December 6, 2009

Monta Ellis, Superstar... er?

You may have heard the kerfluffe in the hoops world about Kevin Durant's plus-minus numbers this summer. For all of his prodigious scoring and highlight-reelin', Durant didn't seem to be making a bad team better: the Thunder got outscored by over eight points per 48 when he played, and essentially played even with their opponents when he didn't. If Durant was so good, why were his teammates having better success without him? Was his defense that bad? Was he stifling his teammates? Did his absence make the coaching staff run simpler plays for the more limited Thunderers, actually improving their efficiency? Whatever the reason, Durant has made the whole conversation more or less moot with his results this season... he has a +9.2 plus-minus to go with his huge individual production, and has clearly been the main engine driving the Thunder to their suprising 10-9 record. His meteoric rise has continued unchecked, leaving many thinking it was a mistake to worry about him in the first place. Maybe a plus-minus of -8.3 isn't all that big a deal.

How about a plus-minus of -20.2?

That's what Monta's sporting right now. When he's on the court for us -- which is 82% of the time, tied with Gerald Wallace for first in the league -- we get outscored by 7.8 points per 48 minutes. Only truly awful teams get outscored by that much per game; some seasons go by without any team sucking to that degree. Whenever Monta, our star, our rock, our beacon of hope, has been out there this year, we've played like a 20-62 team.

"That's depressing," you think, "but we do suck. It'd be even worse if Monta wasn't busting his ass out there." Here's the thing: when Monta's not been on the floor, we've outscored our opponents by 12.4 points per 48 minutes. The record for point differential in a season is held by the '71-'72 Lakers, the Jerry West/Wilt Chamberlain 33-wins-in-a-row world-champion Lakers. Their point differential was... 12.3. When Monta's been on the bench, we have outscored our opponents at a pace that would make us the most dominant team of all time.

When you combine the two numbers -- -7.8 with Monta, +12.4 without him -- you get a total plus-minus of 20.2. Monta's plus-minus is more than twice as bad as Durant's much-discussed one last year. Monta's plus-minus is, in fact the worst of any NBA starter by far. The five worst plus-minuses for guys who've played at least 50% of their team's minutes so far:
1. Monta Ellis, -20.2
2. Gilbert Arenas, -15.3
3. Jeff Green, -13.6
4. Shane Battier, -12.2
5. Derrick Rose, -11.4

Some players on this list, no question, but there are pretty big question marks about all of them. Arenas is still not 100% and getting torched on D. Jeff Green can't rebound or defend enough to hang at his position. Various metrics suggest that Battier's losing a step and not the impact defender he once was. Rose is calling himself out for poor play. Most of these guys are getting tons of crap from their respective fanbases. And yet we see Monta as our big bright spot, as a guy taking it to the next level, and he has a worse plus-minus than any of them.

What the hell is going on here?

Let's start with this: that +12.4 sans-Monta number is absurd. Monta Ellis is not the only thing holding us back from being a 68-14 team. A couple big garbage-time runs in games we lost by 30 are distorting this data; in the 154 non-Monta minutes we've played so far (not many, because he's played so much), the team has scored 127.1 points per 100 possessions, an unsustainably high number. The +12.4 is nothing to worry about...

...but the -7.8 is, as it bespeaks some big offensive problems. Last year, for all of our injuries, dysfunction and lack of playmaking, we scored 109 points per 100 possessions. (Monta's contribution to this was negligible, as our offense with him was about the same as our offense without him.) This year, when Monta's on the court, we're scoring 103 points per 100 possessions. Even if you call shenanigans on the 127.1 PPos of our non-Monta lineups, there's no way you can reconcile that number all the way down to 103; with Curry aboard, there's no reason to think it'd be any be lower than 109. Our offense has genuinely functioned better when Monta's been off the floor. And while Monta's looked pretty great defensively at times so far, overall there's no evidence he's moving the dial at that end... we've given up only 0.6 points more per 100 possessions when he's sat. As with so many of our numbers thus far, that may have to do with our weirdly small lineups. Still, if you're looking for proof that Monta makes a significant difference on defense, there really isn't any. We have been a really, really bad team with him on the floor.

The calls to trade Monta were frequent and loud as recently as a couple weeks ago. Those calls have subsided in the wake of Jack's departure and a couple of spectacular performances on Monta's part, and I'm glad... now is no time to trade him. But his wretched plus-minus numbers have not changed. Since Jack left, here's how we've fared:

@CLE: -8 in 46 minutes with Monta, +2 in two minutes without
@BOS: -16 in 40 minutes with Monta, +2 in eight minutes without
POR: +14 in 48 minutes with Monta
@DAL: +8 in 48 with Monta
@SA: -21 in 44 minutes with Monta, +7 in four minutes without
LAL: -27 with 35 minutes with Monta, -6 in thirteen minutes without
IND: +12 in 42 minutes with Monta, +7 in six minutes without
DEN: -24 in 30 minutes with Monta, -4 in eighteen minutes without
HOU: -2 in 47 minutes with Monta, 0 in one minute without
ORL: -8 in 48 minutes with Monta

In this post-Jack, new-look world, since the Warriors became The Monta Ellis Show -- including his masterpiece in Portland, his two 40-point games, everything -- we are -72 with him and +8 without him. In every single game where he's sat, we've been better on a per-minute basis with him on the bench. We've been outscored by 8.1 points per 48 when he's been on the floor, actually doing worse than before... now we're playing like a 19-63 team.

The sample sizes are tiny. The guy's tired. He only sits when things aren't competitive, so of course the other guys will do well in garbage time. His numbers will be fine when the bigs get back. Maybe. Maybe. Maybe.

But there is cause for concern here. The Monta Ellis Show, as entertaining as it's been, has not been working. And if we're going to pin our hopes and dreams on this guy, we would do well to figure out exactly why we're struggling so much when he plays.


Kai said...

i think these numbers are misleading for a couple of reasons. Obviously, we have a better +/- when monta is off the floor; he only gets taken out when its a blowout, scrubs are in, and we score more. These numbers are off even more due to the fact we have such a small bench. In these blowouts, were playing at least 3 of our starters when the opponent isnt playing any. It would be interesting to see as the year goes on because as we get healthier and better, i think we will see a drastic rise in montas =/-. its hard to do everything by yourself(as montas finding out right now with such little health)

Owen said...

Some good points, especially that we tend to have more starters in in non-competitive situations, thanks to our short bench. And another data point in Monta's favor that I neglected to mention: he's been on the court for almost every single one of Mikki's wretched minutes. That alone probably hurts his plus-minus by four or five points.

Still and all, Monta is not the only good player getting big minutes on a bad team. Gerald Wallace plays as many minutes as Monta does, also on a bad team... the Cats play their opponents about even when Wallace is in, and his net plus-minus is +6.9. Rudy Gay's averaging about 40 minutes per for the Grizzlies, and while they're not good with him, they're a whole lot worse without him. His net plus-minus is +6.2. Context can shrink the gap between Monta and guys like them, but it can't eliminate it. Something is really not working about our current strategy.

Something we can probably agree on: there's not much point in wearing Monta like this, in a season that's really already lost. It's time to dial him back to 34 minutes, so he can expend full energy when he's in and teams will have to re-adjust to him a couple times a night. If we do that, I'm sure his numbers will improve. He's a good player -- he's just not the iron man we want him to be.