Tom Haberstroh recently wrote a piece about the value of different types of assists. As someone who has always found it annoying when people talk about how "he would've made more assists if they hit their shots", I think this is pretty enlightening. It's a long read though. http://www.hardwoodparoxysm.com/2010/03/10/nba-hd-adjusting-how-we-measure-and-view-assists/ To sum up for those who aren't interested in reading, basically he's advocating for making each assist at the rim more valuable than 1, because it's up to the passer to make such an opportunity available. Whereas perimeter assists are generally less valuable because it relies more on the shooter to make the shot. In other words, great passers CREATE EASY SHOTS for his teammates. They don't just distribute. It's one thing to pass to a player and have him make a jumpshot for an assist. But another to pass the ball to him under the basket for a layup(or even fouls which doesn't get counted) Another note: I must admit. I'm not a fan of Lebron in general. But I must admit more and more I simply find the man is simply AMAZING as a basketball player.
There's the graphic that coulda been added to maybe get people talking... "See, Kyle Lowry has BETTER ASSIST than Aaron Brooks. He plays like a TRUE point guard...." Some of it is a little dependent on who's receiving the passes. Like he pointed with Kidd passing to Dirk who's a big time jump shooter. Compared to passing it into Shaq who's an inside dominator who can help anyone's close-in adjusted-assists. Jason Williams dumps it into Dwight Howard, which might explain how he's #2 on the list. Its plainly stated where he says Anyone who's played ball can know its takes a different knack for passing into the INTERIOR of a defense. Driving to the congested paint or passing through all the congestion. Requires a little more court vision and creativity there.
I hate to bring up TMac on this board. But I remember Morey once said that McGrady was the best wing player other than LeBron in passing the ball to the most efficient scoring shots. I think what Morey tracks is even more than what they are doing here in the article. They are talking about MADE shots. Morey probably tracks the passes to set up GOOD shots even when the shooters miss or are fouled.
while i like the idea of not treating all assists as equal, i feel the methodology is off, if we're just essentially trying to find the efficiency of each pass by the passer. his method weights made field goals (assists) by the average value of each attempted field goal and this doesn't make sense to me . it also gives too much weight to inside shots. i would think a truer measure would be points created per assist attempt. for an assist attempt, you would just take the number of assists at each spot and divide by field goal percentage. so on inside shots and 3's, lebron would have 3.9/.608=6.4 attempted assists inside and 2.6/.354=7.3 attempted assists on 3's. they would be worth 3.9*2+2.6*3=15.6. his points per attempted assist would be 15.6/13.7=1.13 just for those 2 areas. doing this for all of the passers, lebron easily leads everybody in the top 10 in assists with 1.056 Pts/Att Ast (PAA). devin harris is closest with 1.022 PAA. so lebron is significantly more efficient passing than the other big time assist guys. in the top 50 only stephen jackson (1.087), jason williams (1.063), mo williams (1.065), jameer nelson (1.060), and raymond felton (1.060) beat out lebron. good news is lowry is essentially tied with lebron at 1.056.
Players who excel at getting assists are the ones who recognize their teammates' talents and how to utilize them to the best advantage. So a PG who recognizes that one of his best perimeter shooters is open and passes to him for a wide open jumper that the teammate will most likely make is every bit of an assist as an alley oops pass for a dunk.
I'm pretty sure Morey even commented before about tracking missed layups and fouls. But since these data aren't available to the public, it's hard for fans to use them. If the NBA does track passes that leads to shooting fouls, I'm sure he'd have included them somehow. As for T-Mac, I also LOVED his passing while he was here. It's a shame about his health. *sighs* I'm not in the calculating mood right now, so I'll come back to this. But upon cursory glance it doesn't seem like you're doing anything different from him. I'm not sure how he's overvaluing inside shots, when they're converted at such a high rate.
Meh. His reasoning seems sound but the problem here is how he overvalues shot at the rim. A simple rebuttal would be his Cp3 example. Sure an alley oop would probbaly be more difficult than a simple pass to the 3 pt shooting dude, but Peja isn't exactly Mr. creative offense. He has to be wide open or else he won't be able to get that shot off. Cp3 isn't just passing the ball to him and watching rafer style, he's actually making the defense collapse then hitting Peja for the open shot. If Peja accidentally steps on the line suddenly that's a "low quality" pass and it's Cp3's fault? It doesn't even take into account player bias. A good example would probably be Dirk and Kidd. As the article stated, generally the mid range shot is the worst shot, however players like Dirk and PP live on mid-range shots. That's their bread and butter, they can hit those "difficult" shots at a high percentage, and probably hit it higher than then the 3 or in the post. In cases like that is it Kidd's fault he's giving the ball to Dirk in his comfort zone? Mo Wiliams has a higher % increase than Kidd, does that mean he's actually a better assister than Kidd? Hell no! It just means that Kidd passes to Dirk, Terry and Butler while Mo Williams passes to Shaq, sideshow Bob and LBJ. IMHO why are you weighting based on difficulty anyway? At the end of the day the mid-range shot and the alley oop all result in the same 2 pts. Should you start weighting those over the backboard shots Kobe took as well? Certainly they're more difficult than the layups he usually makes
roslolin: So in the following extreme example, do you consider both players equal passers? Player A passes the ball on 20 possessions to shooters for a long 2. They make 8 shots. He gets 8 assists. Player B passes the on 12 possessions into the paint, and the finishers makes 8 layups/dunks. He gets 8 assists.
And before anyone pipes up, understand that that "8 makes in 20 attempts" hypothetical is the usual success rate you anticipate for shots from this area.
Seems like the assists are already weighted, though, because you only get credit when the ball goes in the hoop. If you pass to a perimeter guy for a long 2, you're really only getting, on average, .4 assists, because that guy will miss the shot more often than not. If you pass for a dunk or a layup and the percentage chance of that shot going in is 90%, you're gonna average .9 assists on those passes. If anything, I'd like to see a system that rewards passers for "points as a result of the pass". Player A passes to player B in the post, and the defense is forced to foul him on the shot to prevent the easy 2, player A gets no assists even though his pass led to 2 free throws. Player A drives, draws the D, and then kicks out to player C outside for the wide open 3, player A only gets 1 assist even though his good pass led to 3 points rather than 2.
I don't watch every NBA team play, but I'm sure that there are some pretty terrible ones. With that in mind, as another extreme example, if you have 4 Alston-type finishers in the paint, why wouldn't you pass it out to the perimeter? You can't assume for example, that because a guy didn't pass it to Chuck Hayes in the post and instead hit Battier on the perimeter that it was a bad play, assuming that Hayes is guarded at the time and Battier isn't. It's also a combination of the roster, and as a result, what the defense gives you.
Well the problem here is you're penalizing player A for something out of his control, and rewarding player B for something which may or may not be his good fortune. Player B might have been more efficient but I would counter that Player A created 20 possessions or "opportunities" for his team. The most player A can do for his team is give his guys a good look, he can't take over their bodies and make the shot for them. On the other hand Player B might have been playing with Hakeem or Kareem Abdul Jabbar. If you're playing with players who have elite post skills then this distorts your data. You end up with 8 "in the post" assists but all player B is really doing is passing the ball to Kareem and watching him sky hook the opponent to death. A good IRL example of this would be the % change between Jameer Nelson and Jason Kidd. According to the data, Jameer Nelson makes higher quality assists than Jason Kidd, however everyone knows Jameer Nelson (more known for his scoring ability than his passing) is half the passer Jason Kidd is (considered one of the best passers of his generation). The difference is the Mavs are composed of jump shooters while the Magic have Dwight "give him the ball and watch him dunk" Howard. The point of doing this analysis is to determine which assists are better; however if the rosters skew data this much then IMHO its not worth doing. I'm not saying all assists are equal because they're not, but IMHO this methodology is the wrong way to go about it.
he's weighting all attempts only by the successful conversions. it would be like figuring out the most efficient shooter by only seeing how much each made shot was worth without taking into account how many attempted shots were taken from each position. a guy who went 2-6 from 3 would be considered more efficient than a guy who went 4-6 from 2 because all his made shots were worth 3 to the other person's 2, but player 2 clearly scored more per attempt. this is the problem with the article. because inside shots are so efficient (60.8%), each assist to the inside represent a relatively small number of passes compared to say 3 pointers or long two which are converted at less than 40%. lets say you get 1 inside assist and 1 "long 2" assist per game. by this guys methodology, your efficiency would be calculated by giving a 50% weight (1 out of 2) to inside shots and a 50% weight (1 out of 2) to long 2's. but what really happened? to get that 1 inside assist, you threw 1.6 passes. to get that 1 "long 2" assist, you threw 2.5 passes. so in reality, you are really throwing 61% of your passes to set up "long 2's" and 39% of your passes for inside shots. that's how it should be weighted because that's the ratio at which you are creating these opportunities. but they are 20 low efficiency opportunities. that's the point. while it's true that you can only throw passes to guys where they are and if someone like nowitzki is at low efficiency spots, then that's where you have to give it to him and that might hurt you quality wise, but it helps you quantity wise. jason kidd can get assists that aren't hard to create by just giving it to a guy 20 feet from the basket and letting him do the hard work of making it. why should he get the same credit as lebron for driving and drawing everyone before handing it off to varejao for an easy layup?
I just want to clear up something for a few of you who aren't fully understanding the nature of the article. I think an analogy should be made to points per game. You know how we tend to differentiate between volume shooters who put up empty stats vs guys who gets their points efficiently? How Morey didn't just look at Kevin Martin's ppg, but rather his TS% which shows how few shots he needs to get his 20+points? Well, people don't tend to do the same thing with assists. A player make a lazy pass to a guy standing 20 feet away from the basket. He chucks it up, makes it, and there's an assist. But we all know from watching the game that this was hardly a pass worthy of such a distinction. While people break down points scored, very few people do so with assists. A player who averages 8 assists will be compared similarly with another player who averages 8 assists. The article did not mention the best alternative, as pointed out by many. But I wish to put this article so people can at least see that there are many types of passes that results in assists, and they shouldn't be counted the same. On another note, I believe Morey once wanted to track a state for the pass before the pass that leads to the basket. For us older Rockets fans, the stat which shows how great certain post passers are like Hakeem and Barkley. Guys who passed out the double team perfectly, but never has anything to show for it because it's the next perimeter pass that leads to the basket. Basically, many players' passing skills are overrated or underrated as a result of people only looking at the assist stat. I see what you're saying now. Nice work.
But the point is not everyone in the low post is a cutting varejao. in the same way that not everyone can be a nowitzki and get the ball 20 feet from the basket and do something with it. The opposite is true as well, if the shooter is a stiff who can't create for himself and the low post option is an capable scorer than obviously the higher quality assist is the outlet pass. As I said already, Cp3 who drives in the middle of the lane then hits Peja for wide open long 2 is much more difficult then say Tony Parker throwing the ball to TD who does his patented turnaround bankshot. Assists aren't cut and dried like pts because you only get one if the player you passed it to makes their shot, if they missed the shot then its like your assist never happened so IMHO if you're gonna use advanced statistical analysis then its better to forego playing around with the assist figures and look for a better indicator. I don't even get your bolded part. You yourself agree that assistors (is that even a word? lol) MUST give the ball to where their guys are but you go ahead and penalize/reward on where they give the ball anyway. You can see the flaw in the data itself, LBJ, Mo Williams, Jason Williams and Jameer Nelson are all among the league leaders in the percentage increase. Considering 2 of these guys are known more for their scoring abilities and Jason Williams doesn't really strike me as the efficient passer, don't you think the fact that they have DH or Shaq and LBJ on the squad has something to do with how high their assists % increases are? On the other hand Jerry Bayless, Kirk Heinrich, Derrick Rose, Jason Kidd and Barea are among the league leaders in the % decrease. Doesn't it seem that Chicago and the Mavs being a jump-shooting team has something to do with it as well? What's the point of having statistical measurements whose results seem to be heavily affected by noise? If this is the methodolgy then I'd rather just look at the simple assists.
