Saturday, March 9, 2013

Statistics Can Lie… Probably..

Just wanted to go alengthy this pillar, which I believe does a neatly position of describing why some scientific studies look to differiate from week to week (have some wine! no, don’t have wine! coffee is skilled for you!  caffeine is bad!).  And a fast notwithstandingt on how it applies to sportss below the leap.
With exercises, I feel the most importantly of this editorial is the part about generallyizable.  If you regard at any of the NBA all-in-1 statistics (PER, WP, RAPM, etc), they focus to present you a outline of what a athlete is ‘worth’ to a squad.  We force -ise that such a measure will commonize; we carry a sample (the preseason, the at first X tests of a season, 1 year out of a career) and assume/hope it applies to other situations (the regularly season, the other 82-X test matchs in the season, later year).  However, we don’t have a sample of the NBA; we only have what actually happens.  This means that when the circumstances shift for a ring or a runner, what we hold we realize could swing dramatically.  Take, for model, the year-to-year correlation chart near the finish of this column.  It shows that the correlation in Produced Wins for jocks who alter detachments is much lower than for athletes who stayed on the same corps.  Players who vary shifts are those who are most possible to amendment their place on a gang or their circumstances in widespread (perhaps they moved on because they were the injured or over/under-producing).  I haven’t seen a similar conspirech for other metrics, but I visualization similar results would near out.
You don’t have to transition crews to have such a transitional still.  Imagine that the Heat switched LeBron and Allen Ray’s duties.  LeBron would jog off screens and be more of a catch-and-fire guy while Allen would handle the ball a shipment, fly to the rim, and (try to) getoach an opponent-crushing engine of basketball destruction.  LeBron would certainly motionless be precious in his novel office; he’s a decently enough open fireer.  Allen could or might not be great in his fresh capacity; he used to be the focal show of an offense, but he was much younger then.  I reckon all of our intuitions would conclude that even if they were stationary productively runners with these the new responsibilities, both Allen and LeBron’s productivity would adjustment.  But we don’t really appreciate how much and it’s strenuous to talk what the impact on the Heat would be as a club, as it’s potential that Wade would pick up some accessory duties to aid Ray out.  That is, it would be tough to universalize.  And that’s with a pretty drastic conversion in side strategy; what if the Heat proper decided to sprint their offense a little differentlyly by adding more LeBron plank sets, or more Wade isolations?  We may surmise at the consequences, but they would certainly be conjecturees.
Another mode to express this is that I be under the im pression that… it’s significant to assume that a load of what we believe we bet about jocks is contextualize-depcloseent.  Related to that, a consignment of what we feel we guarantee about how a athlete would do in a unlike -ise is outrightasee assumption.  4 years ago, if you had asked me about J.J. Redick’s capability to span a line-up, I would have been doubtful; he’s basically a catch-and-launch guy that draws off the bench.  But the later 2 years appear to have demonstrated that he had (or learned) some of what it accepts to be more of a focal end: his minutes and starts are up, but most of his numbers have stayed steadily despite an uptick in usage.  He apparently isn’t downrightitely restricted on offense.  When race discuss about Chandler Tyson’s limitations, I class of assume that they’re claim and he might’t fire hopers.  But cannotbe his coaches and squadmates won’t allow him, because he’s so much superior at playing near the rim?  Or who’s to voice that if there had been the chance, he could’ve learned to open fire springers?  We don’t -ise, and we should’t recognize because the data is never in our ‘sample’ because It'sn’t really a sample.
I’ve used the NBA as an ideal, but the ideas really concern to all pes.  What we perceive about soccer runners presumably depconclusions on their rings and detachmentmates.  And the ideas treat to contrasting degrees of a unmarried p. e. as well; I wouldn’t assume that what we sense about the NBA applies to the WNBA or college tie outside of a few basic things.  Even then I would be hesitant; I would assume that launching threes is preferable than yearn twos at any scale, for role model, but I could be wrong.  There could be a league somewhere where accuracy drops dramatically enough at the 3 tip line that it wipes out the value of the supernumerary position.  And if such a league existed, you’d be foolish to treatment what you distinguish about it to the NBA.  You should alprocedures think the conditions under which your information are collected before you acquire in a rush to speak about how they impose to other situations.

No comments:

Post a Comment