Shot Chart and Play by Play Visualization

[Kathoro]

Kpa this is exactly what I was interested in. Can you make a data table that includes Shot Zones in one column, Total Misses in the next, Offensive Rebound %, Points Per Possession After Initial Offensive Rebound, and Opponent Points Per Possession After Initial Defensive Rebound?

This would be the greatest table ever.

[kpascual]

Kathoro, not sure what you mean by total misses and opp. PPP after defensive rebound. For the after defensive rebound stat, how will that relate to the shotzone? Is it the shotzone of the last shot before it got defensive rebounded? This particular analysis looks at the first shot in the possession, and it sounds like that's not what you're looking for.

Likewise, for total misses, do you mean total misses counted within a possession, or total non-scoring possessions?

Anyway, here's what I do have, assuming there was at least 1 offensive rebound within the possession.

Code: Select all

+------------------+-------------------+---------------------------+------------------+
| initial_shotzone | avg_pts_converted | possessions_without_score | possession_count |
+------------------+-------------------+---------------------------+------------------+
|                5 |            1.4040 |                       676 |             1718 |
|                1 |            1.2990 |                       754 |             1729 |
|               13 |            1.2315 |                      3209 |             8085 |
|               14 |            1.2066 |                      3738 |             9176 |
|               12 |            1.1506 |                       874 |             2012 |
|               11 |            1.1466 |                       836 |             1883 |
|               10 |            1.1115 |                      1200 |             2636 |
|                6 |            1.0870 |                      1247 |             2678 |
|                8 |            1.0869 |                      1130 |             2428 |
|                7 |            1.0852 |                      1650 |             3487 |
|                9 |            1.0843 |                      1687 |             3571 |
|                4 |            1.0026 |                      2052 |             3825 |
|                3 |            0.9927 |                      1095 |             1907 |
|                2 |            0.9882 |                      1935 |             3637 |
+------------------+-------------------+---------------------------+------------------+

[Kathoro]

Shot Zone Column:
In this column, all the shot zones

Offensive Rebound % Column:
Offensive rebound % corresponding to each shot zone

Points Per Possession Column:
Points Per Possession with at least one offensive rebound

Defensive Rebound % Column:
100-Offensive Rebound % for each shot zone

Immediate Opponent Points per Play:
Points per play for the defensive team after getting a rebound.

Forget the total misses column. The "Immediate Opponent Points Per Play" column would measure the points per play for the defensive team on the play immediately after getting the rebound. This would show which types of misses lead to the most points per play for the defensive team on the immediate play occuring after the rebound.

I hope that clears things up. For example:

Shot Zone: Off Reb %: Pts/Poss: Def Reb %: Opp Pts/Play:
Zone 1 21.5 1.02 78.5 1.25

Would this be possible for you to do?

[kpascual]

Is this what you're looking for? Here's the definitions I used:

Pts per possession - average # of points scored by a team whose initial shot occurs in that zone.
Off reb pct - percent of possessions that result in at least 1 rebound within the possession, whose initial shot occurs in that zone.
Def reb pct - percent of possessions that end in a defensive rebound, whose initial shot occurs in that zone. You could change this to "the zone in which the defensive rebounded shot got shot from", and you'd be maybe a 1 percentage point bump, but the data is essentially the same between the two.
Pts per possession after def reb - The average # of points scored by the OTHER team, after a defensive rebound, where the shotzone represents the location of the shot that got defensive rebounded.

Code: Select all

shotzone	pts_per_poss	off_reb_pct	def_reb_pct	pts_per_poss_after_def_reb
1	1.245	0.1845	0.5295	1.1179
2	1.221	0.1932	0.5442	1.0829
3	1.1022	0.2185	0.5813	1.0974
4	1.2438	0.1959	0.5358	1.0942
5	1.2776	0.1911	0.5143	1.0782
6	1.0058	0.1677	0.4979	1.0976
7	0.9912	0.1513	0.5072	1.0877
8	1.0012	0.1428	0.4993	1.0943
9	0.9911	0.1498	0.5055	1.0815
10	1.0202	0.1692	0.4915	1.0834
11	1.0812	0.1886	0.4777	1.112
12	1.0709	0.1901	0.4788	1.114
13	1.402	0.1782	0.3325	1.129
14	1.4536	0.1734	0.3114	1.139

I can sort of see your thinking behind the defensive rebound stat. But at first glance, the data doesn't seem to suggest there's a huge defensive consequence when you choose to shoot in a particular zone. I would guess that there are more important determinant factors to scoring than the location of the shot that got defensive rebounded.

[Kathoro]

Yes this is amazing. The only thing is I would like the offensive rebound % and defensive rebound % to be on missed field goal attempts only. The reason being is I use data on the offensive rebound % on misses for calculating my own player ratings. But this is great. I was wanting to see if misses from particular shot zones lead to increased efficiencies for the opposing team once a defensive rebound was secured.

One major problem I have, however, is that I don't know which zone is which. Is there a chart that you have that labels the locations corresponding to the different zones? I see the charts where the areas of the court are divided into 14 zones, which are each labeled with shot frequency numbers. But I don't know which zones they are according to your naming system.

Thanks for the table.

[Kathoro]

Nevermind, I found the shot zones on the 82games.com article.

[Kathoro]

I do think that these results show that shots in the high paint and low pain result in more fast break opportunities and thus more efficient scoring for the defense on the immediate play after the defensive rebound.

Of the top five zones in defensive points per immediate possession after rebound, four of them are the zones in the high and low paint. This effect would serve to partially offset the higher offensive rebound and put-back rates associated with field goal attempts in these areas.

[Mike G]

I wonder if this is caused by the fact that shots taken in the paint are more often taken by "bigs" -- who are then later getting back on D when the shot is missed.

[Kathoro]

I think the bigger effect might be from penetrating guards or forwards who are incapable of acting as "free safeties" on defense after missing a shot. A big who misses presumably still has his teammates in the back-court waiting to slow down any attempt at a quick fast break. In addition, many times a driving guard or forward will throw up a wild shot and even end up on the ground, causing them to get back even later than otherwise. Not to mention many of the guards or forwards who penetrate and miss a driving layup will often spend time complaining to the official instead of getting back on defense.

[Kathoro]

It would be interesting to break down the opponent points per possession after defensive rebound for zones 11-14 into the five different positions. Like PG, SG, SF, PF, C in one column and the opponent points per possession after defensive rebound for zone 11, zone 12, zone 13, and zone 14 in four other columns.

[DSMok1]

kpascual, this is some very useful data. Thanks for posting it! Can be applied to all sorts of things (directly calculated linear weights for events, for instance, like EZpm uses.)

[Crow]

kpascual, Could I interest you in preparing a team shot chart for Eastern Conference Finals series? I think Boston probably took a lot more mid-range shots than the Miami but would like to check it.

[kpascual]

Apologies for the late reply, I had been traveling for fun the past week or so.

Here are a couple PDF files around the Eastern Conference Finals. This first one shows the field goal percentage by zone:

http://vorped.com/bball/docs/shotchart_ecf_2012_fg.pdf

And this one shows the shot count by zone:

http://vorped.com/bball/docs/shotchart_ ... tcount.pdf

[Crow]

Thanks. Turns out Boston only took a few more mid-rangers than Miami and the percentages are close.

[Crow]

... but it is also true that the team with fewer mid-range shots won 5 of the games and had only 1 less in a tight game loss. So only in 1 game of 7 did a team win with more than a negligibly higher amount of mid-range shots attempted.