Thursday, December 4, 2008

Tracking fast break points

A question came up on HoyaTalk about the perception that Georgetown is running (and scoring) more this season.

Fast break information is available in the play-by-play data, but not in a very satisfying form. Here's what I had to say about it in that thread:
I'm not a big fan of the way fast breaks are tracked by the play-by-play for a couple of reasons:
  • fast breaks are only noted on made baskets, so if a player misses a shot or turns the ball over on a fast break, it's not recorded as a failed fast break
  • more importantly, players who get fouled on a fast break (without making a basket) don't have their made FTs noted as resulting from a fast break, so fast break points from the play-by-play are generally lower than the actual points generated by the fast break in a game

I also promised to have some more info about fast break points later this week. I guess it's getting to be later in the week, so I've tried to come up with something. Any feedback on this would be greatly appreciated.

At the bottom of each HD box score, I'll add two additional lines that I hope will give some additional information about how each team performed on offense.

As an example, here would be the revised box footer for the Georgetown/Drexel game:
Efficiency: Georgetown 1.141, DREXEL 0.736
eFG%: Georgetown 0.585, DREXEL 0.311
Substitutions: Georgetown 29, DREXEL 28

2-pt Shot Selection:
Dunks: Georgetown 3-3, DREXEL 0-2
Layups/Tips: Georgetown 16-20, DREXEL 3-11
Jumpers: Georgetown 3-9, DREXEL 3-22

Fast break pts: Georgetown 16 (0.258), DREXEL 5 (0.152)
Seconds per off. poss: Georgetown 17.8, DREXEL 15.6

The second-from-last line is labeled rather unambiguously, but the numbers will need further explanation
  • The first number is indeed points scored on plays indicated as fast breaks in the play-by-play. However, as I noted above, the play-by-play doesn't track points scored on FTs scored when a player is fouled in transition.
  • The second number is my feeble attempt to deal with this. This is the ratio of fast break points divided by total field goal points (i.e. total points - FTM). Here, I assume that the team will score roughly the same percentage of points shooting FTs from transition as from half-court offense. So, the value in parenthesis is a very approximate estimate of the percentage of points scored off of fast breaks in the game. Very approximate.
The final line is much simpler. It's merely the amount of time each team spent on offense, divided by the number of possessions each team had on offense. Teams that score a lot on transition will have shorter time/poss than teams that run a deliberate offense. However, there is a confounding effect since turnovers will also shorten possessions. [The work-around would be to report time per scoring possession, but that's not something I currently track, so this will have to suffice for now.] I'm not sure exactly what we could learn from this stat, but it's easy enough to generate; I'll start spitting it out along with the fast break tracker.


Finally, I thought I'd take a look at fast break points for the first five games from last and this season, to see if the team really is getting out more in transition, and how much so:
2007                                   2008
Opponent* Fast br pts (%) Opponent Fast br pts (%)
William & Mary 0 (0.000) Jacksonville 10 (0.213)
Michigan 0 (0.000) Drexel 16 (0.258)
@ Ball State 10 (0.204) Wichita State 16 (0.432)
Fairfield 4 (0.077) Tennessee 6 (0.102)
@ Alabama 5 (0.094) Maryland 4 (0.068)
Total 19 (0.058) Total 52 (0.143)
*PBP not available for the game at ODU

And there it is - five games into the season, the Hoyas are scoring about 2.5 times as many points on transition as they did last year. Or, another way to look at this: last season, Georgetown outscored their opponent in transition twice in the first 5 games, while this year the Hoyas have done it 4 out of 5 games.

Edited: Numbers now corrected after finding bug in fast break retrieval (see comments).

4 comments:

  1. Interesting that its really just two games.

    I wonder about the 4 points in the Tennessee game. Seemed like more.

    ReplyDelete
  2. I just went back and checked the Tennessee game by hand, and it looks like there were 6 points on fast breaks for G'town, so I've got a bug of some sort.

    I'll post a corrected / updated post tomorrow.

    ReplyDelete
  3. Outstanding work as always, gentlemen. This should be an interesting stat to track this season - especially compared with last year's numbers.

    ReplyDelete
  4. (above comment mine)

    ReplyDelete