(10-29-2020, 09:53 AM)NeptuneBeachBum Wrote:

(10-29-2020, 07:56 AM)mikesez Wrote: I didn't say DVO was better than simple calculations like passer rating or yards/attempt.
I said, if you want complex, you can still do better than QBR. Football outsiders have explained enough of their methods that somebody else with 25 years worth of standard play-by-play text files and computer programming skills could replicate it.  ESPN has not done this.  They don't want you trying to replicate or check them.

1.  You said, "DVOA that Football Outsiders made their name on is basically a supercomputer version of what I would do"; showing you changing your mind about QB assessment and your recognition that contextual evaluations it is better than simple calculations.  But it also shows your ignorance that DVOA is philosophically the same as QBR.

2.  You said  and "Football Outsiders have explained enough of their methods that somebody else with 25 years worth of standard play-by-play text files and computer programming skills could replicate it.".  Again, you are displaying your full ignorance here.  Football Outsiders and ESPN have both explained the generalities of how they do their calculations, but neither has released the specifics of their proprietary formulas... its their copyrighted intellectual property.  So when you say Football Outsiders have "put their stuff out there", you are have no idea what you are talking about.  They have not done that any more than ESPN has.

3.  You state your preference for DVOA over QBR by asserting "don't trust something you can quantify if you can't also verify it".  But you can't verify or replicate DVOA.  You don't have their formula.  You are just making another attempt to grasp at straws to save face in a discussion that has gone dramatically off the tracks for you.

4.  Quit while you are behind.  Way behind.

You have two problems. You are misunderstanding a lot of what I say, and you're making a lot of incorrect assertions. Each time I try to correct you, you continue to misunderstand me. DVOA uses 25+ years of actual games to create their baseline.  Creating a baseline like that is extremely tedious, but it's not hard to gain access to the data behind it, and the baseline is stable across teams and across years.  QBR seems to use real time simulation of the result of the current game as its baseline.  That's much more volatile and harder to explain.

TL;DR - just let people disagree with you, dude.

---