Dissecting the Derrick Rose/Carmelo Anthony two-man game

Sunday's unexpected victory against the San Antonio Spurs was probably an anomaly in the long term, but it did provide a microcosm of the preferred offensive game plan during close games. When Derrick Rose and Carmelo Anthony are healthy, Jeff Hornacek loves to call for the Rose-Melo two man game as a source of crunch time offense. They can opt to do so within the Triangle, or they can abandon all subtlety and clear out; we've seen both over the course of the season. On Sunday, they abandoned all subtlety, and it didn't matter; the Rose-Melo playcall, in this instance, directly led to victory.

One game is hardly representative, though, and a relatively stagnant late game playcall like that often rubs people the wrong way regardless of the end result. The complete lack of ball movement involved in this "set" can be grating, especially when the shots don't fall. And, of course, the Knicks have lost six or seven close games in the last month alone -- it must not be working THAT well, right?

Is there a basis for consistently riding one play for the entire season, a season in which they’ve won just over 40% of their games? Is this adherence to a particular playcall part of the problem? Is this a red flag on the resume of Jeff Hornacek?

Before we unpack the ideology behind this play, we'll start by looking at why it worked so effectively against San Antonio.

Here's the first appearance, and it's a simple one. Melo gets to his spot close to the elbow, as he usually does in these situations, and cheekily backs off just as Rose makes his move. This effectively (and illegally, but illegal screens are the catch rule of the NBA) removes Rose's defender from the play, and from there it's a routine finish through David Lee for the bucket. That was pretty easy; more importantly, it helps influence how San Antonio reacts to this set moving forward. This play ended up setting the table for later possessions.

A couple minutes later, the Knicks went right back to it, and this time San Antonio is ready.

After getting burned playing straight up the first time around, San Antonio decided to switch. When you have guys like Kawhi Leonard and Danny Green on the floor at the same time, guarding the two biggest threats, it's a pretty good strategy. Because of the switch, Melo gets a "mismatch"; unfortunately, it's not as big of a mismatch as it would be against other teams (which is probably what prompted the switch in the first place). Doesn't matter; Melo backs down into his office and hits the patented fadeaway over two dudes. This isn’t a great shot, but I’ve seen Melo hit it enough times to feel OK.

So at this point, the Spurs have tried to defend straight up, and they've tried switching. Neither has worked. So, in response, the Knicks went right back to it, needing one more bucket to dagger the immortal Spurs.

This time, Melo cleverly changes the plan, part of the ongoing (and consistently underappreciated) meta-game that goes on over the course of a 48 minute game. Instead of setting the screen, he slips it, and causes an apparent miscommunication between Green and Leonard (Leonard is prepared to switch, but Green wasn't ready for Melo to bail out so quickly, leaving Melo uncovered). And that's all she wrote; a catch and shoot midrange is a good shot for Melo, and he makes it count.

Clearly, this play can be difficult to stop (when executed properly, which is a big caveat). Overplay on the Rose drive, and you're leaving Melo wide open for a spot up shot attempt. Commit to stopping Melo, and Rose blows by for a free run at the rim. Switch, and you're giving both Rose and Melo a mismatch to attack in isolation (more on that later). If the play itself is executed correctly, it puts a ton of pressure on a defense; ultimately, that’s exactly what you want out of a play.

As an aside, you can see in each clip that there is a token amount of off ball movement and "screens" from the three guys who out of the play. That seems to imply that Hornacek wants those guys to move around rather than standing still. The execution ranges from "bad" to "not even trying", so I’m not entirely sure what to make of it. They should do more of that, though. Doing a better job distracting potential help defenders will only make the play more effective.

_____________________________________________________________________________________________

Decent sales pitch, right? Sunday was an anomaly, though; games have been won and lost on the back of the Rose-Melo two man game (pretty sure the defense has led to more losses than the late game offense, but whatever). While this play isn't a traditional isolation, the lack of complexity and subtlety means it ultimately functions similarly to an isolation, even if it's actually a simple pick and roll. And as we all know, isolations can be volatile.

But it's important to remember the counterfactual when discussing those late game playcalls. Simple plays are favored late in games because of what they avoid rather than what they actually accomplish. A simple two man game has a low risk of a turnover, which is a big deal; live ball turnovers in crunch time situations are death. Losing a possession and gifting the opponent a fast break is about as bad as it gets. Avoiding turnovers is important; unfortunately, actively avoiding turnovers also means cutting down on ball and player movement. Is it worth neutering your offense for the sole purpose of avoiding turnovers? Maybe, maybe not; that’s a discussion for another time. The decision becomes easier, however, when you have gifted scorers like Rose and Carmelo.

On top of that, simple late game playcalls generally allow your defense to get set behind the play, forcing an opponent to execute in the halfcourt (which is less efficient, and simultaneously burns more time off the game clock). Simple plays also allow direct control of the clock; if you're running a straight up pick and roll or a clear out isolation, it's much easier to milk the shot clock to the very end. When you're trying to close the door, every second counts, as we all know (related: fuck Reggie Miller).

With that in mind, let's look the options for defending this play to get an idea of how it usually plays out on paper. Then we'll take a look at the corresponding numbers to get a feel for how effective this play should be, in an ideal world. Naturally, it doesn’t work like that, as identical playcalls can end up looking wildly different in practice. Regardless, there’s a lot of reasons Hornacek likes to abuse this particular play in crunch time.

Please note that I did not use “clutch” numbers. Not clutch possessions feature this specific play, and on top of that, the sample size is too small to take seriously. Therefore, it's safe to assume all subsequent points per possession figures overestimate impact in clutch situations (due to factors like reffing, fatigue, and focused defenses), but they'll serve as a decent barometer nonetheless. It’s unquestionably overly simplistic, but I’m not going through hundreds of crunch time possessions, by hand, for a shitty team. That kind of content comes when the Knicks are actually worth diving into. Sorry, fellas.

OPTION 1 - DEFENDING STRAIGHT UP

When defenses play it straight up, Derrick Rose takes on the role of initiating the possession. Fortunately, the one thing he does at a high level is finishing in the pick and roll (0.911 points per possession, 75th percentile). More importantly, the threat of Rose puts pressure on the defense; if Rose doesn't take the shot himself, there's a potential spot up opportunity for Carmelo, as we saw before. In those situations, Melo scores 1.289 points per possession (which is completely insane. Seriously. That's 97th percentile, on high volume. Get the guy a true playmaker, please). Combine the two, and you’ve got a conundrum for opponents. If Melo gets any contact at all on his screen, his defender has to help off, leaving Melo open for a spot up look. If that defender does not help, Rose gets a head start towards the rim, where he’s been finishing at a high level.

Cherry on top? Both Rose and Melo are comfortable enough in these roles to take care of the ball. Rose only loses the ball on 11.3% of his possessions as a pick and roll ball handler (despite a couple of seriously glaring errors). And Melo only coughs up the rock only 2.5% of the time in spot up situations. High chance of scoring, plus low risk of a turnover, plus all the additional benefits of a conservative playcall? Not bad.

If the opponent does manage to defend it straight up, you ideally move right into a real pick and roll. The shot clock can be an issue there, though, and that’s not how we do things over here anyway. This turns into an isolation. And that’s OK, at least as a fallback plan, because...

OPTION 2 - SWITCHING

Switching usually gives both Rose and Melo a mismatch to attack in isolation. Rose is a dynamic player one-on-one, and it shows; he scores 0.92 points per possession in isolation, while turning the ball over on only 6.3% of his attempts. And we all know about Melo — he’s at 0.93 PPP, and turns the ball over on only 5.4% of his attempts. Those are pretty nice figures. This is probably the more palatable option of the two, although matchups also dictate how opponents want to defend this play.

Even then, it’s not always simple; switching can lead to miscommunications if not executed properly. As we saw above, simply slipping the screen can cause defenders to hesitate, giving an advantage to the offense. And that was against the Spurs, one of the most consistently strong defensive teams the league has ever seen. If Danny Green and Kawhi Leonard can screw up a simple switch under pressure, imagine how the Blazers would react.

Of course, if the switch is successful, either Rose or Melo attacks the mismatch in isolation.

Basically, the Rose-Melo two man game can be pretty difficult to stop. There's no easy answer for defending it outside of bad execution by the Knicks (specifically Rose, which is definitely possible), perfect defense (not likely), or perfect personnel (also not likely). As I mentioned before, those per possession numbers are lower in the clutch, but I believe the overall production is enough to justify the playcall. As a game winds down, it becomes harder and harder to score on a team level, not just on an individual level. Combined with the factors I addressed above, this play is a pretty good option considering the personnel on the roster.

The natural counterpoint to all this: guys who are ignored late in games (AKA everyone else) tend to tune out, which is valid. But I don't see many alternatives. Ideally, you'd like to get Porzingis involved, but that presents its own problems; teams have realized they can switch smaller guards and wings onto Porzingis and get away with it, as KP can’t yet beat those guys with ease. Using KP as a screener late in games can be a dicey proposition, one that limits the avenues for attack after the switch is made. Hernangomez (turnovers), O’Quinn (does nothing except rebound and shoot midrange jumpers), and Noah (lol) probably shouldn't be involved at all; after that, you're left with one of Courtney Lee, Brandon Jennings, or Justin Holiday. None of those players inspire much confidence in “must score” situations.

Point being, the constant use of the Rose-Melo two man game makes sense. Despite a ton of warts, both guys are gifted scorers, and they can both take advantage of mismatches (on a good day, at least). Getting your defense set behind the play and lowering the risk of a turnover are additional bonuses, even if that set defense is as bad as the Knicks’ is. At the expense of marginalizing players who don't really have any business creating offense at this point in time? It's a worthy trade off (note: this is from the perspective of trying to win games, as opposed to providing late game reps for young players). Basically, going this route is fine. They run other stuff as well. In an ideal world, though, you could certainly do better.