The Uselessness of OPS

Last week I dusted off my dusty old notes from 2011 which proved irrefutably that Peter Edward Rose's unassailable all-time hit record would withstand assailment from Derek Jeter and updated those notes with data from the 2012 season.

The notes I created for that 2011 analysis were jotted onto the back of a printout that previewed the 2011 Belmont Stakes.  In and of itself, that isn't significant.  What is of interest for us, this week, is that same 2011 Belmont Stakes printout also included notes I made which excoriated once and for all the utter uselessness of that darling new stat of SABRmatricians everywhere, OPS [or On-base Plus Slugging].

Ultimately, baseball statistics are important for two reasons; Primarily to contextualize what a player has achieved and secondarily to project what that same player is likely to do under given circumstances (such as for an upcoming at bat or perhaps extrapolated over the course of a season).

When, during the 1992 season, Reds manager Lou Piniella summoned Rob Dibble from the bullpen to close out the 9th inning, we fans understood that Dibs' 14.1 strikeouts per 9 innings suggested he was probably going to strike out at least one batter that inning, and perhaps more than one batter.  In 2007, Adam "Big Donkey" Dunn's statistical record suggested that during any given at bat he was 6.3% likely to blast a home run, had a 16% chance of being walked and a 26.1% chance of striking out.  These are outcomes that a manager might project, or that we fans may anticipate, during any given at bat due to the simplicity - the clarity - of undiluted statistics.  The moment one blends the results of differing measurable events in order to project a future outcome then any projection will necessarily be less accurate.  If the point of maintaining a statistical record is to understand as precisely as possible what has already occurred and to utilize that record in order to then make as precise a projection about a future outcome as may humanly be possible, the effort to combine past results of differing events adds disparate and incompatible data upon which those projections are based and renders such exercises fruitless.

Now, a situational example:

Bottom of the 9th inning, two outs, a runner on first base and the home team is down one run.  That run must score.  The manager of the team sends up to bat a batter with an OPS of .800.  As that .800 OPS batter digs into the batters' box, on deck is a Drew Stubbs-type batter who the manager and every fan watching knows will strike out on three pitches.  Ergo, the player now standing in the batters' box must drive in the baserunner (after all, OPS relies on Slugging Percentage which measures how many bases the batter advances per at bat).  If that .800 OPS batter is Ichiro Suzuki [career .798 OPS during the 2011 season when my notes were first generated, it's now .787 and I'm too lazy to look up a 2013-comparable .800 OPS replacement for Ichiro], his career 162-game average for doubles was 27 and home runs was 9, his OBP was .370.  But simply being issued a base on balls (a "walk"), or hitting a single (which Ichiro was doing, through 2011, approximately 180 times per season) - key elements determining OBP - would not have the desired result, that being the batter advancing multiple bases and therefore driving in the baserunner.  Is this the .800 OPS batter the manager would send up to bat in this situational example?  If that .800 OPS batter had been Carlton Fisk [career .797 OPS], a batter who average fewer than 100 singles per season yet averaged 27 doubles (identical to Ichiro through 2011) and 24 home runs, more than twice as many as Ichiro, and, therefore, was more than twice as likely to hit a home run than Ichiro.  In the situational example I provide here, Fisk is more likely to drive in that baserunner than Ichiro, and by a significant margin.

Two .800 OPS batters, two different likely outcomes.  

So, what is the point of OPS?  What does it suggest to you about what a batter has done or is likely to do?  Comparing a .300 hitter versus a .275 hitter gives you a clear delineation.   Comparing a .400 OBP versus a .350 OBP gives a clear difference.  Comparing an .800 OPS versus a .700 OPS is meaningless; How many walks and hits added to that OPS as opposed to how many doubles or home runs?  OPS itself does not answer that question.  Since it cannot answer that question then one cannot use it to project - or anticipate - a likely outcome.  There is no doubt that walks and singles do not produce the same results as doubles and home runs.  So why add them together?  Senseless.

Into his 2011 season, for his career Prince Fielder [.927 OPS] and Tris Speaker [.928] shared OPS.  Speaker's OBP was .428, Fielder's under .380.  Fielder's SLG was .540, Speaker's .500.  Speaker averaged 7 home runs per season, Fielder was averaging 38 per season for his career [as of 2011].  Do we project, do we anticipate, similar outcomes for their at bats?  Of course not.  Would you project, or anticipate, similar outcomes for two .300 hitters?  Two .400 OBP batters?  Two .500 sluggers?  Yes to all questions, and certainly with a much greater degree of accuracy than you would for batters with similar OPS.

Were Luke Appling and Vinny Castilla similar batters?  Both had an OPS of .797.  In a twenty year career Luke Appling never hit more than 8 home runs in a season but walked more than 100 times per season three different times.  Vinny Castilla had six seasons with 30+ home runs (three of which were 40+ seasons) but never took as many as 55 walks in a season.  Batting average is the largest portion of calculating OBP; Castilla batted .276 for his 16-year career, Appling .310.  Appling's lifetime OBP and SLG were practically identical; .399 OBP and .398 SLG.  Castilla's OBP was .321 and his lifetime SLG was .476.  Would anybody equate the performance, or project the likely outcome of an at bat, of Appling and Castilla?  No, yet that is precisely the purpose of OPS.

For their careers, Ted Kluszewski [.850] and Minnie Minoso [.848] were both .850 OPS batters.  Klu's lifetime OBP .353, SLG .497.  Minoso's lifetime OBP .389, SLG .459.  Empirically Big Klu and Minnie were different hitters, as were Luke Appling and Vinny Castilla.  Why invent a formula that creates an illusion of similarities which did not (or do not) exist?

OPS lacks clarity which is the essence of statistical record keeping.

Ave Atque Vale

Springfield, Ohio native Jonathan Winters slipped the surly bonds of earth to touch the face of God this week [credit: Ronaldus Magnus] at the age of 87.  Born in Dayton, Ohio in 1925, as a young boy he moved to Springfield, Ohio at just about the time My Dear Elderly Mother was born there.  Winters has long been a favorite of the current and former denizens of The Ranch.  While it would be impossible to cite a single, preeminent accomplishment, Jonathan Winters' performance in It's A Mad Mad Mad Mad World (a film highlighted on the ol' web page numerous times) merits recognition here, today, as a small way in which I can pay my respects.  First up, a one-minute clip which my Obamanomics/Euro-socialist/99%er pals will thoroughly enjoy:

Jonathan Winters' case for paying taxes

And the sequence that never fails to have me laughing until it hurts, the infamous gas station scene:

Gas Station Destruction Scene

Long-range Reconnaissance

Prior to Reds Opening Day, I introduced most of you to the Chicago Cubs old ballpark (not Wrigley old, but older even than that!), the West Side Grounds.   That ballpark featured a grandstand that resembled a section of the Second City's celebrated "El." Here's a photo from the same perspective as you saw before, but from 1904 (instead of being from the 1906 World Series.  See; March 26 post, below):

In 1907 the Cubs defeated Ty Cobb's Detroit Tigers in that season's Fall Classic.  The photo below gives you a closer look at the grandstand during the 1907 World Series:

By 1908, the Cubs extended the grandstand further down the lines and also had, evidently, relaxed their on-field seating policy by permitting fans to bring seats and benches onto the field:

The next chapter in Long-range Reconnaissance will interrupt the on-going baseball theme for one that spotlights The First Saturday In May.