Jeremy Lin has been beyond excellent. It could reasonably be argued that he has been the best Rocket so far this season, and not just compared to expectations. When you combine all of his box score statistics (as is done by PER and WS/48), he comes out pretty much tied for second with Dwight Howard, just a bit behind James Harden. But some custom stats that I’ve been keeping suggest that those metrics are actually underestimating Lin.I have been tallying on paper what most fans informally tally in their minds: how often each player is making good and bad plays. (So geeky, I know.) I loosely define “good plays” as offensive plays that result in a high percentage attempt (e.g., passing to a shooter in the corner for an open three, even if he misses) or defensive plays that prevent the opponent from having a high percentage attempt (e.g., closing out quickly and challenging a shooter, even if he makes it). “Bad plays” are just the opposite. I also give tallies for particularly good/bad rebounds or loose ball plays. I have kept tallies for every game except the first. You can view the raw data in a Google spreadsheet here.
The advantage of these stats over box score stats is that they give a much more complete picture of performance, especially on the defensive end. There are so many things that the box score overlooks, e.g., effectively challenging a shot, forcing a ball handler into the strength of the defense, or making a great pass that the receiver fumbles out of bounds. All these important things are captured with my tallies.The disadvantage of these stats is that they’re more subjective and squishy. Some good plays and bad plays aren’t easily noticeable (e.g., Dwight’s mere presence may create extra spacing for Rocket shooters), and it’s sometimes – but not as often as you’d think – hard to assign individual blame on defensive breakdowns. I assign credits and blames as objectively as humanly possible, but I certainly have my biases. For example, even I don’t trust my own tallies on Asik because I have a hard time accepting that the man has ever done anything wrong.Back to Jeremy Lin. According to my tallies, Lin has been only slightly below Harden in terms of the rate of good offensive plays, and after adding up all “credits” and subtracting all “blames,” Lin comes out looking slightly better, as shown below.(Data objectivity note: I neither particularly like nor dislike Lin as a player or person, so I don’t think my judgment is clouded in this case like it is in the Asik case. If anything, I might be underestimating Lin in comparison to Harden, because I happen to really like Harden.)Harden hasn’t been fully healthy and has been on a bit of a cold streak, so I don’t expect this trend to continue, but I find it extremely encouraging that Lin has been playing at what we can fairly call Max-Contract level. I don’t think it would be crazy to predict that he will be the Rockets’ third superstar this season because I have no reason to believe he can’t or won’t continue at this pace. His major weaknesses from last year – shooting, defense, and left-handedness – all seem to have magically evaporated over the summer. What’s to stop him now?——Two other brief notes related to these data (but unrelated to Jeremy Lin):Pessimism: Dwight Howard is clearly not back to his dominant self. According to my tallies, in terms of both defense and rebounding – where Dwight makes his money – he has been just average, which means that he has been awful.Optimism: According to my tallies, against the Raptors Monday night, Terrence Jones had not only a good game, not only the best defensive and rebounding game of the night, not only one of the best games of his young career, but the best game of any Rocket so far this season, and by a non-trivial margin. I don’t believe there is another player on the Rockets’ roster who could’ve defended Rudy Gay nearly as well as Jones did. Not even Asik. He was remarkable.——P.S. – We could test just how objective (or not) these tallies are if we have multiple people recording them. I encourage you to try it for a game or two. If you send your data to me via a Google spreadsheet (firstname.lastname@example.org) in the same format as I have mine, I will do a reliability test and report back the results.