Counterpart defense in ezPM (EvanZ, 2011)
Posted: Fri Apr 15, 2011 9:44 am
Author Message
EvanZ
Joined: 22 Nov 2010
Posts: 279
PostPosted: Mon Jan 31, 2011 4:18 pm Post subject: Counterpart defense in ezPM Reply with quote
I have implemented (some form of) counterpart defense into my ezPM metric.
http://thecity2.com/2011/01/31/ezpm-v-2 ... he-biggie/
If anyone has ideas for ways to improve or validate what I've already done, feel free to comment at the blog or here. For example, one of the big questions is whether and how to deal with counterpart assists.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
huevonkiller
Joined: 25 May 2010
Posts: 15
Location: Miami, Fl
PostPosted: Tue Feb 01, 2011 7:43 am Post subject: Reply with quote
I really like the counterpart aspect, we need more of that in defensive analysis.
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 816
PostPosted: Tue Feb 01, 2011 2:08 pm Post subject: Reply with quote
Inclusion of opponent scoring (divided equally to all defenders actually on the court when it occurs, entirely based on counterpart or in-between) is good.
One thing it does is focus on the cost of shooting fouls over non-shooting fouls. By charging for opponent points from the foul line you get to the immediate real cost of specific fouls rather than charging all fouls the same as past linear metrics without opponent scoring data have tended to.
Perhaps though non-shooting fouls deserve some charge for possibly or eventually putting teams into the bonus. There might be a reasonable approach to give them a small charge for that consequence when it actually occurs or on average. What about spreading a third or half the average or actual cost of in the bonus non-shooting fouls out to the previous non-shooting fouls? Using one-half would simply be applying the split the difference philosophy used on some other tough calls. Using one-third does something similar but preserves some greater consequence to fouling after the bonus comes into effect. I probably like one-third a little better but it could also be 40/60.
Back to top
View user's profile Send private message
EvanZ
Joined: 22 Nov 2010
Posts: 279
PostPosted: Tue Feb 01, 2011 2:31 pm Post subject: Reply with quote
Good points Crow. I've thought about it myself before. Something like that could obviously be done, however, I would imagine that fouls before and after the bonus would tend to average out over the course of a season.
OTOH, I suppose starters benefit the most from fouling before the bonus, whereas bench players would be more likely to foul after the bonus. Duh, right?
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 816
PostPosted: Tue Feb 01, 2011 2:51 pm Post subject: Reply with quote
Yes the impact on starters and subs would tend to vary. Is it fair to only charge for actual opponent free throws while you are on the court? On the face of it, no; but probably better to recognize the issue than ignore it.
And better to plan your team in light of the different consequence. Maybe there is some advantage in having a less discerning fouler in the game early in the foul count for the bonus and the more discerning guy in there afterwards. Of course early foul trouble and loss of player minutes is a risk of some team consequence even if it does not hurt player ratings.
Splitting the cost of in the bonus free throws tries at least tries to be more even-handed.
Back to top
View user's profile Send private message
BobboFitos
Joined: 21 Feb 2009
Posts: 199
Location: Cambridge, MA
PostPosted: Tue Feb 01, 2011 5:39 pm Post subject: Reply with quote
Many defensive fouls are positive defensive results, though. Many stop a potential basket/fastbreak. Resetting to a possession from a deadball = slight advantage for the defense.
_________________
-Rob
Back to top
View user's profile Send private message AIM Address
EvanZ
Joined: 22 Nov 2010
Posts: 279
PostPosted: Tue Feb 01, 2011 5:56 pm Post subject: Reply with quote
That's a good point. Before the bonus, fouling may help. I'm only debiting players when it definitely hurts, and being agnostic about the rest.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 816
PostPosted: Tue Feb 01, 2011 6:09 pm Post subject: Reply with quote
1. Foul before a shot that stops a likely basket
In many common metrics the fouls would be the standard deduction.
In EZPM, currently, this foul would be no deduction.
2. Foul on a shot that stops a likely basket
In many metrics the fouls would be the standard deduction, no worse than the previous scenario. Whether the foul shots are made or not.
In EZPM, currently, it depends entirely on whether the foul shots are made or not.
3. No foul, shot made
In many metrics, no defensive accounting unless it is a datapoint in the entire team and season defensive roll-up.
In EZPM, defensive blame to somebody- one or all or in-between, however that finally shakes out.
4. No foul, shot missed
In many metrics, no discrete, targeted defensive accounting.
In EZPM, defensive credit to somebody.
In 3 of the 4 scenarios the accounting is different between these approaches. In scenario 2 the impact will be essentially the same only if the opponent hits both free throws.
An average of 1700 personal fouls per team for a season. Deduct applicable scenario 2 cases and you probably end up having 1250+ differences in each team's defensive accounting between most previous metrics and EZPM on this single issue.
Look at other statistical elements and you will find many thousands of other accounting differences per team, per season. It probably comes to several thousand differences in accounting for a player with big minutes.
Back to top
View user's profile Send private message
EvanZ
Joined: 22 Nov 2010
Posts: 279
PostPosted: Tue Feb 01, 2011 6:48 pm Post subject: Reply with quote
Crow, if I wanted it to be the same, I wouldn't have bothered! Wink
But yeah, those are all points of distinction that people should consider.
_________________
http://www.thecity2.com
EvanZ
Joined: 22 Nov 2010
Posts: 279
PostPosted: Mon Jan 31, 2011 4:18 pm Post subject: Counterpart defense in ezPM Reply with quote
I have implemented (some form of) counterpart defense into my ezPM metric.
http://thecity2.com/2011/01/31/ezpm-v-2 ... he-biggie/
If anyone has ideas for ways to improve or validate what I've already done, feel free to comment at the blog or here. For example, one of the big questions is whether and how to deal with counterpart assists.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
huevonkiller
Joined: 25 May 2010
Posts: 15
Location: Miami, Fl
PostPosted: Tue Feb 01, 2011 7:43 am Post subject: Reply with quote
I really like the counterpart aspect, we need more of that in defensive analysis.
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 816
PostPosted: Tue Feb 01, 2011 2:08 pm Post subject: Reply with quote
Inclusion of opponent scoring (divided equally to all defenders actually on the court when it occurs, entirely based on counterpart or in-between) is good.
One thing it does is focus on the cost of shooting fouls over non-shooting fouls. By charging for opponent points from the foul line you get to the immediate real cost of specific fouls rather than charging all fouls the same as past linear metrics without opponent scoring data have tended to.
Perhaps though non-shooting fouls deserve some charge for possibly or eventually putting teams into the bonus. There might be a reasonable approach to give them a small charge for that consequence when it actually occurs or on average. What about spreading a third or half the average or actual cost of in the bonus non-shooting fouls out to the previous non-shooting fouls? Using one-half would simply be applying the split the difference philosophy used on some other tough calls. Using one-third does something similar but preserves some greater consequence to fouling after the bonus comes into effect. I probably like one-third a little better but it could also be 40/60.
Back to top
View user's profile Send private message
EvanZ
Joined: 22 Nov 2010
Posts: 279
PostPosted: Tue Feb 01, 2011 2:31 pm Post subject: Reply with quote
Good points Crow. I've thought about it myself before. Something like that could obviously be done, however, I would imagine that fouls before and after the bonus would tend to average out over the course of a season.
OTOH, I suppose starters benefit the most from fouling before the bonus, whereas bench players would be more likely to foul after the bonus. Duh, right?
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 816
PostPosted: Tue Feb 01, 2011 2:51 pm Post subject: Reply with quote
Yes the impact on starters and subs would tend to vary. Is it fair to only charge for actual opponent free throws while you are on the court? On the face of it, no; but probably better to recognize the issue than ignore it.
And better to plan your team in light of the different consequence. Maybe there is some advantage in having a less discerning fouler in the game early in the foul count for the bonus and the more discerning guy in there afterwards. Of course early foul trouble and loss of player minutes is a risk of some team consequence even if it does not hurt player ratings.
Splitting the cost of in the bonus free throws tries at least tries to be more even-handed.
Back to top
View user's profile Send private message
BobboFitos
Joined: 21 Feb 2009
Posts: 199
Location: Cambridge, MA
PostPosted: Tue Feb 01, 2011 5:39 pm Post subject: Reply with quote
Many defensive fouls are positive defensive results, though. Many stop a potential basket/fastbreak. Resetting to a possession from a deadball = slight advantage for the defense.
_________________
-Rob
Back to top
View user's profile Send private message AIM Address
EvanZ
Joined: 22 Nov 2010
Posts: 279
PostPosted: Tue Feb 01, 2011 5:56 pm Post subject: Reply with quote
That's a good point. Before the bonus, fouling may help. I'm only debiting players when it definitely hurts, and being agnostic about the rest.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 816
PostPosted: Tue Feb 01, 2011 6:09 pm Post subject: Reply with quote
1. Foul before a shot that stops a likely basket
In many common metrics the fouls would be the standard deduction.
In EZPM, currently, this foul would be no deduction.
2. Foul on a shot that stops a likely basket
In many metrics the fouls would be the standard deduction, no worse than the previous scenario. Whether the foul shots are made or not.
In EZPM, currently, it depends entirely on whether the foul shots are made or not.
3. No foul, shot made
In many metrics, no defensive accounting unless it is a datapoint in the entire team and season defensive roll-up.
In EZPM, defensive blame to somebody- one or all or in-between, however that finally shakes out.
4. No foul, shot missed
In many metrics, no discrete, targeted defensive accounting.
In EZPM, defensive credit to somebody.
In 3 of the 4 scenarios the accounting is different between these approaches. In scenario 2 the impact will be essentially the same only if the opponent hits both free throws.
An average of 1700 personal fouls per team for a season. Deduct applicable scenario 2 cases and you probably end up having 1250+ differences in each team's defensive accounting between most previous metrics and EZPM on this single issue.
Look at other statistical elements and you will find many thousands of other accounting differences per team, per season. It probably comes to several thousand differences in accounting for a player with big minutes.
Back to top
View user's profile Send private message
EvanZ
Joined: 22 Nov 2010
Posts: 279
PostPosted: Tue Feb 01, 2011 6:48 pm Post subject: Reply with quote
Crow, if I wanted it to be the same, I wouldn't have bothered! Wink
But yeah, those are all points of distinction that people should consider.
_________________
http://www.thecity2.com