Exactly How Full of S is OPS? (August 6, 2003)
Regression analysis over 30 years at the team level shows that the best estimator is 1.64*OBP + SLG. Coincidentally, taking the complete bare-bones approach I did with my articles a few months ago on "OPS: Begone!", I also came up with this exact same equation.
--posted by TangoTiger at 11:56 AM EDT
Posted 12:10 p.m.,
August 6, 2003
(#1) -
bob mong(e-mail)
(homepage)
Who is this guy? That article is fantastic!
Well written, humorous, informative, and easy to read - get him to write some stuff for Primer!
Posted 3:20 p.m.,
August 6, 2003
(#2) -
Rally Monkey
That's what's wrong with the Angels. Should have kept Palmeiro.
Posted 9:11 p.m.,
August 6, 2003
(#3) -
Innumerate masses
Teacher? Can I get my microscope back now?
Posted 2:19 p.m.,
August 7, 2003
(#4) -
Steve
Great article. There is one detail needs to be clarified, though.
DePodesta's model says that an * EXTRA * point of OBP is worth 3 times more than an * EXTRA * point of SLG. The baseline for those extras are never defined. Is there some threshold value of OBP and/or SLG that should be considered? Has anyone done any studies on this?
Posted 3:08 p.m.,
August 7, 2003
(#5) -
tangotiger
I mentioned this in my OPS article, but that doesn't matter.
Value = (OBP - baseX) * 3 + (SLG - baseY) * 1
Value = 3*OBP + SLG - 3baseX - baseY
Value = 3*OBP + SLG - whatever
As you can see, it's irrelevant what baseX, baseY, or whatever equals. The DIFFERENCE among the players will remain exactly the same.
Try it out. Take a 400/500 player and a 300/600 player, and take whatever baseline for OBP and SLG you want. The difference between Value(player1) and Value(player2) will be exactly the same.
Posted 6:20 p.m.,
August 7, 2003
(#6) -
Scoriano
Obvious point: To make comparisons valuable, the product of any of these shorthand formulae still needs be adjusted for park context, etc.
Posted 8:41 p.m.,
August 12, 2003
(#7) -
Michael J
So does this mean that Mientkiewicz (however you spell that ridiculous name) is actually a pretty legitimate first baseman considering he has had a relatively high OBP the last 3 years combined for any player and a Slug for a first baseman. I see him ridiculed occassionally on this site, that's why I bring it up.
Posted 8:57 p.m.,
August 12, 2003
(#8) -
Michael J
that should read relatively low Slg percentage for first baseman.