Is Jeremy Lin's Turnover Problem a Problem? Probably not.

Cheer up, Jeremy. The City's got your back.

Amidst all the Jeremy Lin hoopla the past fortnight, just about the only concern that people seem to be raising about his game are the apparently large number of turnovers he's producing. Through his last game against Sacramento, it was 5.1 TOV/36 minutes. Seems high, right?

Think about this, though. Jeremy Lin has the ball in his hands A LOT. Right? I mean, that much is obvious. That's been the whole problem with the Knicks is that they didn't have anyone who could handle the ball. So Jeremy has had a lot of touches. We don't have a real-time source for touches, so nobody really knows exactly how much the ball is in his hands, but it's got to be up there among the league leaders per minute (or per possession) right now.

So, the question we should really be asking is whether Jeremy Lin's turnovers are *adding* to what we would expect the team to have if he weren't in the lineup. To answer this question, we can calculate his adjusted turnovers (turnovers at the team level per 100 possessions), in the same way that adjusted +/- is done, and just as I've been doing adjusted versions for the various shot locations recently. In fact, Jeremias Engelmann was the first (and only?) person I know that has calculated adjusted turnover rates, which you can find here for 2009-2011. If you look on there, Lin has a 0.0 offensive rating and a +0.7 defensive rating. To give a comparison, Stephen Curry (also playing for the Warriors) had a +0.1 offensive rating. So, at least, through his rookie season, there was nothing to suggest that Lin had a terrible problem with turnovers (and was in fact a net positive on the defensive end).

Those figures were through last season, so I decided to run an updated version from 2010 through yesterday's games (basically 2 1/2 years). This would include all the playing time for Lin this season in NYK. Instead of running a ridge regression as Jerry did, I used a regression technique called lasso which is used to select factors in a model that tend to have the largest effect. All other factors are set to zero.

What I found is that Jeremy Lin's offensive adjusted turnover rating was set to zero by the regression, meaning that he doesn't have a significant effect on the turnover rate at the team level. However, on the defensive side, his rating is +1.14, which suggests not only does he not have a major problem with turnovers, he's actually helping his team more often than not in *creating* turnovers by the opponent. In fact, a lot of the Knicks are. Iman Shumpert, Jared Jeffries, Josh Harrellson, and Toney Douglas are all net positives on defense and rank in the top 20. If people would pause long enough to look, they would see that while New York is near the bottom of the league in coughing the ball up (my guess is much of that due to pre-Lin play), they are actually second in the league behind Memphis in opponent turnover rate. So I'm really not that worried about Lin's offensive turnovers, and I think you shouldn't be either. And that's not even considering that he is getting the first major minutes of his career, and has (presumably) time and talent to improve.

Here are the results of the regression, for those who are interested in where everyone stands.

Offense

Google Spreadsheet

Defense

Google Spreadsheet

14 thoughts on “Is Jeremy Lin's Turnover Problem a Problem? Probably not.”

  1. I think you'd have to agree that the 'average' estimate is a low bound on his current turnovers, though. The 5.1 TO per 36 is over double his rate from last season when he was getting garbage time, and his TO% is up as well. He's had at least six turnovers in every game he's started except for against the Wizards. Now, he obviously brings enough to the table that it isn't anything to sweat too much, but it's certainly a weak point.

    1. "I think you’d have to agree that the ‘average’ estimate is a low bound on his current turnovers, though."

      Why would I *have* to agree with this? He's already played more possessions this season than he did all last year. And for a team that generally creates a lot more TO's than GSW.

      1. Well, I guess I should have said "I imagine you'll agree...". Your data set covers over two years, which includes time in which Lin certainly appeared to be less turnover-prone than he is now. Including that time means that your estimate is probably a lower bound on his current ability. What we're interested in is if he's turning over the ball too much now, as the starting point guard for a team. In that sample, albeit small, the answer would appear to be yes.

          1. Your analysis is estimating his turnovers including time in which we're pretty sure he didn't turn the ball over as much as he does now, whether we use TO per 36 or TO%. So it seems safe to assume that right now, as a big-minute point guard and not as a garbage time guy, he's turning the ball over more than your method suggests. Maybe it's only a little bit, since his current streak makes up about half of his data, but it still seems like the likely conclusion to me.

          2. Given the small absolute difference in tov% between this season and last season, and the rather larger ball handling responsibility and touches he's getting now, I find it difficult to agree with you that this is a lower bound. If anything, I expect that when Melo comes back and handles the ball more, Lin's turnover "problem" will be even less of a problem than it already is.

            Time will tell.

  2. I'm not sure I trust this yet.

    One of the reasons the rest of the team's TOs are not up is that the two biggest culprits on the Knicks (Amare and Melo) have been out. When the Knicks have Amare, Melo, and Lin on the court, some of the TOs previously committed by those other guys may shift to Lin because he'll be handling the ball so much, but I suspect the overall TOs will be up unless Lin does a better job.

  3. Amare and Melo both turn the ball over a lot relative to the guys that have been playing in their place. Lin's TOs will probably go down when both are playing because he'll handle the ball less often, but they may not go down as much as Amare and Melo go up relative to without them.

    1. Carmelo is 8th in adjusted offensive turnover rate. Might want to reconsider your opinion of "a lot". As for Amare, the regression found him to contribute 0.0 adjusted turnovers. Not good, not bad.

  4. All I can say is that I watch every Knicks game. Every time Amare gets the ball in ISO and puts the ball on the floor he's a huge threat to get a TO. Melo's TOs are different, but he gets a lot of really stupid ones making plays for himself also.

    Since both these players will continue doing pretty much the same things on offense they have been doing even when Lin is on the court, IMO they will continue getting "close" to as many TOs as they get now. It's not like they get TOs trying to make passes to set other people up or to penetrate to makes plays for others like Lin does. If that was the case, their TOs would shrink because Lin would be doing that job.

    IMO unless Lin improves, the team total will rise. The question I have is by how much. There will be some tradeoff just because Lin handles the ball more, but it's not going to be 1 for 1 when you have Amare and Melo

    1. W.C. it's not clear to me that you understand the point of doing the regression. But here's the bottom line. Take whichever 5 Knicks you want (presumably the starting unit) and add together their offensive ratings from the spreadsheet. That total is a good prediction of how many TOV they will have per 100 possessions relative to the league average. The worst Knick on the entire list is actually Tyson Chandler at +1.9. Carmelo is -0.96. Lin, Amare, and Fields are 0. The total is +0.94, meaning that unit should average about +1 TOV more than the league average. And hey, what do you know, that's just about where they are right now. League average is 14.2 per 100, and the Knicks are currently at 15.4. How many do you think it will be when Melo comes back? The worst team in the league is OKC at 16 TOV/100 possessions.

  5. Can't really compare what Melo and Amar'e did before Lin playing without a PG to what they will do with Lin running the floor.
    I've suspected that Lin's turnover problem is not nearly as bad as it seems, and not terribly reflective of his true ability. He is running D'Antoni's system which is very point guard-centric and inflates both their assists and turnovers. The thing is he has not been working with the perimeter shooters D'Antoni requires for his system to blossom. The Knicks are one of the worst 3 point shooting teams in the league at about 30%, and Lin in this stretch has been playing without Carmelo, who is a threat to hit from any range, Josh Harrellson who has shown 3pt ability at 35% this season and Toney Douglas/Baron Davis who are (or were) good 3pt shooters. The only consistent 3 point shooter during this run has been Steve Novak and you can't count on just one shooter day in and day out. Bill Walker and Landry Fields were black holes from range - in fact Lin himself was probably the most efficient 3pt shooter besides Novak during this time. Having "scouted" Lin somewhat in his performance this season I found that he is extremely efficient at finding open men behind the 3 point line, with 32 assists to 4 turnovers, but he should have a lot more assists on a healthy Knicks squad that will be welcoming J.R. Smith. Improving 3pt% will allow the team to spread the floor, making it easier for Jeremy to drive into the lane when necessary. As of now it is where he gets the bulk of his turnovers - over half of the time he's coughed the ball up, Jeremy has either mishandled it due to defensive pressure when trying to dribble to the hoop or he's made a poor pass in traffic to a big man in the lane, generally leaving his feet to do so and try to make the ill-advised acrobatic dime. The Knicks' defense is too one-dimensional as it stands now and since Lin is the man designated to operate it he is the one who pays in his personal statistics.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>