Usually only playing 1-2 leagues per year (for the last 10 yrs), maybe U guys can
clue me... It's about what eye would like 2 see, & whether or not any of U see any
chance that something like this could actually be accomplish'd here, & be seen...
The word 'normalization' seems 2 be used as an excuse, for 'estimating' precise
historical numbers... Does the '23 BRuth exactly resemble itself after 10,000 PA's
or after 10 million PA's (?)... A home team in Coors, w/ 3 BRuth clones, batting in
the line-up, pretty much means 3 hitters w/ 100 homers each, which is where the
'normalization' (although extreme in this example) seems unreal & useless, & as
a continually failing experiment, for that particular type of league...
It was not that long ago, when MMantle & RMaris became the 1st team-mates with
more than 50 homeruns each... A dozen yrs later, HAaron, DaEvans, & DvJohnson,
became the 1st trio of team-mates 2 hit 40 or more homers each... Then, late in the
'70's, the LADodgers became the 1st team w/ 4 team-mates, who hit 30 homeruns,
or more, w/ SGarvey, RCey, RegSmith, & DuBaker... Realistically, -this, -just seems
MORE real 2 me... Not outrageously guessed by a path of old-Wi$ 'normalization'...
What seems a more rational & logical approach from my point of view, would be 2
just simply 'normalize' EACH LEAGUE individually, truly based upon those pitchers
& batters within THAT PARTICULAR LEAGUE, prior 2 it's commencement...
The stud pitcher, AJoss, seems 'normalized' across the entire database, handling
RHornsby's & BRuth's in successive PA's, pretty routinely... In other words, -teams
fill'd w/ AJoss's & BRuth's should be RE-normalized against itself... An example is
also applicable 2 those Prog-Reg leagues using a 'normalization', that is intend'd
ONLY for whatever pitchers/batters used over it's sole duration... Whether in OL's,
or Themes. or Prog-Reg types, there still seems a better way to achieve realism...
That's my point, & question... Can a different type 'normalization take place here (?)
9/25/2013 5:07 PM (edited)