Charting our Defense: Western Michigan

[tommy]

So this is the 3rd game I've charted (4th actually - just haven't posted UNCG yet) in an attempt to quantify our defensive performance. I'm continuing to try to refine my techniques and provide more detailed and accurate information. Obviously, I have a long way to go, and I know my methods are open to all sorts of questioning. If you have any suggestions for me going forward, I'm all ears. But a lot of guys on the board found the previous analyses of the OSU and Washington games helpful and/or interesting, so I thought I'd continue with Western Michigan.

I fully acknowledge the limitations of this type of charting. That being said, I did watch this game and re-watched many plays, forward and back, forward and back, to make sure I was getting down everything of consequence that happened. I'm not perfect, but I think I got just about all of it.

I know this kind of thing isn't for everyone, but it is for some, so I hope some of you find it useful.

Not all the numbers are going to match up with the boxscore. This is because sometimes I analyze a play differently than others. Or because credit or blame for a particular play belongs to more than one player (that's why there are halves and thirds and even sixths -- that's what you get when you add halves and thirds) and because some plays are most fairly attributable to "team" rather than any individual player(s). Or because I missed something. I don't think this happened much, but it might've happened here or there.

Also, I know that in a blowout game like this one was, numbers tend to be skewed, and sometimes the conclusions that can be drawn are less certain. And of course this is only one game, just a snapshot.

So the major change I've made here since last time is analyzing the game in terms of possessions. So for each possession, I'm marking down the following:

1. Which players were on the floor? If you're on the floor for a given possession, it counts for you, if not, not. Obviously.
2. Were you engaged in the outcome of the possession in my judgment? Shows general level of activity, but also perhaps how involved in the opponent's offense your man was.
3. Forcing a missed FG attempt, either a 2 or a 3. I tracked 3's separately, but lumped them together in the table below.
4. FG's allowed, again both 2 and 3 pointers.
5. Forced turnover. Many of these are shared. Also, turnovers include charges taken, but not blocked shots, as the latter are forced FG misses.
6. "Creating" a missed free throw.
7. "Creating" a made free throw.
8. General catch-all for good defensive play that doesn't fit into other categories. I call it deflection/peskiness/disruptiveness. DPD. Might be able to capture some of the "intangibles" that have interested many on these boards lately.
9. Ball denial, both on the wing and in the post. Good denial gets you a plus. Failure to deny when you could've/should've gets you a minus.
10. "SIF" My shorthand for staying in front. These are only counted when your man makes a definitive move to the hoop. Stay with him, you get a plus, lose him you get a minus.
11. Help. Good help gets you a plus; failure to help when you could've/should've gets you a minus.
12. Catch-all for defensive lapses not otherwise covered, sort of the flip side of #8. I just call this "got beaten - other."

Then below I did one additional analysis: on what % of plays that a guy was on the floor did the team get a stop vs. what % of plays that he was on the floor did we give up points? How did guys measure up against each other and compared to the team as a whole? I thought this might address a little bit the issue of "intangibles" as well, as if you're doing things to help the team make stops, even if they don't show up in other areas of the charting -- like how you move, your talk, being in the right spot, getting other guys in the right spot, leadership, etc., that might show up in the team's success defensively while you're on the floor. So that's the second chart below.

OK so the main chart is first.

On floor
Engaged
Forced miss
FG Allow
Turnover
FT miss
FT Allow
DPD
Deny +
Deny -
SIF +
SIF -
Help +
Help -
Beaten-other

Curry
50
3
1
0
1.5
0
0
2
0
1
1
1
1
1
0

Rivers
49
6
1
1.5
1
0
0
0
1
0
2
0
0
0
1

Dawkins
32
5
2.5
1
.33
0
2
2
0
0
0
1
1
0
0

Mason
33
16
7.5
4.5
1.833
0
0
0
0
0
0
0
4
1
0

Thornton
34
10
1.5
2.5
2
0
0
2
1
0
4
4
0
1
0

Kelly
40
12
5.5
2.5
1
0
0
0
2
0
0
2
2
1
1

Miles
51
21
10.5
4
2
1
3
3
2
0
0
2
8
1
1

Cook
50
9
3
4
1.833
0
0
2
0
0
6
3
0
0
0

Gbinije
36
6
2
1
0
0
2
0
1
0
0
1
0
0
0

Hairston
36
4
.5
2
.5
1
1
0
0
0
1
2
4
0
0

A couple of things that I see. First, our bigs were very good giving help when appropriate. Thornton stayed in front of his man when his man made a move to the hoop 50% of the time. Cook, 2/3 of the time. Much more in there to consider.

OK the second table is below, showing the % of possessions that each guy played that resulted in stops, and the % that resulted in scores by WMU.

stops
scores
% stops

Curry
29
17
63.0

Rivers
27
18
60.0

Dawkins
19
13
59.3

Mason
18
14
56.2

Thornton
21
10
67.7

Kelly
20
18
52.6

Miles
28
20
58.3

Cook
26
23
53.0

Gbinije
19
17
52.8

Hairston
22
13
62.9

Team
46
33
58.2

Here, Thornton clearly outshined Cook, and none of the bigs were very good. Hmm.

OK, have at it! And again, welcome to suggestions for future improvement on the methods.

[moonpie23]

this is an interesting way to group data...as i said, when you're at the game, the perception is different. One thing i noticed was that, whenever i THOUGHT we didn't do a very good job on defense, Coach K had an unhappy look on his face. Whenever i KNEW we didn't do well on defense, Coach K and the entire coaching staff had unpleasant looks on their faces coupled with tradmark gestures.

the perceived "quality" of the opponent is a factor as well. Had we been doing this to another top ten team, i'd be saying we demolished them...

[Bob Green]

Thanks for continuing to take the time and put forth the effort to chart and analyze the team's defense. Miles' Help+ number (8) quantifies what my eyes tell me: Miles Plumlee continues to be our best low post defender because he is best at rotating so he is where he needs to be to counter the opponents. I find it interesting Miles recorded the highest (3) number in the Deflection/Peskiness/Disruptions category. I would've expected that honor to go to Tyler "The Pest" Thornton.

Thanks again for the great work!

[Kedsy]

Nice, thought provoking stuff, Tommy. My eyes told me Tyler got beat a lot off the dribble, and that Quinn stayed in front of his man better, and your numbers confirm that. But I concluded based on my observation that Quinn played as well or better defensively than Tyler did, and your stop percentage numbers seem to deny that emphatically, as Tyler was by far the best on the team in that category. Perhaps you have captured some defensive intangibles here. Although it's also possible the defensive stop percentage is similar to plus/minus in that there are too many other variables to make it reliable, especially in a one game situation. It would be interesting to track this percentage going forward, perhaps as it relates to overall team percentage to take the opponent into account. Might also be interesting to compare your stop percentage to plus/minus to see if they correlate.

I do have one question for you: how is "creating" a missed free throw different from creating a made free throw? Doesn't seem like the defensive player would do anything different for one vs. the other.

Also, seems like it was an off-game defensively for Michael, especially if the stop percentage has validity, but also in the first table. Other than his very athletic blocked shot, I didn't notice him much on defense, and maybe your numbers tell me why. One would think a potential defensive stopper would be more active, although on the other hand perhaps his "engaged" number was down because his off-ball defensive was good and he prevented his man from getting the ball? Or is that covered in "ball denial"?

[MChambers]

I do have one question for you: how is "creating" a missed free throw different from creating a made free throw? Doesn't seem like the defensive player would do anything different for one vs. the other.

In most cases, sure would be. I suppose it could be different in the sense that you might have fouled a poor foul shooter (e.g., Mason) as opposed to a strong one (e.g., Seth). Seems unlikely in most cases.

By the way, Duke's free throw defense has not been good this year (73.4%). We need to work on cracking our knuckles.

[loldevilz]

Thanks for continuing to take the time and put forth the effort to chart and analyze the team's defense. Miles' Help+ number (8) quantifies what my eyes tell me: Miles Plumlee continues to be our best low post defender because he is best at rotating so he is where he needs to be to counter the opponents. I find it interesting Miles recorded the highest (3) number in the Deflection/Peskiness/Disruptions category. I would've expected that honor to go to Tyler "The Pest" Thornton.

Thanks again for the great work!

I agree. Miles has quietly broken out in the last few games. 15 boards and 4 blocks in 25 minutes is absolutely ridiculous. He looked like Zoubek a bit getting the rebound on the defensive end and then the offensive end. The guy has a nose for the ball and is in the right place at the right time.

Just like Quinn Cook I expect him to eventually find his way into the starting lineup.

[Kedsy]

Just like Quinn Cook I expect him to eventually find his way into the starting lineup.

I agree Miles is playing great right now. But who do you expect him to start ahead of?

[tommy]

Nice, thought provoking stuff, Tommy. My eyes told me Tyler got beat a lot off the dribble, and that Quinn stayed in front of his man better, and your numbers confirm that. But I concluded based on my observation that Quinn played as well or better defensively than Tyler did, and your stop percentage numbers seem to deny that emphatically, as Tyler was by far the best on the team in that category. Perhaps you have captured some defensive intangibles here. Although it's also possible the defensive stop percentage is similar to plus/minus in that there are too many other variables to make it reliable, especially in a one game situation. It would be interesting to track this percentage going forward, perhaps as it relates to overall team percentage to take the opponent into account. Might also be interesting to compare your stop percentage to plus/minus to see if they correlate.

I do have one question for you: how is "creating" a missed free throw different from creating a made free throw? Doesn't seem like the defensive player would do anything different for one vs. the other.

Also, seems like it was an off-game defensively for Michael, especially if the stop percentage has validity, but also in the first table. Other than his very athletic blocked shot, I didn't notice him much on defense, and maybe your numbers tell me why. One would think a potential defensive stopper would be more active, although on the other hand perhaps his "engaged" number was down because his off-ball defensive was good and he prevented his man from getting the ball? Or is that covered in "ball denial"?

There are some other folks out there who do defensive charting and use complicated formulas to try to come up with defensive ratings, that sort of thing. They chart "creation" of a missed free throw vs. a made one, so I kinda piggybacked on that. I think the only way it can really be meaningful is to show if a guy is fouling good free throw shooters it may be indicative of "bad" fouls vs. fouling weak free throw shooters not harming the team nearly as much. But I acknowledge it's one of the least compelling aspects of the charting I'm doing.

I think I noted in my introductory comments that "engagements" can indicate some different things. In my mind, a high number of "engaged in the outcome of the possession" relative to total possessions a player is on the floor can indicate a high level of defensive activity; it can indicate that the other team forced the ball to your man a lot; it can indicate the offensive player worked hard to be involved in the offense, among other possibilities. A low number of "engageds" could mean the opposite: low level of defensive activity or intensity; the other team didn't look to get the ball to your man much; or your man didn't work very hard to get the ball or otherwise be involved in many plays.

As far as Mike G's "engaged" number being down, and that being indicative of good off-ball defense, it could be. But that's why I'm charting denials as well. If Mike's man was working hard to get the ball and Mike was denying him, I'd be marking that down as a denial for Mike. If, for instance, his man just kind of cut through now and then, made some perimeter passes, but didn't really get all that involved, then Mike wouldn't have much of an opportunity to raise his "engaged" numbers. And in this game, WMU forced the ball down low a lot. That's why our big guys' engaged numbers are high, and why our bigs forced both a lot of missed shot attempts and allowed a lot of makes. Most of WMU's offense ran through their bigs.