Pages Navigation Menu

Is Controlling Pace Conducive To Winning?

US Presswire

US Presswire


 
You hear it often on pre-game broadcasts. A chyron will appear on the screen, stamped with the words “Keys to the Game”. And then the analyst will posit that in order to win, Team X must “control the tempo”. Or they might say “Slow it down” or “Run, run, run”. However this permutation of speed-related words is stated, this apparent “key” to the game refers to Pace.

Esteemed basketball writer John Hollinger developed Player Efficiency Rating over a decade ago, a metric littered with the nuts and bolts of the traditional box score. Its backbone though, was Pace Factor, which aptly estimates the number of possessions a team uses per game.

Because certain teams thrive on end-to-end transition sprints while others embrace the half-court duel, the resulting distribution of varied counting statistics tells a misleading story. By scaling them with Pace Factor, Hollinger provided a standard comparison method for fast and slow teams. Those efficiency stats are central to the analysis we see today.

For this writer’s debut piece though, I’m interested in pace as its own entity, and if controlling pace – fast or slow – can translate to victories, on a game-by-game level. After all, if a TV commentator says it repeatedly, it must be true!

Data

I used the friendly neighborhood Basketball-Reference to collect the pace number for every regular season game since the 1997-98 season. That’s 18,239 entries, up to this year’s All-Star break. Then, I grabbed each team’s pace from each season, as an estimate for “true team pace”. Thank you, Basketball-Reference.

Method

Each game has four pieces of information: the pace, that season’s teams’ “true paces”, and the winner. For example, the first game of the 2012-13 season, the Celtics at Heat, was played at a pace of 94.5 possessions (per 48 minutes). The Heat’s “true pace” is 90.7, while the Celtics’ is 91.4. Finally, Miami won, 120-107.

I’ll address my use of “true pace” first. Obviously, this isn’t a perfect figure (for reasons acknowledged at the end), but I feel comfortable using it as an estimate of how fast a team is comfortable playing in an ideal setting. The true pace difference between the two teams in my example is 0.7 possessions, which isn’t substantial. It translates to saying Boston has played 0.77% faster than Miami this season. That’s an extremely small margin and it’s difficult to say that either team wouldn’t be comfortable playing at the other’s “true pace” since they are so close together. To mitigate the inaccuracies of these similarities, I developed a strategy to throw out a majority of the games where the difference in pace was not significant.

I calculated each of these matchup differences in true pace for each game, from each season, and took its average, which came to 2.57%. I discarded every game whose true pace difference came below this (such as the above). Now, I have games with a distinct difference in true team pace. For example, coming just above the 2.57% “margin” is the Jazz 91.2 pace versus the Hornets’ 88.9. Their November 2nd game is hence, included. Of the entire data set, this method of filtering retains about 38% of the games.

Finally, how do I define if a team “controlled” the pace? The first of two ways is simpler: the game’s ultimate pace number resided above or below both team pace numbers. That Hornets’ game resulted in a pace of 84.8. Being below both team paces (Hornets’ 88.9, Jazz 91.2), I’ll deem the Hornets the controller of the game’s pace. Should the pace number lie between the two true team paces, who is it closer to? If it’s closer to Team A’s true pace, while still more than the 2.57% margin apart from Team B’s true pace, Team A is the controller. If it’s closer to Team A’s true pace, but still within 2.57% of Team B, the controlling team is indistinguishable, so the game is thrown out.

The first scenario accounts for about two thirds of eligible games, while the second accounts for a third.

With all of that said, let’s see how many times the pace-controlling team won.

Results

Screen shot 2013-02-20 at 8.56.07 AM

Our derived pace controllers won 50.7% of our 6,961 games. With no consistent trend emerging over the 16 season sample, the results imply that controlling the pace doesn’t necessarily result in winning more often. There also aren’t extreme fluctuations to either side, suggesting that minor randomness is the only culprit in the variations. The proximity to 50% is almost uncanny, as if it was the winning percentage of teams that won the opening tip.

Let’s take a glance at playoff games, using the same methodology.

Screen shot 2013-02-20 at 8.56.26 AM

We’re obviously plagued by small samples here, but the data doesn’t show any sign of agreeing with the hypothesis here either.

To rehash, this counts games where pace sided closer to one team. As I mention below, we can’t truly account for intent to control pace. A game may just wind up being slower or faster than usual, with neither team deliberately forcing their pace identity. By my definition, 38.2% of games resulted in the game pace shading closer to one team, which means more than 60% didn’t shade in either way, or featured two teams with similar paces.

This doesn’t conclude the study. My blanket approach here is fairly simple, and doesn’t drill down into identifiable subsets that may better depict true control of pace. I’d want to explore those areas before dispelling any theories. It’s possible the pace preaching of TV announcers is, in fact, true in a certain situations (though it’s looking bleak).  Because of randomness, winning might not be the best outcome to measure these pace numbers against, as offensive and defensive efficiency tend to paint a more accurate picture.

Next time, I’ll cut the data set into different subsets that might prove (or further disprove) the assertion.

Assumptions and caveats

  • By using team season pace, I’m using two constantly moving numbers as comparators. A team might speed up or slow down during the season, or make roster changes that affect its pace. I admit this proxy number may be problematic. Over a season however, I’d expect their pace number to stabilize and reasonably illustrate the team’s possessions per game.
  • The margin average, hence, is also shifting constantly. I did try some other methods – the standard deviation of the season paces, or even a fixed number – but they didn’t greatly affect the results of this study.
  • Even though I specifically assign who controlled the pace of the game, it doesn’t equate to that team actually doing so. It just means the game was played at a speed closer to their level – intentionally or unintentionally. The underlying fact is that pace isn’t one of Dean Oliver’s Four Factors, and teams are likely more concerned with executing plays rather than tempo.
%d bloggers like this: