adjust the data for each game to account for quality of opponent, and you have what I call Same-Four Plus/Minus. Before I show you the numbers for Kentucky, a couple notes:
-Based on my experience over the past couple years, the data can fluctuate by .05-.10 points per possession until you get about 500 possessions with and without the player. So, any difference of .05 or less is pretty insignificant.
-This doesn't necessarily rank "best" players; instead, it is a reflection of which players are hardest to replace on the team. If a player has weak backups, he'll look good by this measure, and vice versa. If one player rates as a +.10 and another as +.05, that doesn't mean the first player is better, just that he is a bigger upgrade over alternatives. This does mean that his team will likely suffer more if the first player misses time due to injuries or fouls, however.
OK, so now you have some background on Same-Four Plus/Minus. So, what does the data say about Kentucky so far this year?
UK's Same-Four Plus/Minus stats through the Tennessee game on Jan 24th. Monk, Fox, and Adebayo make the most difference for UK, probably to nobody's surprise. Briscoe trails well behind his fellow guards, and has been a net negative so far.
The table above is probably a little tough to understand if you're seeing it for the first time, so let me walk you through it from left to right. Up first is De'Aaron Fox...he has played 1,153 possessions so far this season. If you look at all of the 4-man combinations he has played with, those groups have played 357 possessions this year with someone other than Fox. During those 1,153 possessions Fox has played, UK has scored 0.07 points per possession more than the 357 possessions the "same four" combos have played without Fox. When comparing points allowed per possession, UK has allowed 0.11 less in Fox's 1,153 possessions than in the 357 "same-four" possessions. Adding the offensive and defensive impacts together, UK has outscored opponents by 0.17 points per game more (with some rounding) with Fox in the game than when the same four players play with anyone else. When you go through each possession and adjust for the rating of the opponents, we subtract .02 from the difference because the possessions where he's sat out have been against a little tougher teams. So, the Same-Four Plus/Minus for De'Aaron Fox is 0.15 points per possession. This means that De'Aaron Fox has made UK 0.15 points better in each possession he's played, instead of whoever UK would play as an alternative. Given that he played about 60 possessions per game, he's making UK about 60 x 0.15 = 9 points better per game.
You can see from the far right column in the table above that Malik Monk has the best rating, at +0.16, with Fox at +0.15 and Adebayo at +0.14. For Fox and Adebayo, the difference in points allowed is bigger than points scored, so they are providing more value defensively over their replacements. Monk is making a bigger difference on offense, although all 3 help UK score more and allow fewer points per possession. Wenyen Gabriel is also a net positive, although his impact is all defensive; UK actually scores fewer points with him in than when someone else fills his spot. Derek Willis is a positive on offense but negative on defense, and roughly neutral overall. Isaiah Briscoe has a negative impact on UK's offense but neutral on defense, while Dominique Hawkins is the opposite.
Some players have large adjustments between their raw +/- differential and their Same-Four Plus/Minus. This happens when there is a significant difference in opponent quality between the possessions they play, and the possessions their same-four combos play. Isaiah Briscoe's raw +/- difference is -0.12, but adjusted is -.06. This is because he missed games against Cleveland State and Tennessee-Martin, 2 of UK's weaker opponents. So, his stats have been put up against better teams, and shouldn't be expected to outscore them as much as Cleveland State or Tennessee-Martin.
So, we have some evidence that UK is better with some players in the game than others. Let's look at two UK players, Mailk Monk and Wenyen Gabriel, and dive into the stats that can explain why they are having a sizeable positive impact on UK's play.
Malik Monk
The Same-Four Plus/Minus data indicated that Malik Monk has a positive impact on UK's offense and defense, with a bigger impact on offense. The above table goes a level deeper, attempting to pinpoint exactly how he's making an impact. So, what does this data tell us?
First, UK shoots much better when Malik Monk is in the game, both on 2's and (especially) 3's. This isn't a surprise, since Monk himself shoots very well on both 2's and 3's. Even though UK shoots much worse on 3-pointers with Monk out, the team actually shoots 3's more often with him out (33.0% of FGAs, vs 30.5% when he is in). Given that UK only hits 27% of their 3's with Monk out, that's a really bad idea. So, UK shoots worse, and shoots more threes when Monk is out...but why?
The answer lies in the way UK is running their offense with Monk out. When Monk is out of the game, UK still shoots well on putbacks and fast break shots (the easy stuff), and gets them about as often as when he's in. However, UK really struggles when they have to run their halfcourt offense without Monk in the game. UK doesn't take any longer to get shots, but their shooting drops substantially because Monk is one of the few Wildcats who can create his own high-percentage shots. On possessions that last between 11-20 seconds, UK shoots 57% with Monk in and only 47% with him out. On possessions taking longer than 20 seconds, UK shoots 52% with Monk in and 40% with him out. Monk's ability to create (and hit) shots in the flow of the offense isn't something that other players on the team do very well, so the offense suffers without him. On a positive note, UK doesn't turn the ball over much more when Monk sits, and they draw free throws at about the same rate.
On defense, Monk's impact is most notable in his ability to hound other team's guards without fouling. While teams shoot about as well with Monk in or out, team shoot fewer three pointers with him in. This is likely driven by Monk's athleticism and ability to close out on shooters. Teams also turn over the ball more with Monk in the game (21% vs 18% with him out). These two stats indicate that Monk is aggressive on defense...but another stat shows us that he's smart about it. Opponents shoot free throws way more often with Monk out of the game (43.7 FTA/FGA, compared to 30.8). Given that teams tend to hit free throws at a much higher rate than other shots, allowing more free throws is a recipe for allowing a lot more points...and that's what's happening.
So, Malik Monk's ability to create shots within the flow of the offense helps UK shoot much better with him in, and his ability to be aggressive on defense without fouling keeps opponents from drawing free throws. Combined, these skills help UK outscore opponents by significantly more with Monk in the game.
Wenyen Gabriel
According to the same-four plus-minus data, Wenyen Gabriel's positive impact is exclusively on defense; UK is slightly worse on offense with him in the game. What is he doing that helps so much on defense, and what's missing on offense?
Let's start with offense...UK clearly shoots worse when Gabriel is in the game than when other players take his place. Given that Derek Willis takes his place for about 85% of the time, and Willis shoots much better than Gabriel (59% eFG% vs 53% for Gabriel), that's not a surprise. Gabriel is a much better offensive rebounder than Willis, but UK's team Off Reb% is only slightly better with Gabriel in the game. The individual difference in offensive rebounding isn;t having much impact on the overall team rebounding. So, Gabriel doesn't offer as much shooting as Willis, the team rebounds about as well offensively with him out, and that explains the slight negative offensive impact.
On defense, however, Gabriel has a pretty sizeable positive impact. In particular, teams shoot much worse on two pointers with Gabriel in the game, as he is an effective interior defender and shot blocker. When opponents miss shots, UK gets 74% of the rebounds with Gabriel in the game, compared to 68% with him out...that means fewer shots for the opponents, especially easy shots. In fact, opponents gets 2.5x as many putback shots with Gabriel out of the game. Given that opponents shoot over 60% on these, it's really useful to limit them. Additionally, it appears that Gabriel's presence discourages opponents from even attempting interior shots, as teams shoot 3-pointers more often with him in the game (34.3 3PA/FGA vs. 31.0). A side effect of shooting more threes is that teams are less likely to be drawing fouls on UK, and UK's foul rate does in fact decline with Gabriel in the game (23% vs 26%). This means opponents shoot fewer free throws. Lastly, teams turn over the ball more with Gabriel in the game than when he sits (22% vs 19%). Oddly enough, Gabriel himself isn't getting a lot of steals; he actually gets steals less often than Derek Willis on a per-game and per-possession basis. However, Gabriel's ability to cover ground on defense is limiting opportunities for opponents to shoot or pass successfully, and leads to missed shots or turnovers.
Gabriel's impact on turnovers is a good example of why it's important to look at team stats and not just individual stats. Per KenPom.com, Gabriel gets steals on 1.3% of opponents possessions, and Derek Willis gets steals on 1.6% of possessions. Just looking at this would give you the impression that Gabriel isn't as aggressive and athletic on defense as Willis, but it's only part of the picture. Clearly, Gabriel is doing a lot of things to force turnovers that aren't credited to him.
In summary, Wenyen Gabriel isn't as good of a shooter as the other options UK has at PF, and that hurts UK's offense a little. However, he has a much bigger positive impact on defense by limiting easy shots, securing defensive rebounds, and forcing turnovers with his activity and length. The net impact is that Kentucky is more likely to outscore their opponents with him in the game.
Thanks for reading my newsletter. If you have any questions, want to argue a point, or have some feedback, feel free to reach out via email at sean@hoopsinsight.com, or on Twitter @hoopsinsights. If you liked this, let me know as well, and tell your friends to subscribe at www.hoopsinsight.com.
You can forward this to others, but please ask them to subscribe as well so I can keep track of who's enjoying my insight and analysis.