Dissecting the Kyle Lohse Start: Beware! FIP!

Opening Day 1.5 featured a one-game series with the defending world champs, the St. Louis Cardinals, and the league’s latest makeover recipient, the Miami Marlins. Righty Kyle Lohse earned the Opening Day honors for the Cardinals on the merit of being not recently or presently injured, and much to the surprise of many, Lohse took a perfect game no hitter into the 7th inning.

His line from the game:

7.1 IP, 1 ER, 3 K, 0 BB, 8 GB, 10 FB, 2 LD

All told, that comes to a 1.23 ERA, 1.49 FIP, 2.94 xFIP, and… a 4.22 SIERA

Everything but that SIERA number suggests Lohse had a great start. Let’s find out why.

The most important item to remember — and not just with Lohse, but with Josh Johnson, Felix Hernandez, and every pitcher who has pitched in this extra-young season so far — is that the FIP constant here at FanGraphs is not, well, constant. The Dark Overlord has some magical algorithm in place to keep the constant (usually around 3.2) variable according to the league’s run environment. Because the regular season has all of three games of data right now, that constant is unusually low — in the ones last I heard.

So Lohse’s 1.49 FIP is actually less impressive in the context of the four team “league.” His FIP- is 109 and his xFIP- is an even more troubling 164. With a 3.2 constant, the comes out to a 2.79 FIP and a 4.56 xFIP. Suddenly his SIERA, which sits near his career levels, does not look nearly as strange.

Which is all to say: Beware of FIP right now! Just because a number looks intrinsically good, it’s best to reference FIP- or SIERA to ensure the number is not biased by the current run environment.

So what was it about Lohse’s start troubled SIERA? Lohse did not walk anybody — though he did hit Emilio Bonifacio in the 4th inning (before removing him with a double play) — and his strikeout rate (about 12%) was not bad. Presumably, it must be his balls in play.

Lohse allowed 20 balls in play, but only 2 hits — good for a .100 BABIP. Last year, grounders went for a .237 BABIP, flies a .137 BABIP, and liners a .714 BABIP. If we apply these percentages to Lohse’s numbers, he comes out with about 2 ground ball hits, 1 fly ball hit, and 1 line drive hit (an expected BABIP of .200). SIERA is sensitive to what kinds of hits Lohse allows, but four predicted hits is not terrible — in fact, the 66% ground ball rate is much better than his career numbers.

However, his career homerun per flyball rate — 9.7% HR/FB — suggests he would have typically allowed at least one homer in a game such as this — which makes sense. A guy averaging 1.10 HR/9 probably would allow a homer in the typical game where he pitches nearly into the 9th inning.

In fact, he almost did, too. In the bottom of the second inning, Giancarlo Stanton redirected a pitch to deep center field, going for about a 405-foot out.

The new Marlins Park should play like SafeCo Field with even fewer HRs to center field, which could be bad news for Stanton, who, more than one time, took advantage of Dolphin Stadiums’ cut out in CF:

Stanton’s Dolphin Stadium Homers

Courtesy of Katron.org.

All this to say: (1) Lohse pitched well, but still looks very much like his Lohsian self, and (2) the new Marlins Park seemed to help him out a bit, but that may mean more for Miami’s home run hitters than for Lohse, and (3) above all, look at FIP- and xFIP- and all the minus stats instead of their traditional counterparts — especially early in the season. None of these stats may not be very predictive right now, but at least the minus stats (and SIERA) are better at being descriptive.




Print This Post



Bradley writes for FanGraphs and The Hardball Times. Follow him on Twitter @BradleyWoodrum.


35 Responses to “Dissecting the Kyle Lohse Start: Beware! FIP!”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Person says:

    Technicality: He hit a batter (I think in the 4th inning) then got a GIDP. So he faced the minimum while not allowing a hit through 6, but he did not have a perfect game.

    Vote -1 Vote +1

    • Person says:

      then I read the full article, where you do mention the hit batter.

      i shouldn’t make such a fuss over a minor point, though

      Vote -1 Vote +1

    • cream says:

      That’s not a technicality. It’s pretty basic.

      Vote -1 Vote +1

    • FanGraphs Supporting Member

      Oh, good point. I don’t know why that didn’t register.

      Vote -1 Vote +1

      • Guertez says:
        FanGraphs Supporting Member

        Also a great way to explain the basics of the calculations using one game worth of data. Makes the formulas a little easier to understand as well. Great piece.

        Vote -1 Vote +1

    • Lanidrac says:

      He should’ve been working on a perfect game, though. The only reason Bonifacio (sp?) got hit was because he moved into the pitch trying to bunt.

      Vote -1 Vote +1

  2. jrogers says:

    Did Stanton have one inside-the-park HR (that blue dot on the warning track in left)?

    Vote -1 Vote +1

    • dkmin says:

      Maybe it bounced off Canseco’s head.

      Vote -1 Vote +1

      • JDanger says:

        You guys don’t remember that game last year? It was on all the highlights. Out of nowhere, Jose Canseco emerged from the stands and ran full sprint towards the fence in left-center field and amazingly the ball hit him in the head! Again! Then the ball bounced over the fence and everyone was all, “here we go again” and then the camera zoomed in on Canseco and he just shrugged his shoulders and said, “what did I do this time?” and the whole stadium laughed so hard and then the theme music played. I can’t believe you guys don’t remember that game, it was claaaasic.

        +10 Vote -1 Vote +1

      • Matt says:

        Then he went on Twitter and shamelessly begged his ex to take him back.

        Vote -1 Vote +1

  3. sprot says:

    This is a new low.

    Vote -1 Vote +1

  4. adohaj says:

    I wonder why he hit almost all of his homeruns to the northern part of the stadium. Do you have data on the average wind direction in Florida?

    Vote -1 Vote +1

  5. Sean says:

    What’s the point of comparing games within a 3-game sample size?

    Vote -1 Vote +1

  6. steex says:

    If Miami’s park really does suppress offense as it appears it might, Mark Buehrle may really come to love his new home compared to his old one.

    Vote -1 Vote +1

  7. Ricky says:

    All this article did was show me a neat graphic of just how huge miamis field is

    Vote -1 Vote +1

  8. billy says:

    A one game sample for a regression-based ERA estimator. This is fucking pathetic.

    Vote -1 Vote +1

  9. Jason says:

    ….the complete absurdity of this column brings back fond memories of the infamous official position on UZR article.

    Vote -1 Vote +1

  10. Jason says:

    Bradley,

    You note that many of the balls in play could very well have been hits and suggest that sampling might play a role (!). However, you fail to follow this to its logical conclusion that sampling might play a role in everything about this start (BECAUSE IT IS ONE START AND YOU HAVE NO DATA!!!!). If you had considered the full implications of your insight (!) about some balls in play being hits on another day you surely would not have written this column.

    For example, perhaps on another day some of those hitters where he induced weak contact would have struck out. …imagine what three more strike outs would do to his FIP in about 25 batters faced….

    The best way to describe a single start is to actually watch it and see how the guy pitched and how the batters reacted.


    Vote -1 Vote +1

    • CJ says:

      This article has very little to do with saying “omg, Lohse looks good based on FIP, but I’m going to be cutting edge and say he’s bound to regress”.

      It’s saying “uh, yup, FIP’s broke until about the start of May, you guys”.

      Vote -1 Vote +1

  11. Lanidrac says:

    You need to realize that looking at Lohse’s career ratios is nearly meaningless in his case. Dave Duncan turned around his career in 2008, and then he was pitching hurt for most of 2009 and 2010. You should really just be looking at 2008 and 2011 to see what kind of continually underrated pitcher this guy truly is. For example, his expected HR/9 rate should be a lot closer to the 0.81 and 0.76 marks he posted in 2008 and 2011 rather than his 1.10 career rate.

    Vote -1 Vote +1

    • Ender says:

      True, if you take out all of his bad years and just include the (somewhat luck induced) good years he is a really good pitcher. I imagine the improvements in his BB rate are for real so he is a tiny bit better than his career stats but not by much. Most of this ‘reinventing’ was playing in a pitchers park in the NL with a good defense, not much that he actually changed.

      Vote -1 Vote +1

      • Lanidrac says:

        Like I said, those bad years were only because of injury.

        You underestimate how much of an influence Dave Duncan can have on a pitcher’s career. Busch Stadium III is not a pitcher’s park. It’s a neutral park that leans slightly towards pitching, plus the Cardinals’ infield defense was pretty bad last year.

        Even if it did have a lot to do with it, he’s still pitching in the NL and the same ballpark and with improved infield defense, so he should still have another good year.

        Vote -1 Vote +1

  12. Eron says:

    I think this shows how Matheny is getting better results with Lohse than LaRussa did, he seemed more relaxed as did the rest of the ballclub. Miami on the other hand looked tight,made some mental mistakes, and overall looked like a reflection of their manager (a lot of bluster, but not much substance)

    Vote -1 Vote +1

  13. Bob says:

    Yawn.

    FIP, and all its myriad variations are based upon several false premises–but to cut to the most basic, which FIP does NOT acknowledge: Across MLB, every year without exception, pitchers ahead in the count permit lower BABIP than pitchers even in the count, who in turn permit lower BABIP than pitchers behind in the count.

    But. But isn’t just the BABIP. It’s the percentage of doubles and triples among balls-in-play, too.

    Think on it. What we all know intuitively as a truism…is actually true. And ignored by SIERA, et.al..

    Specifically regarding Lohse: If memory serves, he had the *highest* percentage of first-pitch strikes in the N.L. last year.

    It really ain’t that complicated. Watch a tape of the game. See how often *induced* weak contact happened (something that doesn’t exist within the pristine confines of theory-based, FIP, xFIP, and SIERA premises).

    Of course it wasn’t a “real & genuine” outstanding performance on the bump. Stanton should’ve had at least one tater. But all in all, Lohse — via his control of the count — pitched the kind of solid, stolid game that winds up in a win 80-90% of the time (90% with an offense like StL has this year).

    Good enough for me. ;)

    Vote -1 Vote +1

    • y says:

      I just don’t get why people who obviously refuse to see the merit in advanced stats continue to come on this site and complain about how their favorite player is made to look bad because of them. Do you really need attention *that* badly?

      All that “induced weak contact” has helped Lohse beat his FIP/xFIP over his career, just like Matt Cain, right?

      Vote -1 Vote +1

  14. Bob says:

    I do need attention, and quite badly.

    The advanced stats, so called, like FIP, xFIP, and Siera Madre Treasure or whatever it is, are mere data points to be considered when evaluating pitching performances — be it one game (which was all I was talking about, when it comes to Lohse), one season, or an entire career.

    Again, every year without exception, MLB pitchers as a whole permit lower BABIP (and, as a percentage, fewer doubles and triples on balls in play) when ahead in the count than even in the count. And lower BABIP when even in the count than when behind.

    Since this is true, every single year, of pitchers as a group, it stands to reason that some pitchers are better able to exploit this advantage than others — whether for a single batter, or a single game, or a season, etc..

    In short, FIP theory presumes that getting strike three on a batter is entirely based upon a pitcher’s skill…but simultaneously presumes that a pitcher being ahead, or even, or behind in the count when a ball is put in play has no bearing whatsoever on the outcome of the play. And the latter is provably false.

    Vote -1 Vote +1

  15. Lex Logan says:

    Wouldn’t it make sense to regress (pad) FIP’s baseline with previous year data until there’s enough current year data to make it useful? I mean, it makes sense to report BA, SLG, ERA, etc. based only on this year’s data, but it’s silly to have a baseline constructed from three game’s worth of data.

    Vote -1 Vote +1