About Me

My photo

Nice guy.  Have some blogs.  Do baseball research.

Thursday, May 10, 2012

baseball-reference.com math: confusing, convoluted, elitist.

Radical Baseball: The Art Of WAR 2.0, 2.1, … May 9th, 2012

You really need to read about WAR 2.0 for yourself.  There’s a lot to it and Sean Forman and his staff at baseball-reference.com are to be commended for making a real attempt to explain WAR to it’s subscribers.


But I have some suggestions.

1. Parenthesis.  Many equations seem to lack parenthesis that would make them correct.
2. One equal sign per equation.
3. No greater than sign in an equation.
4. Define terms in each document.
5. Explain everything.

There's a tendency to explain relatively easy stuff at 5mph and then the complex stuff at 100 mph.  Leaving the reader with that "what the heck?" feeling.

WAR 2.01: Renaming ndWAR to afWAR Posted by admin on May 4, 2012

The change from oWAR and dWAR to ndWAR and dWAR is confusing folks, so I'm making an additional change.

dWAR was confusing to begin with since I think it dealt with fielding.  Defense to me is fielding and pitching.  Why isn't it fWAR, f for fielding?

ndWAR, I guess is non-defense WAR?

afWAR => Average-Fielding-WAR.

This is a definite "Say what?".  What's with the greater than sign?  And what the heck is to the right of that?  Are those dashes?  Subtraction signs?  And what does fielding have to do with what very recently had been oWAR, offensive Wins Above Replacement?

afWAR = batting + baserunning + DP's + replacement + position

I'm guessing that the word position means that players are compared to other players at their same position but I still don't see the fielding in the term Average-Fielding.

afWAR + dWAR does not equal WAR as you'll be double counting the position adjustment.

Too bad, just when my common sense might have kicked in.  Argh!


Baseball-Reference.com WAR Explained

an equation should explain what we are doing here.

Players Runs over Replacement = Player_runs - ReplPlayer_runs = (Player_runs - AvgPlayer_runs) + (AvgPlayer_runs - ReplPlayer_runs)

This gives us runs above replacement (RAR).

An equation might explain it but not that equation.  Why are there two equal signs?

Pythagorean win-loss records

I happen to know this one but a link or simple explanation should be included.  Same with PythagenPat.  Come on, give the average fan a fighting chance.

Data Coverage: This link is useful except the headings disappear as you page down and baseball-reference didn't even have the sense to repeat them at the bottom, which would have helped somewhat.  Those helpful popup definitions that are elsewhere on baseball-reference.com are missing here.


wRAA For Position Player WAR Explained

we also exclude IBB's from the count of BB's. Our view (and those of the creators of these stats) is that SH and IBB's are managerial decisions, and in general the best way to handle these is to exclude them from the rate stat (wOBA)

I don't get excluding IBB.  The team derives the same benefit as a regular BB but it's a minor point.

wOBA = (0.70 x uBB + 0.73 x HBP + 0.89 x 1B +1.27 x 2B +1.61 x 3B + 2.07 x HR +0.25 x SB + 0.50 x CS) / (AB+BB-IBB+HBP+SF)

Two things about this mess:
1. Shouldn't there be parenthesis around each two items multiplied?
2. Wouldn't the whole thing be more readable if items were placed on a separate line?

wOBA =
((0.70 x uBB) +
(0.73 x HBP) +
(0.89 x 1B) +
(1.27 x 2B) +
(1.61 x 3B) +
(2.07 x HR) +
(0.25 x SB) +
(0.50 x CS))

There.  Much better.  And why exactly are stolen bases (SB) and caught stealing (CS) included in weighted On Base Average (wOBA)?

There's too much that's too confusing here but one more item deserves mention: wRAAA.  What the heck is that and where did it come from?  It suddenly appeared in:

Top 30 Seasonal Increases from wRAAA to wRAA_adv

There's no mention of wRAAA  prior, nor of wRAA_adv for that matter.

You see the problems.  Even if an average fan tries to learn this stuff it quickly becomes overwhelming.  It does not need to be this confusing, convoluted and elitist.  Fix it.

No comments: