THE BOOK--Playing The Percentages In Baseball

I’m not sure the OPS+ article is helpful wrt addressing misunderstandings about the stat. 

1.  It never explains why its (relative OBA + relative SLG - 1) rather than /2.

2. It says 110 is 10% above league average...but it doesn’t explain that means 10% above average in estimated runs.  The reader is still left to believe its 10% above average in OPS, but it’s really more like 5%.

3. “OPS+ does share the core weakness of OPS — namely, it gives equal value to both OBP and SLG, though sabermetricians agree that, point by point, OBP is a more valuable stat than slugging.”

OPS+ undervalues OBA, no doubt--but it is not weighting it equally than SLG, it’s weighting it at approximately 1.2 times SLG. 

Perhaps I’m just incapable of seeing it through the eyes of its target audience.

This statement concerning BAPIP is simply wrong:

“If the batter’s line drive rate remains constant, BABIP can be a good indictator of “luck” or random chance, which tends to even out over the course of a long season. If a lot of hits are falling when put into play, a hitter is usually due for a dry spell or a pitcher is usuall due for a break. If very few hits are falling, a hitter may be nearing a hot streak, while a hot pitcher may be due for an upcoming reality check.”

This is like saying in roulette, a black number is due after 5 red numbers…

I always hate when I hear people say “a pitcher has no control over his BABIP”. I really don’t think that’s true. I believe pitchers with good peripherals (thus, with low FIPs) tend to have below-average BABIPs.

For instance, I tried to think of the best pitches of my lifetime, and I looked up their career BABIPs. I looked up Pedro, Johan, Randy Johnson, Clemens, Sabathia, Rivera, Hoffman, Mussina, Peavy, Maddux, Smoltz, Glavine, Wagner, Mark Buehrle, David Cone, David Wells, Dwight Gooden, Kevin Appier, and Bartolo Colon (I got tired of thinking of people at this point). And every one of them had a BABIP below what b-r.com has as the MLB average (.299).

The only good pitchers I found who had above-average BABIPs were Chuck Finley (.300) and Roy Oswalt (.304)

My guess would be that there’s at least some correlation between BABIP and a defense-neutral stat like FIP.

Tom: read these:

http://www.tangotiger.net/DIPS175.htm
http://www.tangotiger.net/DIPS20.htm

Yes, you are right.  But, the impact is fairly muted.

Excuse the n00b here, but isn’t wOBA the equivalent of OPS+ that does take into account weighting OBP over SLG?

Very cool. Thanks Tango!

Yea Yahoo!

I assume the last piece in the series will somehow feature Tony Randall and Burt Reynolds.

Wells/#5, wOBA is more analogous to OPS with proper weighting than to OPS+ with proper weighting, because wOBA shares none of the adjustments that differentiate OPS+ from OPS.  wOBA, unlike OPS+, has no adjustment for park, and the only adjustment for league is to make average match up with average OBP for that season.  So a .400 wOBA in one season or era might not be the same as a .400 wOBA in another season or era, and a .400 wOBA in one park might not be the same as a .400 wOBA in another park, whereas OPS+ strives to put all of those contexts on the same scale so you can make direct comparisons.

Kincaid/#8 - thanks. Would it make sense for someone to look at adjusting wOBA for park and league factors? I guess you’d lose the benefit of wOBA looking similar to OBA, which is one of the niceties of the metric.

Wells/#9: It can and has been done. And it really doesn’t take away from the scaling, as it looks similar either way. StatCorner does have park-adjusted wOBA’s based on NL/AL averages, as well as “wOBA+,” which works similarly to other “+” metrics.

Depending on what you want to use it for, adjusting wOBA for those things would be an important step.  For example, FanGraphs bases its run values off of wOBA, but in converting from wOBA to offensive value, they apply a park adjustment.  There is also a factor unique to each year (usually around 1.15)that adjusts (wOBA - lgwOBA) to runs above average, so in converting wOBA to runs, you would be using a league adjustment of sorts.

In fact, one of the reasons that wOBA contains no adjustments is that it allows you to make whatever adjustments you want for whatever you want to use it for.  Since there are different ways to apply those adjustments and different people have different preferences for how to handle them, leaving the adjustments out is a feature because it let’s you know exactly what adjustments are happening because you have to do them yourself.  Generally, as you say, a large part of the appeal of the rate stat itself is that it is scaled so it looks like OBA, so making adjustments to the rate itself aren’t all that important, but in converting wOBA to runs or wins or value of some sort, those are definitely adjustments you’d want to make.

Very nice that wOBA is being covered next week!!

While wOBA does measure something similar to about (1.8*OBP) + SLG, its a completely different calculation.  When you calculate OBP and SLG, they are based on different denominators.  OBP is plate appearnaces, and SLG is based on at bats.  Walk are included in only one denominator.

wOBA uses Tango’s linear weight to give a value for every different possible result of a plate attempt, sums up the values of a players contributions, and then divides by plate appearances.

Great to hear you will be covering wOBA.  Anything like this that helps catch the mainstream up on the more current standards is much appreciated.