Saturday, December 22, 2012

± and beyond

I just had to give it a great title, to see what happens when a computer tries to alphabetize that character.

Right now, individual skater stats usually end with the ±, or plus minus.  That's number of points scored by that skater's team whilst she was on the track, less number of points scored by the other team whilst she was on the track.
  • Advantages
    • Easy to understand: a skater with a positive helped her team win, a skater with a zero just broke even.
    • Easy to calculate: for-against, could be done without computer assistance.
  • Disadvantages
    • Hard to calculate with: ± is centred on 0, and 0 is very difficult to factor in to future calculations.
    • Boundless: there is no minimum value or maximum value.  Theoretically, a blow-out win could be orchestrated by 1 jammer, who's ± would be something like +80.
In my opinion, those advantages make the ± a great thing for coaches and individual skaters, on a bout-to-bout basis.  However, it's not as easy for derby journalists, statisticians, and fans of the game.

For those groups, there is another option: the jamming/blocking average.  This is the percentage of points that were scored whilst the skater was on the track, that were scored for that skater's team.  Thus, if the jam score was 4 to 1, the lead jammer's jamming average would be .800 and the other jammer's would be .200.
  • Advantages
    • Easy to use for calculations: it's centred at .5, so a 0 is exceedingly rare and refers to 0 impact on the bout.
    • Bounded: all skaters will have a value between 0 and 1.  There is no possible greater or lesser value.
    • Easy to understand: we're used to percentages by now.  A jammer has .525, she just slightly won.  .725, and she's got good control.  Easy!
  • Disadvantage
    • Difficult to calculate: this one requires a computer to come up with.
I think it's a much better option for fans, journalists, and statisticians.  It gives you, at a glance, how much a jammer or blocker affected the scoreboard in her favour.  As well, it can be compared more easily between bouts and skaters.

There is, as well, a third option.  The first two are concerned with every point scored.  What if we instead looked at benchmarks?  (Thanks to The Mighty Bush for help on this one.)
  1. A jammer's job is 1st to get lead jam.  Thus, let's look at the percentage of jams where a jammer got lead.
  2. A lead jammer's job is to record a full pass, denying her opponent any points.  Thus, let's look at the percent of lead jams where the lead jammer scored at least 4 points more than her opponent.
  3. A non-lead jammer's job is to deny a full pass to her opponent.  Thus, let's look at the percent of non-lead jams where the lead jammer scored less than 4 net points.
  • Advantages
    • Easily to understand: this battery of three stats tell right away how successful a jammer is at those three aspects of the game.
    • Bounded and easy for calculations: centred at .5, bounded between 0 and 1, so easy to apply further.
  • Disadvantage
    • Difficult to calculate: this one definitely requires a computer to calculate all the "if...then..." portions.
This option is great for coaches and skaters, as it gives simple yet detailed feedback on three vital aspects of the game.  As well, it's great for journalists, statisticians, and fans as it breaks down a bout and can give a great "why" for a team's success/failure.

So, there are a few good options to measure jammers.  How to measure blockers?  Maybe later I'll argue for why I like to use the exact same metrics to measure their performance.  What do people think?

No comments:

Post a Comment