Science Does more team assists increase eFG%?

[Wizard Mentor]

*******Post Substantially Changed, Error Fixed********
One of the criticisms of the Blazers is that we didn't get enough assists last year, and that was a major reason for our decreased efficiency. The data from the 2017-18 season shows that assists have no effect on eFG%. More specifically, the coefficient of determination was 0.14 % (much less than 1%). The best line fit was virtually flat, and in fact had a slightly negative slope. (i.e. the more assists you had, the lower your eFG% was!).

BlazersEdge Article suggests we just need to improve our assist rate:
https://www.blazersedge.com/2018/8/...zers-assists-damian-lillard-cj-mccollum-stats


Notes:

• The previous graph expertly showed that there is no correlation between how many assists YOUR OPPONENTS get and YOUR eFG%. It sucks and it's embarrassing that I made the error, but I'd rather fix it and learn something than hide it and pretend it didn't happen.
• New graph fixes the error (hopefully). It also attempts to adjust for pace by using assists per 100 possessions.
• Data taken from https://www.basketball-reference.com/leagues/NBA_2018.html
• The Blazers were dead last in assists at 20.1/100 possessions. They were 23rd in eFG% at .511.
• The Coefficient of Determination is 38.9%, so pretty good for this type of data.
• The slope of the curve is positive, so on the whole, teams with more assists/100 possessions do indeed shoot better.
• It would be interesting to relate eFG% to number of passes, but I don't know where such data is.
• If one were interested, one could easily compare TS%, Ortg, and other stats of interest to Total Assists.

 

[e_blazer]

That's interesting. I think that intuition would expect that shooting percentage on assisted field goals would be higher than non-assisted shots. You would think that the assisted shot would hit a player who is open or in better scoring position and, therefore, the shooting percentage would be higher. It would be interesting to know if your analysis holds across the league or if the Blazers are an anomaly. It could be that Dame and CJ, being the Blazers best scorers, it's more efficient for them to take the shot than to pass to a guy who's less likely to make the bucket. They're also more likely to take and make a three point shot than the guys they would pass to.

 

[BonesJones]

One of the criticisms of the Blazers is that we didn't get enough assists last year, and that was a major reason for our decreased efficiency. The data from the 2017-18 season shows that assists have no effect on eFG%. More specifically, the coefficient of determination was 0.14 % (much less than 1%). The best line fit was virtually flat, and in fact had a slightly negative slope. (i.e. the more assists you had, the lower your eFG% was!).

BlazersEdge Article suggests we just need to improve our assist rate:
https://www.blazersedge.com/2018/8/...zers-assists-damian-lillard-cj-mccollum-stats

Notes:

• data taken from https://www.basketball-reference.com/leagues/NBA_2018.html
• It doesn't mean assists are unimportant, just that they do not immediately lead to higher eFG%.
• The Blazers offense is not as efficient as it could be, but blaming their troubles SOLELY on low assist numbers is wrong.
• It would be interesting to relate eFG% to number of passes, but I don't know where such data is.
• If one were interested, one could easily compare TS%, Ortg, and other stats of interest to Total Assists.

View attachment 22106

The problem with this study is that its comparing seperate situations. For some teams with great players, letting them go 1-on-1 is just as good as passing the ball. However, for teams like ours, passing the ball would greatly improve the offense. I've always said this type of offense struggles in playoffs, so I want to see this study done for the postseason.

Also, the graph is flawed. We're not that close to 29th.

 

[Wizard Mentor]

The problem with this study is that its comparing seperate situations. For some teams with great players, letting them go 1-on-1 is just as good as passing the ball. However, for teams like ours, passing the ball would greatly improve the offense. I've always said this type of offense struggles in playoffs, so I want to see this study done for the postseason.

Also, the graph is flawed. We're not that close to 29th.

Flawed how?

 

[BonesJones]

Flawed how?

We're not that close to 29th....

 

[Wizard Mentor]

The Blazers were 23rd in eFG% at .511
The Blazers were 29th in total assists at 1655.

This data is wrong, fixed in new graph.
Thanks, @BonesJones

 

[BonesJones]

The Blazers were 23rd in eFG% at .511
The Blazers were 29th in total assists at 1655.

You're wrong.

https://www.teamrankings.com/nba/stat/assists-per-game

 

[PtldPlatypus]

The Blazers were 23rd in eFG% at .511
The Blazers were 29th in total assists at 1655.

Ahh...you were looking at opponent total assists rather than team.

Also, since total assists is impacted but pace, I'd think a normalized figure (like assists per 100 possessions) would paint a more accurate picture.

 

[BonesJones]

You're wrong.
View attachment 22108
https://www.teamrankings.com/nba/stat/assists-per-game

The last 4 playoff teams in assists other than us:
- Russel Westbrick...
- ISO Harden...
- A Nate McMillan (ISO-Heavy)
- Utah Jazz (Defensive Minded)

 

[blue9]

I would expect a higher rate of assists to increase your FG%, but not necessarily your eFG%. Why? Fouls/free throws off of dribble-drives increase eFG% and never involve an assist.

 

[PtldPlatypus]

I would expect a higher rate of assists to increase your FG%, but not necessarily your eFG%. Why? Fouls/free throws off of dribble-drives increase eFG% and never involve an assist.

Incorrect--FTs increase TS%, not eFG. eFG is simply FG% factoring in the additional point off of 3's.

 

[Wizard Mentor]

Ahh...you were looking at opponent total assists rather than team.

Also, since total assists is impacted but pace, I'd think a normalized figure (like assists per 100 possessions) would paint a more accurate picture.

Yikes! I'll revisit this.

 

[Wizard Mentor]

Yikes! I'll revisit this.

Fixed now! See original post, or if you're too tired for that, here's the new graph:

 

[Wizard Mentor]

Ahh...you were looking at opponent total assists rather than team.

Also, since total assists is impacted but pace, I'd think a normalized figure (like assists per 100 possessions) would paint a more accurate picture.

Thanks for pointing this out, it helped me fix the graph and data.

 

[blue9]

Incorrect--FTs increase TS%, not eFG. eFG is simply FG% factoring in the additional point off of 3's.

Ah! Always get those two confused.

 

[Wizard Mentor]

Incorrect--FTs increase TS%, not eFG. eFG is simply FG% factoring in the additional point off of 3's.

Yeah, I wanted to keep FTs out of the analysis.

 

[PtldPlatypus]

Fixed now! See original post, or if you're too tired for that, here's the new graph:

Big shock, the two biggest above-the-line outliers are the teams with James Harden and Lebron James on them.

 

[oldmangrouch]

Chicken, meet egg. Maybe having a high eFG% increases assist totals, rather than the other way around? The greatest passer in the league can't help a brick-layer make shots.

 

[PtldPlatypus]

Chicken, meet egg. Maybe having a high eFG% increases assist totals, rather than the other way around? The greatest passer in the league can't help a brick-layer make shots.

Of course, the tired and true "correlation/causation" argument. However, this is a pretty clear case of causation. It is inarguable that better ball movement leads to more open shots. It is also inarguable that NBA players hit open shots at a higher rate than contested shots. And it is also statistically demonstrable that on the whole, NBA players shoot better off the catch than off the dribble.

This is like when people attempted to argue against the likelihood of steroid use in baseball by saying "Steroids can't turn some average Joe into Babe Ruth!", but the claim ignored the fact that it wasn't "average Joe's" who were 'roiding, but the top 1% of the 1%, who were simply taking already world-class talent to the next level. Similarly, we're not talking about great passers turning you or me into Kyle Korver; we're talking about players who are already world-class basketball players, who already demonstrate the ability to hit open jump shots at a high rate.

 

[riverman]

Ball movement leading to assists wear down defenses...GS is really good at it..the defenders have to work way harder ...5 player scoring 20 pts apiece usually beats 2 players scoring 50