Beyond Assists: Why Derrick Rose Is A More Efficient Playmaker Than You Think

@obeast. Thanks and well said. People might be interested in the track this blogger is doing. He calls it "Opportunities Created". http://elgee35.wordpress.com/2011/02/22/opportunities-created-oc-leaders-2010-playoffs/ Also, for those Rose supporters out there. His '10-11 OC numbers look quite good for Rose so far. Note that this is just one dude, so while his playoff tracking was complete, this stats here for this regular season are based only on the games he's had time to track: Player - OC (Games) - Ast/36 in those games Nash 18.4 (4 games) - 12.4 Rose 15.6 (7 games) - 8.0 Kobe 14.3 (12 games) - 7.0 Westbrook 12.5 (5 games) - 9.9 D. Will 11.8 (5 games) - 10.5 (And no, Rondo's name wasn't forgotten. He's way down on the list, which both makes sense, and should make everyone personally consider exactly what this stat measures.)

An assist that leads to a 3 pointer is worth 3 points. An assist that leads to a layup or dunk is worth 2 points. You don't count assists for unconverted attempts. That is all.

Synergy tracks like every play, don't they? I mean assists are already subjective, and a fairly meaningless stat with respect to evaluating NBA players (as opposed to the eyeball test). What it's likely more useful for is poring over the stats of a bunch of prospects you simply don't have the time to see, and determining who deserves a closer look, with more subjective but more useful passing stats, sorted better categories than just trying to manipulate a box score in every conceivable, instead tracking "good passes", "passes leading to a shot", "passes leading to a contested shot", etc. Unfortunately, those stats will not be tracked for unscouted players. If you really need more for NBA players, I just don't think you can get it out of any combination of the available statistics discussed above.

How would you respond to this? http://www.basketball-reference.com/blog/?p=8946

Christian, you're missing his point -- the assist metrics in question can't distinguish between, say, 5 assists on 5 possessions from easy layups and 5 assists on 15 possessions from contested jumpers. It's like the difference between a scorer going 5/5 and 5/15. Efficiency matters. What's missing is some way to account for "assist attempts per assist", which definitely will vary with what you're caling "difficult of shot created".

Sorry Matt (Johnson) but you're simply incorrect. A pass creating a bad shot opportunity will result in lower assist numbers automatically as fewer of those shots will go in. So counting points created is without a doubt the right thing to do. We're talking statistics here, not beauty contest. There are data one might consider controlling for - particularly team (minus assist-er) quality however you'd want to measure that, but "difficulty of shot created" ain't one of them.

@Roboto: You're talking about a different point than I was when I responded to G-Dogg. He didn't understand why you wouldn't simply say that the value of the assister's action is in direct proportion to the amount of points it produces. So I laid out that the true value of the assister's contribution is trickier to gauge. I did not mean to imply that all 2 point assists are open layups and all 3 point assists are contested.

@MattJohnson: why not just compare it this way? Would you rather have an uncontested 3-pointer or a semi-contested 18-footer or a fully contested lay-up at the basket? Obviously, that depends upon who's shooting, how likely they are to get fouled, and how likely they are to make the FT shots they may take, and so if the uh-sister played an equally larger [sic] role in creating any of the three situations, it's the would-be basket plus any FTA*FT% that results in the most points that should get more credit. When D-Rose gets people to collapse on him and gives Korver an open 3 point shot rather than passing it to Omer Asik for a contested layup/dunk attempt that may or may not result in a bucket and/or one or more 56% likely FTAs (and/or just a missed/blocked shot with no foul), I'm hoping he passes it to Korver EVERY TIME! @RyanSchwan: Rose may average "only" 2.32 assists [not passes] per turnover but here are the scrub PGs who average fewer A/TO: Westbrook, Stuckey, Jennings, Billups, Curry...none of whom are the #1 scoring option for their respective teams...so they SHOULD be passing more than Rose as a percentage of their possessions. Also, the following PGs have lower usage rates than Rose, but have higher TO/48 Minutes: Nash, Westbrook, Wall, Rondo, D. Williams and D. Harris...none of whom are primary scorers for their respective teams, so they should be LESS likely to bounce the ball off their knees or feet and out of bounds (a.k.a. the Ben Gordon move). Their turnovers likely come from bad passes. Furthermore, the following PGs have higher TO Ratios (TOs/player possession) than Rose: Rondo, Wall, Nash, Curry, Billups, Mo Williams, J. Holiday, D. Harris, Collison, Kidd, Evans, Nelson, Westbrook, A. Miller, Ridnour, Felton, D. Williams, Lawson, Arroyo, Sessions, Calderon, Conley, Lowry, Barea, Parker (and 23 others). Rose is 49th of 65 qualified PGs.

Why not come up with the 'points resulting from assist' stat to show which player's assists mean more?

This is an interesting article, but a couple "hold ons": "Rose sees a 13.6% increase in his assist numbers when factoring in three’s" Using % here can confuse matters. That increase means an improvement of 1.1 assists. Among the guys in the league who get more assists than Rose, all of them have an improvement of 0.8 or higher except Calderon. So his net gain over the weaker elite guards on this from is 0.3 APG or less. In and of itself, such a gain is not particularly important - though I'm not saying it can't be used as part of the more general argument. @G-Dogg. The ideal measure of an "assist-er" isn't whether the ball actually went in, which is very much dependent on luck and another player's skill. It's the opportunity that the passer helped create or improve that's important. Obviously in practice there's a variety of issues, but just compare it this way, would you rather have an open layup at the basket or a semi-contested 3-pointer? Obviously the former, and so if the assist-er played an equally larger role in creating both situations, it's the layup assist-er who should get more credit.

Matt's comment makes no sense. You don't get an assist unless the shot goes in. If you pass to a 3-point shooter, you get an assist 34% of the time, so yes you have to make more passes to get credited for an assist there. But, every assist on a 3-point shot is worth 3 points. So if you are looking at assists, those assists resulted in points scored of between 2x and 3x, depending on how many were 3-point baskets. Ryan's comment also is off base. Rondo in fact leads the league in turnovers per game at 3.8. Nash is at 3.6, Rose at 3.5. (Paul has one of the lowest PG TO rates at 2.3.) Turnovers come not just from passing, but also from driving to the basket, which Rose does more frequently than most guards, and which is reflected in his higher scoring average. So simply comparing turnovers to assists is not meaningful.

Until stats become more detailed and analysts determine how to assess the value of being able to create ones own shot, you'll really have to rely on just watching for players like Rose more-so than other types of players. The general way to put this is that giving Rose the ball seems to generally lead to pretty good scoring opportunities for his team. They might not be the result of a Rose assist, or bucket, they might come at a low percentage when the alternative was even less likely to work, they might work because Thibbs is a great coach and Rose is the primary beneficiary, etc, but it becomes very clear when watching him play repeatedly. That said...Dwight Howard is my MVP. He's just playing insanely good basketball, and none of his teammates are pulling the weight of Deng, Wade, and the other teammates of MVP candidates.

Assists+ ignores would be assists that lead to fouls. While Rose's assists may lead to a plethora of 3 point opportunities, shots taken from beyond the arc are converted at a much lower percentage than point blank shots and rarely lead to fouls. Point guards that create layups and dunks also create lots of fouls which have, perhaps, an even larger impact on the game than 3's.

As a Bulls fan, this data confirms what my untrained eye tells me while watching the games. I've also noticed it seems like Rose makes a lot of passes to Boozer et al. inside that don't lead to a basket but to foul shots. Is this a stat that anyone tracks? I'd be curious if Rose does this more than his peers...

AST+ is a fun stat, but it hardly means anything. The average increase due to AST+ is around 9% for PG in the league - so Derrick Rose's 13.6% is worth all of a fraction of a point per game over the average. Then, of course, there's the fact that for every 2.2 passes he makes, he turns the ball over once. That's a number that doesn't even rank in the top 30 among point guards in the league. Players like Rondo, Paul, and Nash generate an extra assist over Rose before generating a turnover - and, of course, a turnover costs your team points, the chance to offensive rebound, and frequently generates easy points going the other way.

You have be careful when doing this. Is an assist that leads to a 3-pointer (which is completed with a probability of about 34%) really more valuable than an assist that leads to a dunk or lay-up (which are completed with a probability of about 65%)? No. They're not. The 3-point shot has an expected value of about 1 point while the lay-up/dunk has an expected value of about 1.3 points.