The Wood Shop: Benching the Starters

This piece has become a quarterly tradition of sorts. I look at the lineup data for the Cavs and come up with all sorts of crazy conclusions. One year, I said Matthew Dellavedova was better for the Cavs than Kyrie Irving. How the times have changed. Another time, I posited that DeAndre Liggins should be starting over J.R. Smith. Well, this year, I’m upping the ante. The Cavs shouldn’t close games with their starters. They need to run out Dwyane Wade, LeBron James, Kyle Korver, Jeff Green, and Channing Frye. It’s a group with an average age of about 33.5. They’re old, they’re pretty slow, they’re professionals, and they’re awesome. This group gives the Cavs their best chance to win on a nightly basis. Let’s see why.

Methodology

If you’re interested in the methodology behind this analysis read this brief explanation in the following paragraph, otherwise, skip ahead. The process of comparing five-man groups is simple. To start, I found the lineup information, specifically plus-minus data, and then converted it to a per 48 minute rate, or actual plus-minus per 48 (pm/48). And, before someone even says it, I obviously know lineups aren’t going to put up per 48 numbers because their production would decrease as they grew tired throughout a game. The rate just makes it possible to compare lineups that have all played different amounts of minutes throughout the season.

The next step to figure out which lineups are really working deals with figuring out what a lineup should be doing plus-minus wise if the players are looked at as individuals who just come together without chemistry or other on the court stuff factoring in. This number is called expected pm/48. To figure this number out, I looked up all the players involved in the lineups and calculated their pm/48. Then I added the pm/48 of each player in a lineup and divided it by five to figure out the expected pm/48 of a unit.

Once I got the actual and expected pm/48 numbers, I compared them. If a unit’s actual numbers were better than their expected numbers, then the unit was making basketball magic. If a unit’s actual numbers were lower than the expected ones, the unit was underperforming.

The Breakdown

1. Thankfully for the Cavs, their current starting lineup is actually outperforming expectations by 6.76 points over a 48 minute period. Tyronn Lue is doing something right starting Jose Calderon along with James, Jae Crowder, J.R. Smith, and Kevin Love.
2. And, it’s almost a miracle this lineup happened. If Iman Shumpert hadn’t gone down, Jose would have never gotten his chance. It’s possible Lue might have realized that that original starting group wasn’t working with Shump, but it surely wouldn’t have been from looking at his individual plus-minus/48 (pm/48) stats and comparing it with Jose’s. Shump’s pm/48 is 5.5 while Jose’s is just 0.5.
3. Their swap out results in a staggering actual 19.1 difference in points over 48 minutes. And, the expected performance difference is 20.34/48 better.
4. That leads me to this point. The best Cavalier lineups are half-versions of their bench. The best group is Love, Wade, Korver, Smith, and James. This group has a pm/48 of 54.4 and is outperforming their expectations by 49.34 points over a 48 minute period.
5. The Cavs’ second best group just swaps out Smith and Love for Frye and Green. That group is outperforming expectations by 18.44/48.
6. Lastly, the two bench groups that don’t even have James in them are treading water. They’re putting up positive real pm/48 numbers. Although it must be noted that swapping out Cedi Osman for J.R. Smith when playing with Frye, Korver, Green, and Wade results in both an over performance and better real floor numbers.
7. There are a few simple conclusions to be drawn from the way the lineup data plays out.

Conclusions

1. LeBron thrives when he is surrounded by guys who don’t need to dribble and know their role. Swapping Jose for Shumpert really shows this. Jose dribbles but he dribbles with purpose. He isn’t doing vine length crossover sequences and then launching a mid-range jumper. I’m hoping Shump sees this piece and takes note because he really should be better than a 36 guy who has in-shape-dad-bod.
2. Furthermore, the fact none of the top lineups have Crowder in them hints at something else. For the Cavs to be effective everyone on the floor needs to be purposeful around James.  For a unit to succeed, the players must shoot 3s (Smith 36%, Wade 36.2%, Love 41.5%, Korver 44.3%, Frye 31.7% but coasting because of reputation), cut effectively (Green, Wade, Osman), move off the ball (Wade, Korver, Frye, Smith) or run the floor (Green, Osman). Notice that I’m not mentioning defense; the Cavs aren’t going to win anything with the bottom of the league defense they usually display. More on that later though.
3. Being purposeful is something that’s a must for all Cavs units. If Shumpert, Crowder, or Tristan Thompson were on a unit for this exercise, the unit was one of the worst that was looked at. They don’t shoot 3s, don’t finish that great, and can’t create.

Don’t Finish Games with the Starters, or Even the Best Unit from this Excercise

Now, that proclamation in bold is sort of nutty on the surface. You’re probably thinking, “Come on, David, those late nights blogging have fried your fragile brain. Smith, Wade, Love, James, and Korver are outperforming expectations by 49.34 points per 48. That’s crazy. And, the starters aren’t bad either. They’re outperforming expectations by 6.76/48. That’s really good considering they have played so many minutes this season. That group isn’t a fluke.”

Well, you’re wrong. Here’s the thing about basketball. You actually have to play defense to win. I know I said different above, but you really do. The Cavs won their only Finals ever beating the Warriors 93-89 in game seven and holding the Warriors to 101 points in their game six win. The two units that generic hater one mentioned above are sporting defensive ratings of 134.2 and 106.7 respectively. Even the 106.7 rating places the Cavs right around 24th in the league.

Guess what group would lead the league if they were representative of the Cavs defense as a whole: Wade, Korver, James, Frye, and Green. That group sports a defensive rating of 89.7. That’s brutal. I haven’t had a chance to dive too deep on why that is yet, but all of those guys are veterans, and, with the exception of Green and sometimes Wade, they don’t gamble (James gambles but he turns it to ten on the text book defense level when he needs to). They’re a group of guys that has their hands up and will try to move with their man.

That unit is the real deal on offense too. They post a rating on 124.1. Again, that’d be number one in the league if you were wondering.

That unit also functions well as an end of game group for another reason. It doesn’t rely as much on the pick and roll. I don’t have access to data that can support this but just watching games I’ve noticed this. Whenever Korver and Frye are on the floor together, they are constantly moving and receiving off-ball screens to get open (See the videos above and below). Unlike when James shares the floor with Love, the 3s that are happening from Frye and Korver aren’t of the pick and pop variety as often. They’re coming because LeBron is penetrating or Wade is penetrating and finding them open. Or the threat of them driving has guys that are supposed to be covering the perimeter ball watching. Then Green is moving off the ball, setting screens, and diving to the hoop. The brilliant thing about this lineup is that it forces teams to defend James one-on-one.

James is a great pick and roll player. He’s one of the best in the league at reading those situations, but a pick and roll is usually defended by two players. Those two players often end up on LeBron. This power unit avoids this problem. And, if opponents try to cheat and double James, they’re leaving open two deadly 3-point shooters, a crafty cutter in Wade, and a workable man in Green. It’s a death sentence to double him because he will find the right man and chances are they’ll score.

And, if you can’t buy that this power unit should be on the floor, at least concede this point: Kyle Korver needs to be on the floor at the end of games. Of all the players James shares the floor with, Korver improves his shooting percentage the most of any other guy. He increases it by 5.4%. And, when James shares the floor with him he averages 10.1 free throws per 48.

The reason is simple. If Korver is in the corner, his man isn’t helping off onto a driving James. No way, no how. Shooting 44.3% from deep will do that. In the next week or so keep checking back, I’m working on a piece about how Jeff Green contributes with his off-ball movement.

*Stats are accurate as of 12/27/2017 and from NBA.com