Projecting the post-season rotation for the Raptors

The playoffs are finally here! I’ve been quite absent for a long time now (haven’t we all), but with a few warmup games under our collective belts, it’s about time we take a look at the upcoming post-season for the Toronto Raptors.

Now, I don’t tend to focus much on the opposition, instead looking at what has worked for the Raptors this season as far as their lineups and rotation go. For now we’ll focus less on the recent discovery that the guys occupying spots 11-16 on the roster are all superstars and look more at what has worked with the rotation players throughout the entire season. Big sample sizes and all (where we can get them).

Those sample sizes will be tricky. So many key Raptors missed so much time that there really isn’t a massive sample for any lineup on the team. It’s part of the reason I wanted to wait so we’d have a few more meaningful games with a largely healthy roster. Now we’ve edged up those minutes totals for each likely rotation group before we looked at this.

So, where to start? Well…

Go With the Starters

This section should be easy. The Raptors have a fairly solid starting group, even with so many of them missing games. In every game they were available in the Bubble, Toronto started the same group, settling on using their five best players together for as much of the game as possible.

Really, the Raptors have a top seven from which to realistically choose a group of starters and primary rotation players, so we’ll take a quick look at the options they have.

Lineup | Minutes Played | ORTG | DRTG | RTG

KL-FVV-OGA-PS-MG: 361 MP, 110.9, 98.1, +12.9
KL-FVV-OGA-PS-SI: 210 MP, 102.0, 105.8, -3.8
KL-NP-OGA-PS-SI: 120 MP, 114.0, 114.4, -0.4
KL-NP-OGA-PS-MG: 111 MP, 115.9, 102.5, +13.4
KL-FVV-NP-PS-MG: 56 MP, 81.7, 100.0, -18.3

ORTG: Offensive rating; points scored per 100 possessions
DRTG: Defensive rating; points allowed per 100 possessions
RTG: Net rating; point differential per 100 possessions

I’ve limited the lineups listed to those with at least 50 minutes played. The least played lineup there was used across 12 different games, so they should have a decent spread of opposition. And these lineups have enough starters in them that they would likely be facing starter-heavy opposition.

A few things immediately jump out. There is no way the Raptors should even consider Serge Ibaka starting over Marc Gasol. The difference is astonishing in the four most used lineups above, all with over 100 minutes of sample. Something to also keep in mind when debating who the Raptors should be attempting to keep this “summer,” but that’s a discussion for another time.

Norman Powell in for Fred VanVleet seems viable, but in the one lineup they tried for a significant stretch with him in for OG Anunoby, it didn’t go well. This may be a feasible swap at some point, but for now our baseline looks pretty obvious. We can write down Lowry, Pascal Siakam, OG, and Gasol in pen, and with how large a sample the main starters have, we’ll leave Norm on the bench and start Fred, at least for now and unless we find a better way to use them as we explore bench and transitional lineups.

Considering the Bench

During the regular season, bench units can be asked to carry themselves, buy the starters rest, and expand their games where they can. That’s fun and helpful in the long run, but come the playoffs that tends to stop. Other teams will be riding their best players, and the Raptors will likely do so in return. Certainly coach Nick Nurse has alluded to having seven players (who we’ve already mentioned) who will be receiving “starter minutes” and we can expect those minutes to be higher in the playoffs than they were in-season.

My favourite approach to bench units is to consider them purely in terms of who they buy rest for. In an ideal world the Raptors would play their starting unit 48 minutes and they’d be fresh as a daisy in every one of them. In reality, each player needs time off the court, and these bench units’ primary function is to find a way to survive (or thrive) in the minutes those players are not playing together.

The Raptors have a few players who they will rely on for heavy minutes, so we’ll look at each of them to find out if we can patch together units without them to buy them rest.

Subbing for Kyle Lowry

We’ll start with the toughest one. In more ways than one. Lowry saves every unit he is in, and has done so for basically his entire Raptors career. Which means the lineups where he is sitting are usually the toughest to find good results for. Let’s quickly take a look at the most used lineups without Lowry this season.

For these bench units, we’re dealing with smaller samples, so I’ll show any lineup with more than 30 minutes to their names. Keep in mind that means these results are little less trustworthy than units with larger samples.

Lineup | Minutes Played | ORTG | DRTG | RTG

FVV-NP-OGA-PS-MG: 201 MP, 114.1, 97.3, +16.8
TD-FVV-OGA-PS-SI: 39 MP, 120.0, 102.4, +17.6
FVV-NP-RHJ-PS-MG: 36 MP, 93.5, 96.1, -2.6
FVV-NP-OGA-PS-SI: 35 MP, 104.9, 93.8, +11.2
TD-FVV-RHJ-PS-CB: 30 MP, 111.9, 93.8, +18.0

So, that’s good news. Unlike in other years, seems the Raptors have found plenty of looks to stay afloat without Lowry. The main one there is the starting lineup they ran with while Lowry was hurt — hence the large sample — and it looks pretty good. Seems like a strong argument for Lowry subbing off early in games and carrying a bench unit for a while to buy some of the other starters rest.

But the other lineups being big positives is a nice surprise too. That last one with the smallest sample is part of a trend that lasted all season long, the three non-top-7 bench guys in Terence Davis, Rondae Hollis-Jefferson and Chris Boucher made a heck of a trio all year and caused enough chaos against opposing benches that the Raptors tended to win their minutes. In fact, on the year, the three played 305 minutes together to the tune of a +8.6 net rating. Not too shabby for guys not guaranteed to play in the playoffs.

Now, that’s a good card to have in the back pocket, but as we noted, this playoff rotation will likely be constrained to the top-7 plus maybe one more player in any given lineup. It’s a bad sign for Hollis-Jefferson that the other lineup he appeared in with more starters really struggled to score. This is a problem lineups will tend to have with him in them, as he’s not great with the ball in his hands and you can’t take him off the ball because defences will ignore him to pack the paint. So for the moment, RHJ doesn’t seem a great candidate for lineups that buy Lowry minutes.

One big positive here is that Ibaka seems much better able to replicate Gasol’s success in bench looks like these compared to the drop off we see with the starters. The starting lineup with Serge and Norm in for Lowry and Gasol does very well in the small sample they have, and so does the one with Davis and Serge in their places. Toronto has options.

One last check. Because samples are so small, it is good to get an idea which players and player pairs fare best away from Lowry, in case these more zoomed in looks at individual lineups are missing a bigger picture trend. I’ve looked at each player without Lowry. For player pairings, I used the two likeliest heavy minutes candidates (Siakam and FVV) and took their most played pairings (>300 MP, or part of top seven) without Lowry.

Player | Minutes without Lowry | RTG without Lowry

FVV: 848 MP, +5.6
OGA: 595 MP, +6.5
PS: 767 MP, +8.3
MG: 499 MP, +11.6
NP: 782 MP, +4.4
SI: 394 MP, +1.7
TD: 742 MP, +4.0
RHJ: 635 MP, +7.1
CB: 500 MP, +4.9
PM: 362 MP, -0.6
MT: 339 MP, +7.0

Pair | Minutes without Lowry | RTG without Lowry

FVV/PS: 570 MP, +9.2
FVV/NP: 512 MP. +5.2
FVV/OGA: 466 MP, +8.0
FVV/TD: 370 MP, -0.5
FVV/MG: 367 MP, +11.4
FVV/RHJ: 333 MP, +4.8
FVV/SI: 300 MP, +5.3
PS/NP: 519 MP, +4.8
PS/OGA: 406 MP, +10.4
PS/MG: 398 MP, +12.4
PS/TD: 315 MP, +9.7
PS/SI: 185 MP, -4.6

Right away, there are a few concerns the smaller sample lineups didn’t pick up on. Ibaka really is not doing so great away from Lowry in general, and specifically beside Pascal, who is needed to bolster non-Lowry lineups. If the Raptors are going to use Serge and Pascal together, they should probably be careful to use them in starter-esque lineups like we looked at in the first part of this section. The bottom line here: Ibaka is one of the weaker bench players away from Lowry in terms of results. With any luck, when we look at Lowry lineups while Pascal is resting, we will find some useful minutes for Serge, which should combine with the transition lineups full of other starters to round out his minutes.

The other is the general ineffectiveness of VanVleet and Terence Davis. That’s a lot of minutes of limited effect. It sure looks like Davis should play beside Pascal while Fred is resting, if they can manage it. This is probably a result of too much offense going through two guards who are better playing off a more ball dominant player — and not enough offense being run through Siakam in those looks.

Everyone else notable has decent to strong results individually away from Lowry, which is a refreshing change from the past. Add another data point to the “never play Patrick McCaw again” tally. Let’s move on for now.

The Presence of Pascal

Siakam is going to be carrying a heavy load on offense and will still be asked for plenty on defense even if OG is taking the tougher assignments this year. So they’ll need to buy him time for sure. Let’s do the same exercise as for Lowry, starting with some lineups.

Lineup | Minutes Played | ORTG | DRTG | RTG

KL-FVV-PM-OGA-SI: 135 MP, 99.0, 107.7, -8.7
KL-TD-NP-RHJ-SI: 45 MP, 83.5, 119.2, -35.7
KL-TD-PM-OGA-SI: 37 MP, 125.3, 93.1, +32.3
KL-TD-MM-RHJ-CB: 31 MP, 146.9, 93.9, +52.9
KL-PM-OGA-RHJ-SI: 30 MP, 100.0, 113.8, -13.8

So. Problems. Small samples, wildly different results with small changes to lineups, and none of these lineups played in more than 10 games, some as few as three games. Which will mess up the average opposition something fierce. We’ll have to patch something together in the second part of this exercise.

Player | Minutes without Siakam | RTG without Siakam

KL: 756 MP, +0.4
FVV: 581 MP, -1.8
OGA: 627 MP, -0.6
MG: 155 MP, +7.7
NP: 465 MP, -0.1
SI: 728 MP, +0.7
TD: 772 MP, +4.9
RHJ: 722 MP, -0.1
CB: 552 MP, +1.2
PM: 649 MP, +0.2
MT: 296 MP, -1.6

Pair | Minutes without Siakam | RTG without Siakam

KL/SI: 519 MP, -2.5
KL/OGA: 439 MP, +0.2
KL/PM: 387 MP, -3.2
KL/RHJ: 346 MP, -4.8
KL/TD: 345 MP, +10.8
KL/FVV: 304 MP, -2.4
KL/NP: 202 MP, -5.3
KL/MG: 55 MP, +8.1
FVV/SI: 383 MP, -1.6
FVV/OGA: 341 MP, -1.8
FVV/PM: 309 MP, -1.7
FVV/RHJ: 272 MP, +0.5
FVV/TD: 227 MP, +0.2
FVV/NP: 161 MP, -3.1
FVV/MG: 60 MP, -10.1

Oof. That’s not pretty. Not a lot of shining examples of how to make it work while Pascal is resting. We’ve got a tiny sample of Gasol success, a very large sample of solid success from Davis, and basically nothing else. Looking at their pairings, Gasol has basically split his time with Lowry and Fred, doing well with Lowry and poorly with Fred. Davis does well with Lowry and barely gets by with Fred. Oddly, as they seem so in sync, Lowry and Ibaka do not do well away from Pascal, and Fred and Serge don’t fare any better. The implication here is we have a large sample of Ibaka not working away from Pascal, so maybe Gasol needs to play while Pascal rests? We don’t have a strong sample of it working, but we also don’t have a big sample of it not working, which is a step in the right direction. Even the tertiary option of Chris Boucher isn’t doing anything special without Pascal on the court.

OK, tough spot here, and Pascal appears to be this year’s Lowry — this is where the work will be, finding rest minutes for Siakam.

Let’s start with using Gasol at centre and Lowry at PG, since he’ll be resting in the Siakam-led bench looks we touched on earlier. And with Davis having success in the above section, he’ll play beside Lowry, giving Fred some rest. Norm kind of needs to play in these bench units, or what is the point of him being in the rotation at all? Already we’ve got combinations of players that literally haven’t played together. Look for 5-man lineups with two of those players away from Siakam and you’ll find only tiny samples. For example, Marc Gasol’s most used lineup without Siakam played 16 measly minutes.

But we know what doesn’t work. So we build this lineup by process of elimination. And hope the talent gets us there. Lowry, Davis, Norm and Gasol. plus OG at the 4 — because the best bet for getting an unproven bench group to succeed is to throw starters at it. The foursome of Lowry, Davis, Norm and OG have played only 14 minutes together away from Siakam, but the results are at least very good (+17 RTG). This one is basically going to be a case of hold your breath and hope having enough good players on the floor gets you positive results.

Bet on VanVleet

Subbing VanVleet should be the easiest, as the Raptors can use both Lowry and Siakam when Fred rests. With how much we had to use Gasol above, this is a lineup that will need to use Ibaka to buy Gasol his rest. So, let’s start off simple, we already have three of the players we will use. And based on the sections above, we know that TD is already playing, but no other player outside the top seven has been selected yet. We’ll leave them all off for now, since we know Nurse plans to stick mostly to that top seven.

That means we have limited options to fill out the lineup. Lowry, Siakam, and Ibaka means we are looking for a SG and a SF. Our rotation options are basically down to just Anunoby, Powell, and Davis. The default here is to get Anunoby some rest, but we’ll look at all three options to be sure.

Lineup | Minutes Played | ORTG | DRTG | RTG

KL-TD-NP-PS-SI: 24 MP, 127.3, 92.6, +34.7
KL-NP-OGA-PS-SI: 120 MP, 114.0, 114.4, -0.4

Sadly the other option had a miniscule sample not worth listing. That first lineup also has a tiny sample, so we should check the individual impacts on the three guys we know are playing, considering how great that result is. The decent sample on the OG lineup suggests it’s not worth eating his minutes in this lineup when there is a more desirable option from a rotation perspective that also might be better competitively.

Player | Minutes with Lowry, Siakam, Ibaka | RTG with Lowry, Siakam, Ibaka

TD: 56 MP, +36.8
NP: 220 MP, +9.7
OGA: 408 MP, +3.1

Sure looks like we chose the right two players — Davis has incredible results in a small sample, and Norm has very nice results in a large one. Have to feel pretty confident this lineup could work.

Settling the Rotation

All that’s left to do is consider the Raptors rotation that comes from these lineups.

Starters: KL-FVV-OGA-PS-MG
Lowry rest: FVV-NP-OGA-PS-SI
Fred/OG rest: KL-TD-NP-PS-SI
Fred/Siakam rest: KL-TD-NP-OGA-MG

If the Raptors need to keep their three mainstays in for 38 minutes, and OG as well (he was regularly in the high 30’s prior to the bubble), that means each of the rest lineups needs to function for 10 minutes, or 5 minutes per half. So, in each half, 15 minutes is actually a rest lineup, while nine minutes is for the starters. That gives Toronto 18 minutes of the starting lineup together for the game. The starting lineup averaged 16 minutes in games they played together this year so that’s about right.

There would need to be some quick transitional lineups between these, to allow OG and Pascal to avoid playing more than a full quarter to start. But that seems like the backbone of successful lineups that can provide enough rest for the heavy minutes players on the team.

Meanwhile, our key bench players project to between 20-30 minutes a night, Gasol projects to 28 minutes (which can certainly edge up if needed, he fits in basically any lineup Serge fits in or even beside Serge in some matchups), and the Raptors don’t run out a single lineup that has less than two starters on the floor. What’s more, the only lineup with two of the three top ball handlers resting has three starters on the floor. All in all, these lineups have everything covered.

********

What do you think? Have I dismissed too easily the idea of playing additional depth options like Hollis-Jefferson, Matt Thomas or Chris Boucher, who had varying degrees of success in the Bubble? Or do you think one of them should take Davis’ spot in the top eight, as Nurse has suggested himself lately?

Does the minutes distribution make sense or should someone be getting more minutes? And leaving aside what we think we should do, what do you expect Nurse to do?

All stats per NBA.com.