Tuesday, December 15, 2015

Statistical Meanderings 2015

This is an annual, largely analysis-free look at some things that I found interesting when compiling my end of season statistical reports. My whole series of annual posts will be a little late and a little brief thanks to some computer issues that prevented me from working on them for a few weeks. They might be the better for it:

* Minnesota was 46-35 at home and 37-44 on the road, close to an inverse record. Nothing noteworthy about that. More amusing is that they almost had an inverse R-RA, scoring 373 and allowing 323 at home while scoring 323 and allowing 377 on the road.

* Every year I run a chart showing runs above average (based on park adjusted runs per game) for each playoff team’s offense and defense. Usually I do this and get to slyly point out that the average playoff team was stronger offensively, but that is not the case this year, and it would be bad form not to show it even when there are no guffaws to be had:

Although it is interesting narrative-wise that the Mets’ offense wound up being twenty runs better than their defense.

* There were nine teams whose starters had a lower eRA than their relievers, led by the Dodgers (3.67/4.36) and also including the A’s, Red Sox, Mariners, Cubs, Rays, Braves, Cardinals, and Mets. One might note that four of the five NL playoff teams are represented; only the Pirates had a lower bullpen eRA (4.09/3.50).

In 2014 there were eight teams with a lower starter eRA and two made the playoffs; in 2013 seven with two playoff participants; in 2012 five with just one playoff club; and in 2011 eight and two.

I certainly would not claim that this little piece of trivia demonstrates any larger truth about the importance of starters and relievers, but it certainly is the kind of factoid that could be used in the style of a Verducci to do so. Of course, the blessed Royals completely break the narrative as the team with the biggest difference in favor of their relievers (4.74/3.34; that 1.39 run gap was much higher than the next closest team, the Brewers (4.92/3.88)).

It is also interesting to see the Rays on the list given the attention they got for aggressively pulling starters on the basis of times through the order. Tampa was 23rd in the majors in innings/start, but second to last in the AL (TB starters worked 5.65 innings per game, KC 5.63).

* Speaking of things you’re probably not supposed to say about the Royals, they were an excellent fielding team with a .690 DER, fourth in the majors. But the two teams they best in the AL playoffs each had a better DER (TOR .696, HOU also .690; San Francisco led the NL at .694).

* Minnesota starters had a 4.68 eRA, above the AL average of 4.47 and in the bottom third of the circuit. But this was a big improvement from their deplorable pitching of the last three seasons. That leaves Philadelphia as the team that can make everyone else feel good about their rotations. Phillie starters had a 5.59 eRA, much worse than their closest competition, Colorado at 5.08 (these figures are all park-adjusted). Rockie starters were last in IP/S (5.29, .2 innings fewer than PHI and ARI) and QS% (33%, MIL at 39% and PHI at 41%).

* If you’d have given me ten guesses, I’m not sure I would have come up with San Francisco leading the majors in park adjusted OBA (.342). In my defense it was a BA-driven performance as their .278 BA was nine points better than Detroit and their walk & hit batter per at bat ratio was .097, just three points above the NL average.

* Dellin Betances and Andrew Miller were 1-2 among AL relievers in strikeout rate. Granted he and Miller wouldn't have both been in the same bullpen, but David Robertson was third.

* Evan Scribner had one of the craziest lines you will ever see. He struck out 64 and walked 4 in 60 innings, but he yielded 14 homers, so he was sub-replacement level (I have him at -4 RAR, which is based on runs allowed adjusted for inherited and bequeathed runners). Scribner had the best K/W ratio among relievers; the next best was Kenley Jansen at 80/8.

If you rank AL relievers by the difference between strikeout and walk rate ((K-W)/PA), a better metric, Scribner ranks eighth. The seven relievers ahead of him were all at least 15 RAR except David Robertson (6). The next sub-replacement level relievers on the list are Aaron Loup (17th) and Mike Morin (19th), but both of them were hit-unlucky (.352 and .353 BABIP respectively) and comfortably above average in dRA. To find the next sub-replacement level performance you have to go all the way down to 46th and Danny Farquhar.

Scribner's 2.2 HR/G (games based on 37 PA rather than 9 IP) rate was the highest among major league relievers. The top three AL relievers in HR rate were all A's: Fernando Abad (2.0) and Edward Mujica (1.9), but OAK's HR park factor of 93 is tied for lowest in the AL.

* My stat reports set a minimum of 40 relief appearances to be included as a reliever, but sometimes I cheat and let in players I’m interested in. One case this year was Jeff Manship. Manship pitched 39 1/3 innings over 32 games. But if you include him, he:

1. Led in RRA (.67 to Wade Davis’ ridiculous .75 over 67 1/3 innings)
2. Led in eRA (1.51 to Davis’ 1.79)
3. was 13th in dRA (2.92, teammate Cody Allen led the way at 2.24)
4. And as you probably surmised by now, led the AL in lowest BABIP (.194, Will Harris was next at .201. Manship’s teammate Allen of the league-leading dRA gave up a .348, eleventh worse of the 95 AL relievers)

Terry Francona frequently used Allen in the eighth inning. Allen’s .37 IR/G was fifth among AL relievers with double digit saves, and Roberto Osuna was the only one of those five with twenty or more saves (twenty on the nose and .49 IR/G). Allen allowed only 4/26 inherited runners to score, lowering his 3.38 RA to a 2.86 RRA

* Does Jerry DiPoto know that Joaquin Benoit had a .190 BABIP? (That's not intended as shot at Jerry DiPoto, Benoit was in the news so it stood out.)

* Ground zero for DIPS intrigue was Toronto. Toronto led the majors with a .696 DER, and their starting pitchers with 15 starts were:

1. Marco Estrada, who had the highest ratio of dRA/eRA (basically, my DIPS run average to a component run average, both based on the same Base Runs formula but the latter considering actual singles, doubles, and triples allowed) of any AL starter (4.73/3.40) thanks to a .223 BABIP

2. RA Dickey, who ranked eighth with 4.72/4.00 and as a knuckleballer falls in one of the first categories of pitchers Voros McCracken carved out of DIPS theory

3. Mark Buehrle, whose dRA/eRA ratio in his final (?) season was an unremarkable 4.52/4.41 but who over the course of his career was an occasional DIPS lightning rod

4. Poor Drew Hutchison, who had the third lowest ratio at 4.46/5.50 and was pounded for a .344 BABIP. On the other hand, he had a 13-5 record despite his BABIP-fueled -6 RAR (second worst in the league, ahead of only...)

* One of the more amusing bits of media silliness during 2015 was Bill Madden's fixation on Shane Greene, which included a caption on an article that asked if Shane Greene was Brian Cashman's biggest mistake, and Madden pondering whether the Yankees would still rather have Nathan Eovaldi and Didi Gregorius than Greene and Martin Prado.

Greene was the worst starting pitcher in the AL with -13 RAR.

I like that as a punchline, but the alternate punchline is that while Prado hit fine (20 RAR) for Miami, Gregorius hit enough (only -3 RAA versus an average shortstop) and gave New York their first good fielding shortstop in goodness knows how long, while Eovaldi chipped in 22 RAR. 36 RAR to 7 RAR, I think Cashman is pretty happy with his choices.

* I will point out that my RAR formula includes no leverage adjustment (which I defend), but then leave this without further comment because you can get chastised for talking about this:

2015 RAR
Jake Odorizzi +36
Wade Davis +30
James Shields +26
Wil Myers +14

* Would you concur that it’s plausible that all five of these seasons could have been produced by the same pitcher?

These are by no means the five most similar in value seasons you could pull out of this year’s pitching lines, but they are broadly similar, no? The reason I like this group so much is that the pitchers are John Lackey, Shelby Miller, Jaime Garcia, Carlos Martinez, Lance Lynn, and Michael Wacha. Not only did St. Louis use five clones as their rotation, they traded a sixth away.

* Ichiro was last in the NL in RAR as a 42 year old corner outfielder. His batting average--Ichiro's batting average--was .229. It is almost inconceivable that he will get another job and that my Twitter feed will react with anything but scorn. But sometimes the inconceivable is reality.

* Speaking of Marlins with terrible secondary averages, Dee Gordon posted a .128, same as Suzuki. The only NL hitters with 250 PA and SECs lower than .128 were Milwaukee's sometimes double play combination of Jean Segura (.110) and Hernan Perez (.101). Ben Revere posted a .128 as well, and both Revere and Gordon were above average offensively, but the next lowest SEC by an above-average NL hitter was .151 (Brandon Phillips). I remain skeptical about Gordon's long-term outlook; it is exceedingly rare for a player to be able to remain an offensive contributor with so little to offer other than singles.

Nonetheless, for 2015, Gordon was a defensible choice for the Silver Slugger, as only Joe Panik had a higher RG and he compiled 220 fewer PA. Still not a good look for NL second basemen.

* The Speed Score trajectories of Bryce Harper and Mike Trout have been something I’ve been watching as much has been made of Trout reconstituting his offensive game as more of a power hitter and less of a baserunning threat. But as of last year his Speed Score was still quite high, albeit lower than when he broke in. With a fourth season under his belt, Trout’s Speed Score sequence is 8.7, 7.0, 7.2, 4.9. So 2015 did mark a significant downturn in terms of Trout’s speed manifesting itself through the official statistics (or at least stolen base percentage, stolen base attempt frequency, triples/BIP, and runs scored per time on base).

Meanwhile, Harper’s sequence is 7.5, 4.9, 2.7, 3.0. If that keeps up Dusty Baker will accuse him of clogging the bases.

* The best season you probably weren’t aware of (which is really to say the best season I wasn’t aware of): Logan Forsythe hit .287/.370/.454 over 609 PA, good for 39 RAR. It was basically the same season as Jason Kipnis had trading some BA for SEC.

No comments:

Post a Comment

I reserve the right to reject any comment for any reason.