Does a Points * TS% stat exist?

[DiscoLives4ever]

As the title says, is there a stat that multiplies points scored by true shooting percentage? I'm not all that well-versed in advanced metrics, but after Deron's big night scoring yesterday at a high efficiency it got me curious. It seems like this would be an easy method of combining volume scoring and efficiency into one stat. Any thoughts?

For example in this game it was 57 points multiplied by .745 = 42.47 Efficiency Points (yes, I just made that up). How would I go about comparing this to other performances?

[mysticbb]

Such a stat does not exist, but not quite sure how useful that is in that sense. Maybe comparing the points scored above average efficiency. The average TS% is 0.521 for this season, that would make 39.9 points expected, while Williams scored 57.

Well, you could use GameScore to compare the games: http://www.basketball-reference.com/fri ... eaders.cgi

[Rerisen]

The problem with game score and offensive rating is they are throwing all other contributions beside scoring in, and it does seem like the OP has found a legitimate gap here for a stat we don't have, where we want to just look at scoring impact in a single number, neither by volume or efficiency apart.

For instance Monta Ellis 22 PPG at .511 TS% vs David Lee his teammate, 18.9 PPG at .550 TS%.

What single number can suggest to us who is helping the offense more purely on scoring, or by how much.

Might call for something like removing the shooting player's own efficiency out of the offense, then taking the remaining team offense (or TS%) and seeing how many expected points they would produce if they shot the same amount of shots as the player. Then adding or subtracting the difference from their actual points.

Similar to Mystibb's idea, but I would use the team efficiency (without the player in question) in order to get at that not every team has the same capabilities to produce offense outside of a player.

I.e. Russell Westbrook was criticized for efficiency far more than Rose, even the years both were below average, due to that there seemed so obviously better options behind him on his team to get points.

So if I run this idea for GS, get something like 20.6 'true value points' will call it, for Monta Ellis per game, and 19.4 for David Lee. It would still favor volume pretty highly.

[nikkoewan]

Hello, first time posting in the Stat Board. Happy to see there are other people like me.

Anyway, I think evaluating players on a pure "scoring" impact is a flawed approach because it assumes players exist in a vacuum. They don't. Unlike baseball (where there are only 3 elements in the game - the pitcher, the batter and the baseball to form a very isolated event), basketball is a team sport. There are 11 elements in the game which are not independent from each other. Therefore analyzing a player's impact based only on his scoring (which this stat seems to imply), is flawed.

However, if we go by what the OP wants (i.e. effect of a players ability to score), you can simply remove the players points and possessions used from the total game, and then calculate ORTG for team. OR you can simply divide the points he contributed and divide it by the total team possessions. That way, you know how many percent of the team's total possessions did the player contribute a score.

[wigglestrue]

Hello, first time posting in the Stat Board. Happy to see there are other people like me.

Anyway, I think evaluating players on a pure "scoring" impact is a flawed approach because it assumes players exist in a vacuum. They don't. Unlike baseball (where there are only 3 elements in the game - the pitcher, the batter and the baseball to form a very isolated event), basketball is a team sport. There are 11 elements in the game which are not independent from each other. Therefore analyzing a player's impact based only on his scoring (which this stat seems to imply), is flawed.

However, if we go by what the OP wants (i.e. effect of a players ability to score), you can simply remove the players points and possessions used from the total game, and then calculate ORTG for team. OR you can simply divide the points he contributed and divide it by the total team possessions. That way, you know how many percent of the team's total possessions did the player contribute a score.

Uhhhh, there are more than three elements. There are fielders. There are teammates surrounding each hitter in the lineup. There is the stadium environment. Baseball is relatively isolated, but way more complex than you gave it credit for. Naturally, basketball is going to be way more complex than you gave it credit for, too. Just 11, huh?

Has the OP's stat been calculated yet? I think it's worth looking at.

[perun]

i did something similar with pts * fg%.

i guess ts% is more accurate.

it would be a good stat for ppl who just look at shooting % n dont look at actual production.

im sure then ppl will see that scoring 30 on lower % > than scoring 20 on higher %

[Alfred]

This is actually a pretty interesting idea for a stat.

[Paydro70]

I'm not sure what this would really tell you... you'd have to square the percentage or something to get a meaningful difference.

[Doctor MJ]

This is related to what I've mentioned a few times as a "shooting profit" stat, which in ****metrics terms relates to VORP (value over replacement). Kevin Pelton is the one I've seen do stuff in this ballpark, but nothing's really caught on, so I don't know where things stand.

Let me break down the stat as I've proposed it and you'll see why you need subtle changes to what the OP suggests in order for it to really be useful.

Profit = Margin X Volume, where margin is how much money you make on every item sold, and volume is the amount sold

Volume = is the true shots taken NOT the points scored. Understand that the fundamental thing here is that a player is shooting a certain number of shots, and that the points themselves are a product of the shot taking and the margin.

Margin = Player's TS% - Replacement TS%. In other words, if you can get just as good a shot by replacing the player with someone else, then it doesn't matter how many shots he's taking, he's not adding anything.

Got it? So in principle, this should be the baseline for deciding how much direct scoring impact a player is having, and of course the Margin itself gives an assessment of ROI (return on investment).

But, there are some issues;

1) How do we decide what Replacement is? In baseball, it's not so hard. For the most part a player will do about the same in fundamental hitting (or pitching) stats no matter the context (once you normalize for known factors), so a replacement hitter is the guy who's just below the folks who are already getting play.

However, basketball is a much more complicated game where different players take on very different roles by definition, individuals change roles with context, and as roles change shooting efficiency (TS%). Hence, to even begin to do this with enough confidence to have major hubris, you have to not only factor in a player's volume & efficiency, but also their role in order to properly gauge the Margin.

Pretty rough.

And of course, beyond that, there's the matter that the presence of one scorer affects other scorers. Typically we think of this meaning that a volume scorer acts to some degree as a Loss Leader allowing other guys more space which gives them open shots and bumps up their efficiency, but there's no way to standardize this, and there's clear evidence that sometimes the stagnancy of the offense when a volume scorer is emphasized can make the other scorers shoot worse (Dantley, Wilt).

Ouch.

Still it'd be fun to see historical data on this. Most simplistically, you could assume league average TS% as replacement TS%. This will overrate efficiency (and underrate volume), but knowing that going in you could do some mental adjustments. Taking it a step further, you could set replacement TS% as a particular sub-50 percentile. So if average 53% is league average, maybe you pick the 20th percentile as replacement and it turns out to be 50%. Taking it one step further, you could set different replacement levels of efficiency at different volume.

Okay, I'll stop. lol. Anyway, I think this is a great thing for people to think about, and I look forward to seeing more stuff along these lines, but it's not an easy thing to do while keeping the concepts involved grounded in something concrete.