Will any boxscore metric add / improve their shot defense?

[Crow]

PER around 20 years old, never had any shot defense and apparently never will. Winshares, Wins Produced, BPM have team shot defense undifferentiated by player or even when player is on or off the court. EWins only has shot defense implicit in opponent scoring. DRE has no shot defense directly, just shot defense value slapped onto other defensive stats invisibly because that value has to find a home. EZPM has by position assumed counterpart defense from play by play data but help defense is either subjectively included or left out. To my knowledge, other lesser known boxscore metrics either have no shot defense or perhaps include it as poorly as the box score metrics discussed above.

PT-PM and Dredge have shot defense in the RPM component individually rated using play by play data and through some of the added player tracking data (but not 3pt defense to my knowledge). RPM has shot defense and everything else completely considered.

Will any box score metric add / improve their shot defense coverage? I doubt it. But that continued refusal is worth noting imo. The play by play data is out there to do better, even if only to get to team shot defense when a player is actually on court, a really low level of accuracy to get to.

[bchaikin]

simulation has individual player shot defense, on a scale of -5.0% for excellent shot defenders to +5.0% for poor shot defenders (and anything in between by 1/2 percentages, -5.0%, -4.5%, -4.0%, -3.5%, etc). this is exclusive of shot blocking, which is handled separately...

see the thread "NBA-ABA careers ranked 1952-2016"...

[Crow]

Good.

Your simulation is a metric wrapped in a method and anonymity for last x years; but points for including shot defense and improving it with new data.

[nbacouchside]

simulation has individual player shot defense, on a scale of -5.0% for excellent shot defenders to +5.0% for poor shot defenders (and anything in between by 1/2 percentages, -5.0%, -4.5%, -4.0%, -3.5%, etc). this is exclusive of shot blocking, which is handled separately...

see the thread "NBA-ABA careers ranked 1952-2016"...

bob how do you choose the shot defense factors you include in the simulations? are you using reputation, eye-test, tracking it yourself or?

[bchaikin]

bob how do you choose the shot defense factors you include in the simulations? are you using reputation, eye-test, tracking it yourself or?

yes, yes, and yes...

reference the above thread where i talk a bit about it. watch 5-6 hours of games a night during the season, where i am primarily looking at individual player defense, and who is guarding who - who are a team's primary perimeter defenders, who is always guarding the opponent's #1 or #2 option on offense, who isn't, etc...

also Synergy, SportVu, etc. are very helpful, but even in these you will find defenders who rarely guard very good offensive players who have "good" defensive numbers, and good defenders who always guard the #1 or #2 option for the opposition but whose defensive stats do not appear good, and so you have to measure that versus what you actually see...

there is now a decade+ of Synergy data, so for example dwight howard's PPP allowed is annually below 0.800, jose calderon's between 0.900 and 1.000, and SportVu not only lists FG%s allowed versus 2s and 3s but also shows a defender's FG% allowed differential between the FG% they allow and the average FG% of the players they guard...

lastly when the simulation is updated and beta tested to get team stats like off/def pts/poss and team eFG% shot and allowed to match the real life numbers, adjustments can be made (as that is the only subjective parameter in the software)...