The Effect of Pace on Average PPG of High Scoring Players

[Chicago76]

I dug this up from a few months back because I think some really important stuff is here. Further support for the pace is not a linear adjustment across eras argument:

Looking at players with a high scoring burden on their respective teams (expressed as % of team pts per min). 35% seems to be a good deliniation of the overall trend.

Shot clock era through final pre-ABA season:
4 35% team point seasons, 3 by Chamberlain and 1 from Baylor.

NBA during the ABA era:
1 35% team point season (McAdoo)

First 10 seasons post merger:
7 player seasons. 2 by Gervin. 1 each by Dantley, World B. Free, Freeman Williams, King, and Aguirre.

Trend continues as pace slows through the late 80s/90s.

Last 12 seasons since the strike shortened season:
29 player seasons. 5 by Iverson. 4 each by Shaq, Kobe, LBJ. 3 by Wade. 2 each by Anthony and McGrady. 1 each by Jermaine O'Neal, Stackhouse, K Malone, Durant, and Yao.

I could have looked at FGA and FTAs as a % of team components of both per minute, but it wouldn't make a difference as the relative efficiency of a big scorer vs. his teammates hasn't changed substantially. For example, low efficiency Baylor might get on the list again, but one of Wilt's seasons would likely fall off.

A few factors to control for:
-the number of teams obviously expanding in later eras exaggerates the difference substantially.
-salary cap probably dilutes talent a bit across teams, meaning better players less frequently play together. Still, duos like Kobe/Shaq and LBJ/Wade had huge team pt share/min. Not enough for both to make the 35%+ list as teammates the same year except 1 yr of Kobe/Shaq, but still at levels where both exceeded 32%, which is unheard of in prior eras.
-3 pt line and its impact on floor spacing and the need to more closely guard perimeter players, who dominate the list since the introduction of the 3 pt shot.

Even with the mental adjustments involved here, changes don't account for the concentation of touches.

I tried to figure out a way to normalize this for statistical player comparisons across eras in a more sophisticated way than I have presented above. I tried looking at measures of scoring touch concentration by teams given their pace across multiple years. I tried looking at the relationship between the same player across time accounting for age. It's all very messy and noisy with no satisfactory method, because it's too difficult to control for the teammate effect, the age effect, etc in order to isolate the pace effect.

A few observations from that more comprehensive study that would take too long to fully address here:
-the notion that player scoring touches can remain constant with substantial pace changes is flat out wrong. If pace drops 20%, the player can't increase his player touches by a factor of 1.2 relative to team to arrive at a net zero impact in scoring. Given small pace changes, it can occur, but this is probably just a function of the variation of player touches year to year, ie, a 3% pace reduction and a 3% increases in touches relative to team really just boils down to general year over year variance.

-the notion that pace and star player touches fall in direct proportin, ie, a 10% reduction in pace results in a 10% reduction in scoring output and attempts is equal is also wrong. If that was the case, we wouldn't see the dramatic increase in scoring share above. Throw out Chamberlain, who was a statistical and athletic freak the likes of which has never been seen again (with all due respect to Russell, Olajuwon, Shaq, and Robinson) and there were only 2 examples of a really high scoring load in over 20 seasons.

-not all parts of the pace curve appear to have the same impact, ie, the curve isn't linear. To illustrate, first note the relationship between pace and time per possession isn't linear. A decrease in pace from 125 to 115 increases the average length of possession by 1 second. A decrease in pace from 95 to 85 increases average length of possession by abt 1.8 seconds. Secondly, there are end of clock effects. Decreasing pace from 125 to 115 doesn't substantially lead to more frequent end of shot clock issues. Decreasing to 85 does. With the clock winding down, teams frequently get the ball to their best shot creator/scorer and get the hell out of the way. Playing at 130 to maybe 105 or 110 doesn't seem to produce a huge change. Slowing things down from a starting point of 105 or 110 seems to have a huge impact. That range seems to be the inflection point.

-the impact of pace adjustment probably isn't felt equally across all player types. A big who can run the floor will probably see his touch share increase in a fast paced environment. The open spacing derived from the tempo makes it easier to receive the ball than running set plays against an organized halfcourt defense. So Chamberlain's 36%, 38.5% and 40% touch share seasons probably are probably 34% to 38% seasons in a modern halfcourt game.

A spot up shooter probably declines a bit. Screens in halfcourt opportunities probably exceed equal transition spacing stop and pops. The end of clock ballhandler/scorer hybrids probably see their scoring touch share go down most substantially in a faster environment (Jordan, LBJ, Bryant, Wade, etc.). At first glance you might think the transition game would be a net benefit, but looking at pace preference of these teams suggests it doesn't work that way. Most of the teams featuring a player like this play at slow to average league pace over the last 20 years. These players are good enough to get the ball in the halfcourt and pick their spots to maximize value. A faster transition game requires passing and a more random distribution of the ball in transition. In a Jordan example, in a faster paced transition game, Jordan would likely lose touches to Pippen/Kukoc/Harper types at a substantial rate. He'd probably even lose touches to guys like Kerr and Paxson as trailers too. This is all relative to distribution rather than scoring touches in absolute terms.

Pace will still be the overriding factor, mind you. So going from a 100 poss/game environment to a 125 poss/game environment will lead to an increase in touches for just about everyone. It just won't be an increase of 25% for ballhandling scorers. It might be half to three quarters of that.

All of this makes adjusting for pace just about impossible. The linear route is probably no more wrong than any other method, but it's not particularly close to being correct either.

Just my 2 cents.

[Chicago76]

Just thinking about an easy way to display the discrepancy taking into consideration league size and the outliers of the really high scoring rate group first 20 years vs. last 12.

Pull out Chamberlain, leaving 2 35%+ seasons. 20 seasons at about 10 teams per season on avg = one 35%+ scoring load player per 100 team seasons.

Pull out top three from most recent era (vs one from earlier era to account for league size). That leaves 16 high scoring load seasons in roughly 360 team seasons, or roughly 1 in 22.5 seasons.

A fourfold increase can't be attributed to the three point shot and the salary cap alone. Doing the same type of calc for the first 10 years after the merger gives up a ratio of 1 per 60 team seasons...in an era when the cap didn't really have an effect (even at the tail end) and the 3 wasn't in high usage (and in certain years, wasn't even around). Some general inverse relationship between pace and the distribution of shots is the only logical explanation.

[C-izMe]

Back in the day players played around 45 minutes a game. You just showed that if anything scoring has gone up as pace went down. Maybe that is because they played a more team game back then and now every other team runs 13 iso plays with one person a game (don't have the exact number in front of me).

[Rerisen]

I've always found it hilarious how people use pace to say that guys like Monta Ellis are not great scorers. Please.

Indeed. That would be more his inefficiency and limp offensive on/off impact.

[azuresou1]

I've always found it hilarious how people use pace to say that guys like Monta Ellis are not great scorers. Please.

Indeed. That would be more his inefficiency and limp offensive on/off impact.

ITT: .536 TS% (.537 career) is horribly inefficient, and on/off measures scoring ability and not ability to play within an offense.

[Rerisen]

It's not 'horribly' inefficient, just inefficient. Among 21 players over 20 PPG last year, Ellis was 2nd to last (Bargnani) in efficiency. The year before, also 2nd to last. In 2009, dead last at 19.0 PPG or better.

It should go without saying that a 'great' scorer should help a team offense a great deal, and not just him own volume numbers.