Studying (really) early season trends of the Raptors

We’re six games into the NBA regular season. The Raptors are 6-0, so things seem to be going well. It is way too early to look too closely at advanced stats and draw any concrete conclusions about this Toronto team.

So, let’s do that! Without the concrete part, that is.

A Note on Small Samples

Everything here will be presented with a minutes total to make sure we’re all clear on the sample size. This really is too soon to be particularly confident in any numbers holding up throughout the year (I like the 20 game point as a good benchmark), but there are still Raptors trends worth keeping an eye on, and we’ll try to touch on a few of those.

So, let’s get to it.

Remember: KLOE

Kyle Lowry has had a fantastic start to the season. There are the obvious counting stats — he’s putting up 19.8 points, 10.3 assists, and 3.7 rebounds in less than 35 MPG. He’s leading the league in assists per game by a wide margin (the second place player is Harden with 9.0 APG).

Then there’s the efficiency Lowry is putting up. He’s shooting 57 percent from the floor — his career high is 46 percent. He’s shooting 53 percent from three — his career high is 41 percent. The combination of those is leading to Lowry putting up a 72.7 TS%. Now, his shooting will certainly regress at least a bit. But his season has started with him absolutely on fire.

(TS% - True Shooting Percentage: an adjusted field goal percentage that accounts for the value of three point shots and free throws.)

Obviously Lowry’s assist rate is at an all time high as well (41% vs. career high of 35%, and recent trend closer to 30%). Meanwhile, Lowry’s turnover rate has not climbed at all, despite running a somewhat new team, with a new coach, and creating so much more offense.

(Assist rate - what percentage of the team’s made field goals are assisted by the player when they are on the floor.)

This all spirals up to some monstrous early season win production numbers from Lowry. Lowry is currently generating 0.319 win shares per 48 minutes. For reference, James Harden and Steph Curry were at the front of the pack last year with rates of 0.289 and 0.267. That was over a full season, obviously, but still. Lowry has generated 1.4 WS total so far this year — making him responsible for a quarter of the Raptors’ 6-0 record all by himself.

(Win Shares - a catch all box score statistic intended to measure how much a player’s production contributes to winning.)

In short, Kyle Lowry Over Everything — for now, anyway. Long term, might want to think about this next guy.

Kawhi Leonard is a Raptor

He’ll be rusty. He missed lots of time. He’ll miss shots. His defense will take time to mesh with the rest of the team.

Heck, that’s probably all true, which makes what I’m about to write so scary.

Leonard has averaged 26.6 points per game (career-high) in 34.8 MPG (career-high), with shooting splits of 50 percent from the field, 46 percent from three (career-high) and 92 percent from the free throw line (career-high). Throw in 7.0 rebounds per game (career-high), 3.0 APG (second highest of his career) and 1.6 blocks per game (career-high), and you’ve got a good start to the season, I think. As an added bonus, Kawhi is still grabbing one steal for every foul he commits (eight of each so far), a truly ridiculous statistic that holds for his entire career (723 career steals to 706 career fouls committed).

I look forward to the rust coming off. Wow.

The Big Question

The biggest debate early in the season centres around coach Nick Nurse’s rotating centre, um, rotation. Sorry.

Serge Ibaka has been getting most of the minutes at that spot early on, and playing significantly more with the starting group. Some of that is Nurse’s early season desire to match up defensively (which is worth exploring, though I’d like to see the opposite explored as well), but it goes beyond that as well to get to such an extreme difference in minutes.

Ibaka has played about 50 percent more minutes than Jonas Valanciunas thus far (28 MPG vs. 19 MPG). I’m going to preface the following with a “this is fine” — it’s very early, there have been some smaller match-ups, Valanciunas was in foul trouble in the most recent game, and experimentation early in the season does not mean it will continue throughout the season. But on the off chance that this is a trend and not noise, let’s take a look at whether it’s a good trend or one that may become a problem.

First off, let’s take a look at the most played lineups (>10 total minutes played) the Raptors have used so far.

Lineup - Minutes Played - ORTG - DRTG - RTG
Lowry-Green-Leonard-Siakam-Ibaka: 55 MP, 116.1, 94.9, +21.2
Lowry-Green-Leonard-Siakam-Valanciunas: 35 MP, 129.7, 95.9, +33.8
VanVleet-Powell-Miles-Anunoby-Valanciunas: 19 MP, 77.8, 106.5, -28.7
Lowry-Green-Leonard-Anunoby-Ibaka: 14 MP, 110.8, 105.9, +4.9
Lowry-Green-Powell-Anunoby-Ibaka: 12 MP, 151.9, 108.3, +43.5
Lowry-Green-Leonard-Anunoby-Valanciunas: 11 MP, 134.8, 109.5, +25.3
Lowry-VanVleet-Green-Leonard-Ibaka: 11 MP, 138.5, 107.7, +30.8

(ORTG - offensive rating: team points scored per 100 possessions while a player/lineup is on the floor.
DRTG - defensive rating: team points allowed per 100 possessions while a player/lineup is on the floor.
RTG - net rating: The difference between ORTG and DRTG, essentially point differential per 100 possessions.)

Back to that small sample reminder: A good guideline for when lineup data tends to be more signal than noise is about 50 total minutes played. Giving us a grand total of: one lineup! So obviously drawing too many conclusions from this sample would be problematic. Let’s find some bigger samples — by shrinking down to three-man lineups, we can get an idea how the two centres have performed with, (a) both Lowry and Leonard, as a proxy for “starters” minutes, and (b) both Lowry and Leonard sitting, as a proxy for “bench” minutes.

Lineup - Minutes Played - ORTG - DRTG - RTG
Lowry-Leonard-Ibaka: 101 MP, 116.6, 105.9, +10.7
Lowry-Leonard-Valanciunas: 55 MP, 133.3, 94.8, +38.5
Ibaka without Lowry and Leonard: 27 MP, 80.4, 76.3, +4.1
Valanciunas without Lowry and Leonard: 31 MP, 82.4, 100.0, -17.6

Now we at least have a couple solid samples, and the low minute samples are still significantly better than we had with the 5-man groups. Though it should be noted there are obviously more unknowns with these lineups — such as who is playing in the other two spots. But with Lowry and Leonard mostly leading starting-type units and the bench being a revolving door, this is very likely the way to go.

The bench group struggling to score no matter who is on the court with them makes sense, considering how much time Delon Wright has missed, in addition to Pascal Siakam being removed from the group (leaving it short on ball handling), as well as having FVV and OG miss time lately. While JV usually provides a big offensive boost to most lineups, his greatest value comes as a screen setter and roller, and without real threats on the perimeter his value is greatly reduced. While the defense has hardly been bad in the JV look, with Serge there it really ramps up, based on that small sample. Ibaka does replicate some of the quickness and rim protection that Jakob Poeltl used to provide to a defensively dominant bench unit last year.

But what we should be interested in is maximizing Lowry and Leonard. The bench should sort itself out when it’s back at full strength, and the team has been winning even with the bench being mostly terrible. If that group keeps struggling, the Raptors can start to rotate Lowry and Leonard into the various bench looks to keep them afloat, which we’ve seen a little of already.

And yeah, those numbers do not compare. The two stars can score at will when given a solid screen setter in Valanciunas to work off of, and the defense has surprisingly been better as well. Now, the Ibaka look has been fine. A +10 RTG is a solid number. And I was one of the few who wanted to see last year’s starting lineup replicated, with Ibaka at PF beside Valanciunas, even if not used as heavily. Many shouted down that suggestion, but it is worth noting that that group last season had a +13 RTG over a huge sample and various opponents. If you thought that group wasn’t good enough, you may want to question the current primary starting group too.

As for the two stars and their respective career years, well:

Leonard with Ibaka: 58% TS%
Leonard with JV: 63% TS%
Lowry with Ibaka: 62% TS%
Lowry with JV: 90% TS% (how?!)

It’s too early to be writing off any looks, or to be worried about one look getting more time than another. Nurse should take as much time as he can to feel out different options, and having the ability to match up is a great tool to have in the tool box. But at some point, I hope we get to see the Raptors not try to match up, and force the mismatch by playing their best combinations of players and seeing if the other team has an answer.

Maybe we come back to this in a couple weeks and see what other fairly-small-sample-size theatre we can conjure up to be concerned about before we need to be, as well as checking in on the numbers above to see how much the story changes in that time.

All stats per NBA.com and basketball-reference.com.