Pace's Impact on Good/Bad teams

Home for all your discussion of basketball statistical analysis.
Post Reply
HockeyAssist
Posts: 2
Joined: Wed Apr 20, 2011 7:01 am

Pace's Impact on Good/Bad teams

Post by HockeyAssist » Wed Apr 20, 2011 7:11 am

I have been thinking about something and would appreciate any input. Let us assume a team has a point differential per possession that is constant. This is obviously a big assumption, but if this were the case, teams with positive point differentials would want to play as many possessions as possible, and teams with negative point differentials would want to minimize possessions played. Although it seems clear to me that certain teams will be more effective with certain paces, I am wondering if this idea is generally one that is accurate, that good teams should play as many possessions as possible within reason,( i.e. not going crazy taking bad shots in transition). Has anyone tested to see how a team's point differential is impacted by pace? I've thought about running a regression on the outcome of a game in terms of point differential, with pace as an independent variable. In theory, after controlling for the strength of one's opponent, it seems pace should be positively correlated with point differential for winning teams, and negatively correlated for losing teams. If this isn't the case, then it seems like a case of a team's efficiency being dependent on pace. Just curious if anyone has any thoughts on this issue.

EvanZ
Posts: 912
Joined: Thu Apr 14, 2011 10:41 pm
Location: The City
Contact:

Re: Pace's Impact on Good/Bad teams

Post by EvanZ » Wed Apr 20, 2011 2:48 pm

This is one of those topics that surely existed in the archives before recent events.

I took the past 5 seasons of team data and ran a simple linear regression. Here are the results:

Code: Select all

Call:
lm(formula = Pace ~ Diff, data = advanced_5seasons)

Residuals:
    Min      1Q  Median      3Q     Max 
-5.4634 -1.8392 -0.4418  1.5984  8.0004 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept) 94.58223    0.22147 427.071   <2e-16 ***
Diff        -0.10539    0.04259  -2.474   0.0145 *  
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 

Residual standard error: 2.712 on 148 degrees of freedom
Multiple R-squared: 0.03972,	Adjusted R-squared: 0.03323 
F-statistic: 6.121 on 1 and 148 DF,  p-value: 0.01449
While there is a statistically significant relationship (p=0.0145), the R^2 is very low (0.03) and it is inverse of the relationship you hypothesized. Namely, there is a (very, very weak) tendency for teams with greater differential to run at a slower pace.

Does anyone else remember previous discussions on the topic?

Crow
Posts: 6245
Joined: Thu Apr 14, 2011 11:10 pm

Re: Pace's Impact on Good/Bad teams

Post by Crow » Wed Apr 20, 2011 7:03 pm

bump

HockeyAssist
Posts: 2
Joined: Wed Apr 20, 2011 7:01 am

Re: Pace's Impact on Good/Bad teams

Post by HockeyAssist » Wed Apr 20, 2011 7:17 pm

Thanks for investigating Evan. I think I might run two regressions later, one with teams that have a positive point differential, and one with teams that have a negative. If my hypothesis is correct, pace will have a different effect on these teams, and putting both groups of teams into one regression may change our results. Does this make sense or is my logic fuzzy?

EvanZ
Posts: 912
Joined: Thu Apr 14, 2011 10:41 pm
Location: The City
Contact:

Re: Pace's Impact on Good/Bad teams

Post by EvanZ » Wed Apr 20, 2011 7:52 pm

Crow just re-posted a lot of the pace threads that were deleted. I would advise reading those. It's a bit more complicated than I thought. For example, I just ran another regression of pace on TS%, TOR (turnover rate), ORR, and DRR and find that they are all correlated with pace, especially DRR (this was observed in one of those previous threads, too):

Code: Select all

Call:
lm(formula = Pace ~ TS. + TOR + ORR + DRR, data = advanced_5seasons)

Residuals:
     Min       1Q   Median       3Q      Max 
-5.43373 -1.69447  0.04635  1.32915  6.21516 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) 125.10042   11.30759  11.063  < 2e-16 ***
TS.           0.27533    0.11986   2.297   0.0230 *  
TOR           0.47462    0.20626   2.301   0.0228 *  
ORR          -0.17809    0.08893  -2.003   0.0471 *  
DRR          -0.64204    0.10671  -6.016 1.39e-08 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 

Residual standard error: 2.393 on 145 degrees of freedom
Multiple R-squared: 0.2675,	Adjusted R-squared: 0.2473 
F-statistic: 13.24 on 4 and 145 DF,  p-value: 3.218e-09 
You can see there is a strong inverse correlation between pace and DRR:
Image

Jeff Fogle
Posts: 95
Joined: Fri Apr 15, 2011 2:05 am

Re: Pace's Impact on Good/Bad teams

Post by Jeff Fogle » Wed Apr 20, 2011 8:54 pm

Wanted to pop in to mention that I think there's an emphasis in the college game for teams with better athletes to force faster tempo. Studio and game analysts are talking more about this in the Dance. North Carolina uses that approach. Wisconsin supposedly doesn't have athletes, so they slow way down and try to beat you with execution. Far from universal of course...but I heard that mentioned a lot last month.

Or, in market terms, if you have an edge, increasing the number of possessions magnifies your edge.

In the NBA though, there's so much athleticism that it doesn't work quite the same way. You get the sense that many teams feel their edge is in the halfcourt game trade-off (we'll score more with our approach than our defense will allow)...so they want to deny fast break opportunities and force the halfcourt style. Teams who run and shoot treys (D'Antoni family of coaches, Golden State) are trying to make up for failures to defend in the halfcourt.

In the NBA, it's not so much that slowing down makes you better, but a slower pace is a characteristic of the better teams (or more experienced teams as you slow down with age). Speed is a sign of weakness.

Not sure if I explained that as well as I could. I think we generally hear "speed up to exploit your advantages" in the college game, and variations of "slow down to let your defense win the trade-off" in the pro game...

mtamada
Posts: 161
Joined: Thu Apr 14, 2011 11:35 pm

Re: Pace's Impact on Good/Bad teams

Post by mtamada » Thu Apr 21, 2011 4:22 am

HockeyAssist wrote:teams with positive point differentials would want to play as many possessions as possible, and teams with negative point differentials would want to minimize possessions played.
Solid intuition, I don't know what the empirical work has shown but the theoretical result is exactly what you say. E.g. taking a statistical approach, DeanO addressed this way back in his Journal of Basketball Studies website at MiningCompany.com or whatever it was called.
http://www.rawbw.com/~deano/

But even some TV announcers get it, with talk about "shortening the game" for example. I'd say it's a pretty well-known principle even without the statistical modelling. The main limitation is style of play. If you're Princeton in the NCAA tournament, then of course you're going to slow down the pace. Whereas if you're Houston's Phi Slamma Jamma with Olajuwon, Drexler, etc. you're going to want a fast pace. But those are the no-brainers. Where it gets hard, and the simple statistical models don't guide us, is what if you're Phi Slamma Jamma but you somehow are facing, not a college team, but an NBA team with talent that's superior to yours? Should you still play your fast-paced game, or should you slow down the pace as the statistics say -- but at the cost of taking yourself away from your accustomed style of play?

Crow
Posts: 6245
Joined: Thu Apr 14, 2011 11:10 pm

Re: Pace's Impact on Good/Bad teams

Post by Crow » Mon Apr 25, 2011 6:19 pm

bump

Post Reply