Is there a stat like this?

[-Kees-]

It's hard to explain, but is there a stat that measures how many assists a player WOULD have, had all shots been made that he passed to? Like say a player has a low APG or AST%, but his players around him all shot 30 FG%, but a similar player averaged higher APG and AST% because all of their players shot 60 FG% or whatever. Is there a stat that distinguishes the difference, and takes account FG% to how many assists a player could have?

[Chicago76]

It sounds like what you're trying to do is figure out how many APG a player would have if his teammates could shoot the league average to normalize things a bit. There isn't anything like that out there. A few reasons why:

1-Team FG% doesn't vary that much, so even a poor shooting team with a PG who gets 7 apg may only get bumped up to 7.5 or at the most 8 apg if he played with poor shooting teammates due to the rotation. There isn't a huge difference to normalize there.

2-Part of a PGs job is to get his teammates high percentage shots, so in a sense you'd be penalizing the PG who does a good job of this because it results in higher FG% for his teammates.

3-Teams that shoot a lower FG% also tend to shoot a lot of threes. Threes are easier assists to make than a thread the needle pinpoint pass, so you'd also be rewarding players who make easy assists to a wide open guy 26 feet from the basket at the expense of a really good interior or transition passer.

You're better off just relying upon AST% to help you figure out who the better passers are.

[rrravenred]

What I want is a measure of passes leading to a shot.

If you could narrow it down to individual players and to their "all other players passing to" shot percentages as opposed to when they're passed to by a particular, you could work out some interesting synergies.

[akennedy13]

This is kind of different but I know that John Hollinger did a study this offseason on assist value. Basically finding out which player's assists ped to the highest quality of shots. For example, a player like LeBron had a lot of assists that occurred at the end of the shot clock when he simply kicked the ball out to the corner and had a teammate bail him out with a corner three, resulting in an assist. It's a similar stat that I think you're looking for, but not identical. Although I haven;t found anywhere where you can find this stat used and sorted among players.

[J~Rush]

I'd also like to see assisted free throw attempts.

[EvanZ]

What I want is a measure of passes leading to a shot.

If you could narrow it down to individual players and to their "all other players passing to" shot percentages as opposed to when they're passed to by a particular, you could work out some interesting synergies.

I did something like this over the summer:

http://thecity2.com/2011/07/29/warriors ... g-synergy/

[rrravenred]

Awesome, Evan. The math is unfortunately over my head, but if I'm reading this correctly, you're saying that specific passers have almost no effect on the likelihood of a player making a shot (though my impression's always been that the Warriors don't run traditional PG play)?

[Chicago76]

^That's what I'm getting out of it too...with one caveat. We're talking about spot up jumpers here. Those are passes that are intuitively easy to make where the burden of scoring is on the shooter rather than the passer. Better passers might find more of those opportunities (volume) but that won't be reflected in the efficiency.

This makes sense to me. The real value of a good passer is in threading the needle for better shots.

[EvanZ]

Awesome, Evan. The math is unfortunately over my head, but if I'm reading this correctly, you're saying that specific passers have almost no effect on the likelihood of a player making a shot (though my impression's always been that the Warriors don't run traditional PG play)?

Because the sample sizes are small (overall, I tracked over 1000 shots/passes, but for each player the sample size is much smaller), it's not easy to claim one way or another. In some cases, however, there does appear to be a difference. For example, Dorell Wright's PPP was 1.26 after an Ellis pass (123 shots), and only 0.92 after a Curry pass (88 shots). That's a fairly big difference and fairly large sample size for each combo.

[Chicago76]

I don't get to watch a lot of GS games, so this is more of a question based upon shot tendency that I wouldn't know a thing about:

Do you think the efficiency from an Ellis pass vs. a Curry pass is due to the quality of the shot or the quality of the "look" for the shot due to the way the defense is playing Ellis v. Curry?

Simplifying things to a three man game between Curry, Wright, and Ellis:
When Curry drives, he's flanked by two guys who are roughly .550 eFG% three point shooters. When Ellis does the same, he had Curry at .663 3 pt eFG%, which is a 0.2 pps advantage over Wright. Curry's defensive assignment is most definitely staying home and not helping. If any perimeter help D is coming, it's going to be from Wright's man, which means theoretically Wright is generally more open when Ellis drives than when Curry does the same, so you'd expect his FG% off Ellis passes to be higher than those from Curry.

In that sense, it might not be the quality of the pass so much as it is the quality of the look due to the floor spacing that Curry brings off the ball.

Curry's 3 pt threat could also be part of the reason Ellis more frequently gets inside with the ball (6.6 FGA inside 10 ft per 36 vs. 4.2 for Curry).

It's touch to assign the credit for the PPS differential:
Is it because Ellis is a better passer?
Is it because Curry is a better spacer?
Is it because Ellis is a better penetrator?
Some combination of all three?

Would be interested to hear Evan's thoughts.

[EvanZ]

In the case of Wright I think there are maybe two things going on, and you hit on one of them. Ellis driving and kicking back out to an open Wright. Wright basically can't hit a shot unless he's wide open.

I think there's another thing that is related but also has something to do with the pass. That is that often when Curry passes to Wright, he's simply swinging it around on the perimeter, so Wright is catching the ball more from his side, and then he has to rotate toward the basket. Obviously such a motion is not as ideal as catching the ball from someone who is closer to the hoop or even at the free throw line, because you have an easier shot if you catch the ball in front of you. The extra time it takes to swing around from catching the ball on the side might be those few extra hundreths of a millisecond that defenders need to bother your shot just a little bit more.

I'm not sure which of those is more important, but since I would argue they are related, maybe it doesn't matter.

I can also imagine that not all point guards hit players at the same angle and with as much accuracy. Guys like Nash and Paul put the ball right where it needs to be so shooters are basically ready to go. Guards that aren't as skilled don't hit the target as well, and the shooter needs to make more adjustments. It doesn't seem like much, but it likely is very important.

This was a great post from Sebastian Pruiti at NBAPlaybook that analyzed video of several PG's from last season's draft:

http://nbaplaybook.com/2011/06/23/makin ... int-guard/

Highly recommended reading/viewing.

[rrravenred]

Your next task is to plot passer and pass-receiver positions and relative velocities.

Drive and kick versus a skip pass or a lead pass on the break... the player's own offensive strengths and playmaking (rather than passing) skills come into play. You've obviously done a massive piece of work in getting those numbers, but it doesn't sound like it'd be a difficult thing for statisticians to do as a part of their normal duties to give us a sample size that gives more certainty on the issue.

[EvanZ]

Now they have some new video technologies too, that will help with that. Wish it was my full-time job.