The Effect of Pace on Average PPG of High Scoring Players

[TrueLAfan]

I should preface this by saying I think pace is a garbage stat. I think it's poorly designed and functionally useless, and is rarely used in the way it could/should be.

That being said … I often see/hear/read about the numbers of good scorers in past “being affected by pace,” or “having better numbers because of pace.” And I thought … really? Because, if we take out the one true scoring outlier—Wilt Chamberlain in his monster scoring years—we've had three players that have scored over 35 ppg in the history of the NBA...Rick Barry in 1967, Michael Jordan in 1987, and Kobe Bryant in 2006. We've had a few more years the NBA's leading scorer has scored under 29 a game...Elvin Hayes in 1969, George Gervin in 1979, Alex English in 1983, Michael Jordan in 1998, Tracy McGrady in 2004, and (ta-da!) Kevin Durant this year, in 2011. So, do you know what those years have in common? The common thread in the years where players had big scoring years?

There isn't one. And it certainly isn't pace. Pace didn't change that much between Rick Barry (35.58 ppg in 1967) and Elvin Hayes (28.48 in 1969). Or between Gervin's 1978 and 1980 years...when his scoring led the league at pretty different levels. Kobe averaged over 35 a game only half a decade ago. The 2000s had slowest pace of any decade in the NBA. The league's leading scorer averaged over 30 a game in 9 out of 10 years … same as the 70s and 80s.

So I dug a bit deeper and did some grunt work. I took the average of the leading scorers in every season where have the pace statistic available, from 1974 forward. I increased the number of scorers used for the yearly average as the league increased in size—I used the top 4 from 1974 through 1976, the top 5 from 1977 through 1988, the top 6 from 1989 to 1996, and the top 7 from 1997 forward. I didn't count the strike year, because it's a statistical mess. And here's what I found

Over 27 seasons, the leading scorers averaged 27.45 points per game, with a league average pace of 97.31.

--When pace is higher than 100 (15 out of 37 available seasons), the best scorers averaged 28.07 points per game. The average pace in these seasons was 103.51.

--When pace is lower than 100 (22 out of 37 available seasons), the best scorers averaged 27.03 points per game. The average pace in these seasons was 93.07.

--When pace is really high, at over 105 (5 out of 37 available seasons), the best scorers averaged 27.09 points per game. The average pace in these seasons was 106.26.

--When pace is really low, at under 91 (7 out of 37 available seasons), the best scorers averaged 26.52 points per game. The average pace in these seasons was 90.51.

Pace may have some effect on the scoring of the league's major scorers. Maybe. Perhaps a small effect. It's nothing like as much as the pace differential. And, frankly, there's a lot of evidence to suggest it has little to no effect at all. Of the 37 seasons available, there are 10 seasons where scorers were above the scoring mark with pace average was lower than the norm, or below the scoring mark with pace average above the norm. that's close to 30% percent of all the seasons. That's an awful lot of negative correlation.

And even that doesn't make it clear enough .., the fact is that there are very, very few NBA seasons around the historical pace average. There are only six NBA seasons with pace between 93 and 99.5. The NBA is a league of extremes. But, as the pace/scoring numbers show for very high or low pace, that doesn't seem to impact how much major scorers average. When the league pace is super fast, top scorers score around 27 a game. When the league pace is slow as molasses...they score about 27 a game.

What does this mean? It seems to show that if you're a primary scorer, a go-to guy...the speed of the game seems to have relatively little effect on how much you score. And this is a kind of a refreshing concept. It gives basketball something like baseball has—a statistical bar. It's like being a .300 hitter or having an ERA under 3.00. If you score over 27 points per game, regardless of when you played, you are a great scorer. It doesn't mean all seasons are 100% equal—that's not what I'm saying. But I am saying it looks like that the pace statistic has no real application to recalculate or reassess the averages of really good scorers.

The painful spreadsheet can be found here: https://spreadsheets.google.com/spreadsheet/ccc?key=0AglwuTuVOEwwdDdibXZYSjhyYlhhclhnbEdwZjdmd0E&hl=en_US

[rrravenred]

Nice work, True.

But if the NBA is a land of extremes, shouldn't you also work in whether those extremes are reflected evenly across a league, or are unevenly distributed ACCORDING to the high scorers. If three sluggish teams with 18 ppg top scorers scorers are considered to have the same effect on team-wide pace as the top-5 scorers, all of whom are on lightning-quick teams, then that might throw a bit of a kink into the calculations, wouldn't t it?

[mysticbb]

Well, I think you had an idea in your mind and wanted to prove it, am I right?

Scoring per game for a team correlates with pace with a 0.86 correlation coefficient from 1984-85 to 2009-10. Thus in 26 years of NBA basketball we see a clear statistical connection between pace and scoring per game. Higher paced teams clearly scored more points per game.

Just saying. ;)

[Ryoga Hibiki]

At team level it's true for sure, but his point is that first options take already as many shot as they can and they're marginally affected by the increased number of possession.

Interesting, but I think it should be increased the pool size considering top3/top5 and so on, adjusting also to league size.

[ElGee]

I'm with Mystic here. While that's an interesting group of numbers, I still don't understand the hesitancy to look at pace or statistics on a per possession basis.

The whole idea of "pace" involves more "stuff" being accrued in the box score. There can't be a higher pace without more events occurring. Those events are primarily shots...leading to scores or rebounds.

Looking at your results, the only thing that doesn't support that idea is the "super high" pace period..also known as the 1970s. That was a period when teams and offenses were structured in a way that didn't promote a bevy of one-on-one play -- offenses were more balanced -- and more importantly, didn't have the 3-point line.

Otherwise, I don't see anything here suggesting that higher pace != more events, and the possession estimate is a solid way to calculate that...better than per minute or per game for the purpose of what people are usually looking for from a stat.

Heck, an individual player will get a few more possessions, in the exact same offense, if the opponent simply plays faster. That's relevant information, and it's not included in ppg or pts/36.

[TrueLAfan]

Well, I think you had an idea in your mind and wanted to prove it, am I right?

Scoring per game for a team correlates with pace with a 0.86 correlation coefficient from 1984-85 to 2009-10. Thus in 26 years of NBA basketball we see a clear statistical connection between pace and scoring per game. Higher paced teams clearly scored more points per game.

Just saying.

Actually, no. I don't have a dog in this hunt at all. I think the “pace” statistic is a poor one, but it's obvious that teams have played faster/had more possessions at different times in league history and that team scoring has changed as a result. (The easiest way to note this is to “watch basketball.” )

But the correlation between team scoring and pace in irrelevant here. The points/question is why primary individual scorers don't score more. It would really be much better (IMO) if we could say that there was a significant correlation between team/league pace and the amount scored by great scoring players. Somebody, obviously scores more when league/team pace changes. I have no problem with the idea that great players get more possessions when team scoring/number of possessions is/are higher. Yet, curiously, great scorers don't seem to … score many more points.

btw rrravenred, while the NBA is, indeed, a league of extremes, that doesn't seem relate to team pace as related to top scorers. For example, there was a very good scorer, a certain M. Jordan, whose teams were among the worst in pace in the league when he played. Like I said, mysticbb is actually precisely wrong in his assessment of my thinking. I'd really like to be able to see/say, “Look, top scoring players in the 70s scored a lot more because the game was so much faster.” (or for anhy reaosnm, for that matter.) And we're talking about true outliers here; about 3% of players that play 2000 minutes score 27 or more a game; that's less than 1% of the total players. And I actually do think there is some effect of the overall speed of play on scoring by top scorers. But I think it's much less than the difference indicated by speed or pace or “pace” or whatever you choose to use or call it.

[mysticbb]

I think the “pace” statistic is a poor one, but it's obvious that teams have played faster/had more possessions at different times in league history and that team scoring has changed as a result.

Well, that's basically what I asked. Your interpretation seems like just that you "want" to see the result as proof that pace doesn't matter much. But with the way you are doing it, I can't see that you are able to achieve a seperation between scoring efficiency, usage and pace.

Player's scoring is determined by the amount of opportunities they have and their efficiency. You need to seperate that from pace to get an idea for it. When I ran a regression on those things I came up with 50% efficiency and 50% pace as the reason for higher amount of scoring, when adjusted the points scored for ever since 1985. I did it on 20+ ppg scorers and the conclusion I had: the players are scoring more when they either have more opportunities or they are more efficient. Well, it is really that simple.

Higher pace gives more opportunities which results in more points, that's what the overall team analysis shows. In your example pace seems to have not that linear effect, but I suspect a lot of difference in scoring efficiency right there, means: the guy on the higher pace tend to score less efficient in the end (there are always exceptions, obviously).

[Nivek]

Any control done for efficiency? Overall efficiency was pretty low through the early years of basketball. It began to rise in the 70s and 80s, leveled off in the 90s, and dipped in the late 90s and early 00s. Nowadays it's back to the 80s/90s level.

It could just be that individual scorers didn't score more because they were less efficient overall.

It could just be that the number of scorers you chose to represent each year is too small. What happens if you look at top 20 scorers? Leading scorers for each team? Second leading scorers?

Consider -- in 73-74, the league averaged 105.7 points per game on 107.8 possessions per 48 minutes. That's an ortg of 97.7. This past season, the league averaged 99.6 points per game on 92.1 possessions per 48 minutes. An ortg of 107.3.

So, the 73-74 NBA had an additional 15.7 possessions per 48 minutes, which they used to generate only an additional 6.1 points per game.

Your finding may simply be showing a difference in efficiency between eras.

[Nivek]

Just took a quick look at the top 5 scorers for 8 seasons since 77-78 (the first year there was enough data collected to calculate Dean Oliver's offensive rating). Table below shows season, points per game, offensive rating and usage:

Code: Select all

SEAS    PTS     ORTG    USG
1977-78 26.1    108.4   28.0
1982-83 26.5    115.2   27.1
1987-88 30.3    117.6   31.2
1992-93 28.2    118.4   30.1
1997-98 25.9    111.0   31.1
2002-03 28.6    110.2   32.9
2008-09 27.4    116.0   32.4
2010-11 26.2    113.0   32.2


What this data suggests is that while individual scoring didn't change much as the pace of the game slowed, usage was higher among top 5 scorers, and efficiency among the top 5 was generally better. 97-98 managed a better average offensive rating than 77-78 despite the presence of Antoine Walker as the league's 5th leading scorer with an ortg of 98. Yeesh.

[rrravenred]

btw rrravenred, while the NBA is, indeed, a league of extremes, that doesn't seem relate to team pace as related to top scorers. For example, there was a very good scorer, a certain M. Jordan, whose teams were among the worst in pace in the league when he played. Like I said, mysticbb is actually precisely wrong in his assessment of my thinking. I'd really like to be able to see/say, “Look, top scoring players in the 70s scored a lot more because the game was so much faster.” (or for anhy reaosnm, for that matter.) And we're talking about true outliers here; about 3% of players that play 2000 minutes score 27 or more a game; that's less than 1% of the total players. And I actually do think there is some effect of the overall speed of play on scoring by top scorers. But I think it's much less than the difference indicated by speed or pace or “pace” or whatever you choose to use or call it.

FWIW, in a casual look at 87 (yes, Jordan's 37 ppg season) the three top scorers were Jordan and Nique (on bottom four pace teams) and Alex English (on the typically fast Nugget teams of the era). Another way to look at it (look into it if I have time) is in in terms of individual scorers and their individual share of Usage as a team's pace changes over time (controlling for minutes played). If the usage fluctuates with the individual's personal raw numbers remaining consistent then that might back up your (and, as it happens Jordan23Forever's) view that "star" players will get theirs, regardless of pace.

(FWIW, Jordan and Nique's usage was off the charts that year. English's, whilst high, wasn't quite in that ballpark)

Correction: Nique's usage wasn't huge in 87, being only 55th on the all-time list. I was getting it confused with the subsequent year, which was 11th.

[Nivek]

A better way to look at this might be to look at high-scoring players per game averages as a percentage of their team's points.

[DeeDub]

In comparing players across eras pace has some impact, but not as much as offensive efficiency/FG%. With the recent rebirth of the MJ-LeBron debate, I never seem to see anyone taking that into account. The level of defense in the early part of MJs career was nowhere near what it has been in LeBron's. In MJ's first season, the average NBA team scored 110.8 ppg on .491 FG%. In LeBron's first season, the average NBA team scored 93.4 ppg on .439 FG%. That 110.8 ppg that represented the average NBA offense/defense in MJ's first year would have been the best offense/worst defense in the NBA, by far, in LeBron's first year. That .491 average FG% was 2 percentage points better than the best FG% offense in LeBron's first year. Those are huge differences. And while scoring dropped a little over MJ's first few years and rose a little in LeBron's, there were still significant differences. The fact is it was much, much easier to score in the early years of Jordan's career than in the early years of LeBron's. The average starting NBA wing player (SG/SF) in Jordan's first season averaged 20.6 ppg on .495 FG%. This past season, only 8 (out of 56) starting wing players scored more than 20.6 ppg. Some of that was pace, but most of it had to do with the level of the defensive play.

The irony is that respected basketball people often not only fail to acknowledge those differences, they claim that it was actually harder to score during the Jordan era due to the eventual changes in the perimeter contact rules. But that is demonstrably untrue in light of the differences in how the average teams/players scored in the early part of Jordan's career. Similarly, comparing Jordan's numbers to someone like Oscar Robertson is similarly unsuited for a straight comparison. The league was very different back then. To make comparison across eras, I think the best way to do it is to compare the numbers to the average or replacement level player of each player's era.

[W Z]

Are you arguing that the best scorers' points per game is not affected by the number of possessions their team uses? Like, that the number of opportunities good scorers have to hoist up buckets has no effect on how many they make? I would think you would need some damn good evidence to back that up, much more than this very rudimentary analysis you've done.

It gives basketball something like baseball has—a statistical bar. It's like being a .300 hitter or having an ERA under 3.00. If you score over 27 points per game, regardless of when you played, you are a great scorer.

There is something like that, in fact something exactly parallel to that - FG%.

[richboy]

Are you arguing that the best scorers' points per game is not affected by the number of possessions their team uses? Like, that the number of opportunities good scorers have to hoist up buckets has no effect on how many they make? I would think you would need some damn good evidence to back that up, much more than this very rudimentary analysis you've done.

It gives basketball something like baseball has—a statistical bar. It's like being a .300 hitter or having an ERA under 3.00. If you score over 27 points per game, regardless of when you played, you are a great scorer.

There is something like that, in fact something exactly parallel to that - FG%.

Here some evidence. Watch a game

The conclusions to your findings is not a surprise because I've said this for years. I've said this to John Hollinger for years. Pace has little to do with scoring because as Pace goes down usage of the elite players go up.

This is why I say watch a game. The slower the game is the more likely on that possession a team will go to its best offensive player. Quick shots and quick possessions often benefit role players. The assumption that a stat like PER makes is an elite scorer like Kobe Bryant on a high pace Laker team would see his scoring numbers drop if he played on the slowest pace team in the league. They don't know Kobe Bryant or the mentality of elite scorers then. It thinks that basketball is played in a series of percentages and ratios. If you played a game and told Kobe there would only be 100 possessions. Kobe then would take 60 of them. If you told him there was 80 possessions he still taking 60 of them. Scoring numbers are not greatly effected by pace.PER thinks though that if Kobe uses 60% of the possessions when there is 100 possessions that if there are only 80 he still would only use 60% of the possessions. That 2008 Kobe would go from 29 a game on the Lakers to 23 a game if he played on the Blazers. Lebron and Wade played on much slower pace teams. Because of this pace assumption it showcases them as much better scorers in PER. Its the biggest flaw with pace adjusted stats. The Lakers pace has slowed over the years as they have become a bigger team. In the process his usage goes up.

[ElGee]

^^^ This is really sound logic and it's true. But the question is not whether pace has a linear effect on scoring, it's whether it effects scoring. It obviously does -- there is no way around it really.

If you have more opportunities to score, you have, well, more opportunities to score. The usage may very well increase in slower settings for No. 1 options, but adding extra possessions in transition still gives them extra opportunities. Adding extra possessions from foul bonuses still gives them extra free throws. Adding extra possessions from timeouts or quick opponent scores still gives them extra possessions.

None of this is perfect -- it's an estimation by definition -- but players have more scoring opportunities in 130 pos games than in 85 pos games, and the same can be said about a grinding 77 pos Finals game vs. a 105 pos GSW game.

[W Z]

If you played a game and told Kobe there would only be 100 possessions. Kobe then would take 60 of them. If you told him there was 80 possessions he still taking 60 of them. Scoring numbers are not greatly effected by pace.

Situation A:
Kobe's team plays at a tempo of 80 possessions/game, and he uses 60 possessions (75%).
The teams tempo speeds up to 100 possessions/game, and he uses 60 possessions (60%).

Situation B:
Kobe's team plays at a tempo of 80 possessions/game, and he uses 60 possessions (75%).
The teams tempo speeds up to 100 possessions/game, and he uses somewhere between 60 and 75 (60-75%) possessions (let's say, 67, just to throw it out there).

I understand the reasoning in you post, richboy, but, can you honestly tell me that you think situation A is more likely than situation B? I really just can't take that assertion seriously no matter how hard I try.

[Doctor MJ]

I want to make clear that I think TrueLA analysis is interesting and thought provoking.

With that said, the biggest scorers in history came during the fastest pace era despite weak efficiency. This is one hell of a coincidence.

Bottom line I think is that that if you are determined to get the ball to your star, this can be done rapidly. This what they did in the 60s. However, getting reliably good shots for your team is tougher than that, and so there is some relationship between taking your time and getting the best shot, which more often than random means your star shoots it.

However, let's understand that at this point the big scorers are typically perimeter guys with the ball in their hand early in the shot clock. The slow pace of today certainly doesn't exist solely to manufacture opportunities for ball dominant players. There are other factors involved in this situation, among them are 1) a general tendency to run the offense more through perimeter scoring than ever before, and 2) these ball dominant perimeter stars being the guys who take shots when the 24 second clock is winding down and the offense simply has to jack it up.

[tclg]

I think this is super interesting. I think that for the most part Guys are going to get theres. And the rules now favor perimeter scoring more than they did before. Its just very interesting statistical analysis

[penbeast0]

Richboy had a very interesting way to think of it.

For me, I figured it as part of the historical trend toward clearout/isolation scoring. Phil Jackson in Chicago was sort of a trendsetter with this, before that teams would expect all their starters and maybe a reserve to be 10ppg scorers with stars in the 20+ range. But the modern offense has been more focused on clearing space for the big stars to work their defenders in their comfort zones than spreading the wealth. More iso offense = greater percentage of scoring by the primary star.

Sort of ashamed to post this here though as I have not done any analysis at all to back it up

Oh and DeeDub, your efficiency argument has a huge flaw because it ignores the greatly increased use of the 3 point shot which means that fg% is less and less useful as an efficiency comparisom . . . use efg% which takes the 3 into acount or preferably true shooting percentage which also has an adjustment for foul draw although B-R.com has taken that out of the yearly league analysis unfortunately. (They have efg and foul draw but not the combined number)

[azuresou1]

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