Re: Defensive stats and stats to track
--- In APBR_analysis@y..., "McKibbin, Stuart" <smckibbi@c...> wrote:
Very good choices of projects. Every single one of them is on my
list of things I'd like to get. Some comments follow, based on my
> I'm in California and watch the Laker games. During the coming year
> hoping to have the discipline necessary to close what I feel are
gaps in the
> existing stats. I want to keep track of the following:
> 1. The types of offensive rebs and by who---and whether they result
> tipins, putbacks or if the team eventually scores in the possession.
I did find some indirect but convincing evidence once that offensive
rebounds do in fact increase a team's likelihood to score.
Specifically, if you calculate a Play % (which is what I call a
team's chance to score ignoring offensive rebounds), it is higher
after an OR than before. You would be collecting more direct
> 2. Who shoots "and-1" freethrows, technical foul freethrows and
I have gone through a series of fudge factors on getting free throws
right. Your analysis would help here, I think. Actually a year ago,
I worked out the formulas to a sufficient degree that I thought I had
it all right. Your results could open that up again. Not as
important for me.
> 3. Who recovers blocked shots (offensive, defensive or team rebs)
and if the
> defense recovers whether it results in transition buckets and by who
Offensive team recovers a higher percentage of blocked shots than
regular missed shots. I've measured it offhand a couple times. It
makes sense, too, with guys blocking things out of bounds. See next
thing wrt whether transition buckets happen.
> 4. If steals result in transition buckets and by who
I've looked a few times for whether steals or blocks lead to a better
Play %. I've never seen it. Strange. I'll be curious if your work
shows it. I have seen indirect evidence for it with college teams
since there is a greater variation in skill. I remember that
Pitino's Kentucky teams and the Iverson G'town teams showing
statistical correlations that suggested that transition buckets were
important. But I never did the measurements.
> 5. Offensive fouls (a special kind of turnover)
General breakdown of fouls is one thing I've gone for with the
defensive stats I'm collecting. Basically, I only count those fouls
that lead to FT's. A full breakdown is nice.
> My reasoning for tracking the steals and blocks is to see if it
might not be
> better to track these defensive stats as really offensive transition
> opportunities. For some reason I think that a Tendex-type
(My general editorial comment that there is no real way to fix Tendex
because no one knows what it is supposed to measure.)
> be improved if we convert these defensive stats to an offensive
> (treating a blocked shot as sort of like an assist), so we're
> apples to apples. I hypothesize that there are three components to
> efficiency: transition, half-court, offensive rebounding.
Probably a good breakdown. I find that people evaluate defense
mostly based on half-court offense. I also find that offensive
rebounds are due to defensive breakdown and lead to points easier (as
echoed above). Transition defense/offense is often tough to define,
but if you define it as the play that comes after a steal, maybe it
> contribute in one area, some in all three. Maybe something will
come of this
> approach, maybe not.
> Anyway I'm going back through the Finals and the Laker-Spurs series
> on these stats. Some of this stuff may be readily available
> fouls?) and I don't need to track it. If so, can someone fill me in
> ease my burden?
I don't know of specific references for this. With my focus on
defense these days, too, I don't know if I can add too much with my
stat-taking this year.
Journal of Basketball Studies