Thursday, May 10, 2012

Level of Fit: Carmelo and Amare in New York

Analyzing basketball with the same criteria and weight for every team is, obviously, problematic. What one team needs is not necessarily what another needs. Another way to think about it is that there are diminishing returns for basketball skills on a team. A sharpshooter who can only shoot three's is more valuable to the 76ers than the Magic. Even in more sophisticated basketball studies this can be lost. If there's a team with no reliable outside weapons except one player, then when that specific player is on the court the team gains a benefit that is usually larger than when that player is on the court for a team that already has shooters. In NBA analysis we assume that there's a Platonic "true" level of value for a player, a concrete number we can find with better statistical models. Instead I think the true value of a player is intractably linked to his teammates, and like a fluid that fits the outline of its container rather than its own inherent shape.

One of the biggest unresolved issues in the basketball stat community is the issue of usage. How valuable is the ability to create a shot, even if the shot is not efficient? Shouldn't it vary by the team? First of all, there's the problem of the shot clock and that sometimes you need your playmaker to bail you out with a shot at the last second. Beyond that, how efficient scoring is defined is usually on a league-wide basis when it's actually the team-wide level that pertinent. If player X's entire team shoots below the league average percentages except for guys who can't create their own shots and live off offensive rebounds or open dunks, but Mr. X is the most efficient scorer, then it is more efficient for that player to take more shots. A large number of teams are close enough in their shooting percentages that it isn't an issue, but in some cases it's important.

I think that's enough of the hypothetical player talk. The best real world example of an active player for this discussion is Carmelo Anthony, partly because he's a big name superstar and he's in New York. He's known for his skill in scoring one-on-one, which according to various reports from Team USA and other basketball players he's the best in the world in that particular aspect. While NBA stars aren't always the best assessors of talent -- Dwight wanted Orlando to recruit Glen Davis and Stephen Jackson, and Michael Jordan can now say he was part of the best and worst teams of all-time -- there's no denying his skills.

But how do you best use Carmelo Anthony? If you concede that Anthony is one of the best at playing one-on-one, and thus doesn't need a lot of help to score, then you agree that a more efficiently designed team employs players who are best at impacting the game without scoring. Not saying everyone else should be terrible at scoring; rather I'm saying the rest of the team should be designed relying on the non-scoring aspects. Essentially, surround him with great defensive players who can crash the boards, hit open three's, and give him one good offensive option, a Robin to his Batman. This isn't exactly innovative, but a lot has been written about how the Knicks can get better and few focus on how to use Carmelo's skill by surrounding him with defenders.

2011-12 (
The table on the left displays the non-PGs with the lowest percentage of shots assisted. Carmelo's true-shooting percentage for his career is near 54.4; he's usually a little above the league average. Of course, you'd rather have LeBron creating plays for you, but there's only one LeBron to go around. There's obviously more to this -- turnovers, passing, positive effect on teammates, etc. -- but it gives you a general idea about who's best at creating shots without help. Carmelo in this respect is elite, able to initiate offense by his own will, and on a basketball court with advanced defenses and the shot clock this is useful.

The Knicks, to most people's disbelief, were one of the best defensive teams in the league, surrendering only 98.4 points per 100 possessions a game, ranking fifth in the league. Their offense, ranked 19th, was submarined by an unusually poor shooting season from Carmelo and an atrocious one from Amare. Speaking of Amare Stoudemire, New York was hoping he'd be the perfect offensive partner for Carmelo, but it was a shortsighted pairing. Amare's best at receiving passes and playing the pick-and-roll; Carmelo's best in isolation. Jeremy Lin, a godsend, can operate with Amare, and in terms of playing with Carmelo he can at least be a viable second option if the other team is doubling Carmelo or he's having an off-night.

The problem, actually, is that they built the wrong team backwards -- Carmelo on a defensive team is a good idea, but Amare is the wrong partner. For one, Stoudemire is a terrible defender, and defensive metrics have him pegged consistently as one of the worst over the years. Carmelo's decent when he tries, but it's not like New York will peg him as their stopper. Next to Amare, however, he's more exposed as it's harder to hide both of them at the same time when they play adjacent positions. Also, New York has had success with him at power forward, where he's also played in the Olympics. He's called a one-dimensional player, but he's one of the best rebounding small forwards in the league; sliding over to power forward isn't as damaging to the defense as it seems.

In retrospect, signing Chandler was the perfect move, as last year they were ranked 21st in defense. There are other reasons for the improvement, like Shumpert, but Chandler is the perfect complement for Carmelo because he can guard the basket and clean up missed shots. I'd argue that the amnesty should have been used on Amare, but I know that's not the sort of move an organization can make. Their target of Billups with the amnesty, however, was a sound decision.

New York will want to know how to proceed with its nucleus and build on their 7th seed. Defensive player of the year Tyson Chandler is obviously their defensive anchor. Lin and Shumpert, if healthy, are good pieces for the backcourt. Toney Douglas is also a decent defender, but his offense has been so terrible it's not worth the trouble. There's not too much additional help in the frontcourt. Jeffries is a solid role player and Harrellson could be useful, but Stoudemire will receive most of their minutes. They ideally need to replace Stoudemire with someone who's decent at offense like a reliable jumper and some post moves, but who's also a pretty good defender. Not asking for an all-star here -- just a better level of fit. Lin can provide the scoring/play-making help as well. They also need a wing defender who has a reliable outside shot. Those guys are precious commodities for NBA teams, and champions usually have a couple. Courtney Lee, for example, is one guy you can get for cheap.

If Carmelo Anthony is to win a championship, paradoxically it will come on an elite defensive team, as in top two in the league, where he can use his shot-creating skills on a team who finds his isolation play most valuable. Amare Stoudemire was the first major piece of this incarnation, but he's also the worst fit. The website stats-for-the-nba, which uses a sophisticated model for +/- stats, found that Carmelo and Stoudemire are a pretty good offensive pair at +0.9 for essentially a whole game but lost that value on defense with a -0.8. The same website looking at the players individually gave Carmelo +1.3 overall and Amare -.4. There's definitely evidence they don't work well together, and it's not a minority viewpoint. However, keep in mind that players aren't static vessels, and their interaction and utilization changes how "good" they are. Carmelo Anthony and the Knicks are entangled now, and further improvement will come from coherence, not more star power.

No comments:

Post a Comment