Roundup 9/4/10

In basketball, don't call a timeout at the end of close games. You'll be more likely to draw a foul and win.

A salary cap also needs a salary floor.

The historical distribution of talent in the NFL.

A neuroscientist's struggle to understand regression to the mean.

The Sports Reference sites, including PFR, was named among Time's 50 best websites. Congrats, guys.

Tony Romo is not a mirage.

Advanced Public School Stats. What's next, Fantasy Teachers leagues? ...With his first round pick, Brian Burke selects Jaime Escalante, math teacher, Garfield High School... Stand and Deliver-great flick by the way. Escalante passed away earlier this year.

Strong finishers.

WPA was invented 40 years ago. Helmet knock: Tango.

Leveraging Wisdom of the Crowds for fantasy football. Basically, Drew Conway identifies the players with the most variant rankings in the various rankings. Be wary of Stafford, Lee Evans, Flacco, Breaston, Maclin, Burleson, Shiancoe, Tomlinson, and especially Braylon Edwards.

Planned sports venue development vs. spontaneous sports development.

Does calling a timeout to ice the kicker work? One study says yes:

Psychology professor Nadav Goldschmied reviewed data from six National Football League seasons (2002-08) and found that kickers who'd been iced scored only 66.4 percent of the time (73 out of 110 kicks). By comparison, kickers who were not iced had an 80.4 percent success rate (131 out of 161).

But I'm dubious. The article says the study accounts for experience, game location and score, but suspiciously says nothing about attempt distance. Wouldn't you think that the iced kicks tend to be the more desperate variety, taking place from longer distances? If anyone has access to the study, please forward it to me. (H/K: FO)

Speaking of dubious, Football Outsiders posted a list of its research findings the other day. Some are interesting, some are vapid, some are plain wrong. Among the highlights (my comments in italics):
  • Standard team rankings based on total yardage are inherently flawed. -- Ok
  • Running on third-and-short is more likely to convert than passing on third-and-short. -- Concur
  • If their overall yards per carry are equal, a running back who consistently gains yardage on every play is more valuable than a boom-and-bust running back who is frequently stuffed at the line but occasionally breaks a long highlight-worthy run. -- Not always. If you're the underdog or significantly behind in the game, you want the boom/bust guy.
  • A running back with 370 or more carries during the regular season will usually suffer either a major injury or a loss of effectiveness the following year, unless he is named Eric Dickerson. -- False. As one of my commenters put it, this notion "is a violent assault on good logic and reason."
  • Teams with more offensive penalties generally lose more games, but there is no correlation between defensive penalties and losses. -- Probably false. It's possible they've fallen for the same trap I fell into years ago. The NFL calls penalty yards by an opponent "defensive penalties." But if true, it's interesting. The real question would be why don't defensive penalties matter?
  • Field-goal percentage is almost entirely random from season to season, while kickoff distance is one of the most consistent statistics in football. -- True.
  • Field position is fluid. -- Huh? I think this is just supposed to mean that defenses affect their offenses' field position and vice-versa.
  • Injuries regress to the mean on the seasonal level, and teams that avoid injuries in a given season tend to win more games. -- Ok. Did we need a study for this?
  • A team will score more when playing a bad defense, and will give up more points when playing a good offense. -- And all this time I thought it was the other way around!
  • The future NFL success of quarterbacks chosen in the first two rounds of the draft can be projected with a high degree of accuracy by using just two statistics from college: games started and completion percentage. -- Intriguing, but I'd like to see how this system has performed since it was published. As I recall, it's been off the mark.
  • Championship teams are generally defined by their ability to dominate inferior opponents, not their ability to win close games. -- Interesting.


Who stands out furthest from the pack in their respective sports? Alan Francis does.

One of the questions I get often is 'where do you get your ideas?' John Cleese (of Monty Python fame) says, "We don't know where we get our ideas from. We do know that we do not get them from our laptops." He says the creative subconscious needs long uninterrupted periods.

A different way to look at RB performance.

"...What do you think accounts for that .03 difference?"

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

9 Responses to “Roundup 9/4/10”

  1. Tom says:

    Regression to the mean is such a simple, yet poorly understood thing.
    I have always liked to think about it as the result of limits, and nothing more, but that's the mathematician in me.

    The biggest error made with regression to the mean is to think that a high or low measurement is an indicator of a future sudden fall or rise in the measurement, which i simply the gambler's fallacy.

  2. Tom says:

    I didn't spot this:

    'Championship teams are generally defined by their ability to dominate inferior opponents, not their ability to win close games'

    To be honest that would seem an obvious statement. A game in which one team dominates another tends to be a game where the pre-game probability of one team winning was much greater than another, and as such the few elite teams which can be expected to have such wins more often than not will appear to be dominant in such games, yet when one elite team plays another it is much more of a coin toss as to the winner, so winning close games can be more down to luck, and so not a great indicator of a truly championship worthy team.

  3. Ian says:

    "A team will score more when playing a bad defense, and will give up more points when playing a good offense"

    That sounds more a corollary than a result of a study. If we define bad defenses as ones that give up more points (not contentious, right?), then the study is an exercise in the blindingly obvious.

  4. Adam Davis says:

    "Championship teams are generally defined by their ability to dominate inferior opponents, not their ability to win close games."

    Not only do I agree with this, but I laugh at the college football pundits who, several years ago, demanded that the BCS rankings remove any reference to margin-of-victory in their calculations. There was so much hand wringing about the supposedly evil incentive for coaches to run up the score, that they completely eliminated one of the most accurate indicators of a team's relative strength.

    IMHO, the real solution to the run-up-the-score incentive was simply to ensure that MOV considerations are done in tiers. In other words, if you beat a team by 50 points or by 57 points, you should really get the same bonus in whatever scoring system you are using, because there really is no huge distinction between a 57-point win and a 50-point win. But it's ludicrous to think that a team gets the same credit (in ranking systems) for a 50-point win that they would get for a 1-point win.

  5. Ian says:

    Adam - I had a thought that you could use the normal distribution c.d.f. to work out how many 'ranking points' a result gets. For instance, in the NFL the standard deviation of winning margins is 14.9 - so a 50 point win becomes normalised to a z-score of 3.36, scoring 99.96%. A 57 point win gets you 99.993% - so ever so slightly more. A 7 point win would get 68.1%, and scoring another TD gets you a 14 point win and a score of 82.6% for ranking.

    This would help provide a good ranking for college teams but whilst also removing the benefit for teams to run up the score.

    The problem with rankings like this is that it's always a compromise between a clearly understandable system giving ropey ranking and one that provides accurate ranks but is too complicated for anyone without a degree in maths to understand.

  6. Chris says:

    Hello Brian,

    I've noticed that the running game (and by extension run defense) has received a lot of bad press lately and so I made a post on my blog that I hope might redress the balance just a little. I'd be very grateful if you'd check it out and give me your thoughts.

    http://keepingthechainsmoving.blogspot.com/

    Cheers,
    Chris.

  7. Thomas says:

    Brian,

    I think you missed a "we needed a study for this?" finding that they list: "Highly-drafted wide receivers without many college touchdowns are likely to bust."

    No, really? I wonder how large a data set this is. I can't really imagine an NFL GM drafting a WR in the 1-2 rounds who hasn't had a significant number of TD's in college.

  8. Jeff Clarke says:

    I think sometimes you're a little harsh on F.O.

    I agree that the "Rule of 370" is really gimmicky and a bit of cherry picking numbers but I found their piece on defensive penalties really interesting.

    They did attempt to answer the "Why don't defensive penalties matter". The answer is that off penalties are typically a screw up of some kind, but defensive penalties are typically a result of playing just a little too hard. It does make a certain amount of sense.

  9. Miles Libbey says:

    "It's possible they've fallen for the same trap I fell into years ago. The NFL calls penalty yards by an opponent "defensive penalties." But if true, it's interesting. The real question would be why don't defensive penalties matter?"

    I don't get the trap you fell into. Can you briefly elaborate?

Leave a Reply