Loading ...
Sorry, an error occurred while loading the content.

1630Re: Dilution, balance, and Bob-bashing

Expand Messages
  • Dean Oliver <deano@rawbw.com>
    Dec 31, 2002
      --- In APBR_analysis@yahoogroups.com, "Dean Oliver <deano@r...>"
      <deano@r...> wrote:
      > The standard approach is the Stratomatic approach -- If Team A
      > averages 4 possessions over the league average and Team B averages
      > possessions over the league average, then Team A vs Team B will on
      > average be 6 possessions over the league average. This is roughly
      > comparable (on a first order differential basis) to the Adams-James
      > rule for evaluating how well a 0.300 hitter will do against a
      > who allows opponents to hit 0.300 in a league that averages 0.260.
      > Mechanistically, I've never been sure that this is right. I'd be
      > curious if data exists to support this or whatever Bob says. I
      > I have the data, but never looked at it that way.

      Let me backtrack a little here. This is not necessarily how I'd "do
      it", as Bob says, or "simulate it". It depends on what the goal of
      the simulation was for. If I was trying to do a simple simulation of
      how many points teams would score, this is how I'd "do it". If I was
      trying to simulate the conditions under which Ron Harper could play
      with Michael Jordan and score 20 ppg, I'd only use this as a soft
      constraint on the results and focus on other inter-relational
      aspects. It might be a hard constraint, depending on the mechanism
      of simulation. I have different simulations for different studies.
      I haven't exactly designed a simulation for pure "fun" of simulating
      across generations.

    • Show all 35 messages in this topic