Ricky Romero’s WAR, FIP and the Pitfalls of Selective Saber Stats

 

 

A plethora of sabermetric related articles around the web, brought on by Moneyball’s release and on this website by Callum’s uneasiness with Sabermetric statistics. To set the record straight, since I’ve yet to pen an article on the whole Sabermetrics vs regular stats argument, I’m firmly in the corner of sabermetrics, with one caveat, that is the saber stats themsevles being used within a proper context.

The problem is, proper context is rarely used. And one mathematically based measure or another never tells the entire story.

These days, when Saber stats are thrown about, the source of data is usually based upon a quick look at FanGraphs. With this in mind, fangraphs versions of WAR & FIP are used within this article.

Someone may quickly point to a players WAR and move on. This can be problematic. One really needs to look at six or seven different stats to get a better idea in regards to whats going on.

Ricky Romero‘s FIP & WAR (which is based entirely on FIP) is a perfect example of the need to look at multiple statistics. From purely an FIP standpoint, Romero has regressed dramatically from his 2010 season, to the tune of around 0.50 FIP runs, entirely on account of increased HR/9, as his K – BB rate has ever so slightly improved.

(stats from Romero’s FanGraphs page)


Season Record ERA FIP WAR IP HR K BB
2010 14-9 3.73 3.64 4 210 15 174 82
2011 15-10 2.98 4.1 2.9 217.1 24 173 78

This accounts for Romero’s fan graphs WAR witnessing a dramatic drop from 4.0 in 2010 to 2.9 this season

Fangraphs FIP:

 

= (13HR + (3BB+HBP-IBB) – 2K) / IP

 

 

However, this doesn’t tell the entire story. Additional internal statistics add additional story lines to Romero’s 2011 season:

 

#1  LOB (Left on Base or Strand rate)

An improved strand rate is one of the key drivers in Ricky’s vastly lower 2011 ERA (2.98 vs last years 3.73). Last season his LOB was slightly below MLB avg at 70.2%. This season he’s near the top of the league at an 80.0% LOB rate. It would be easy to dismiss LOB gains as unsustainable but if Ricky has really turned into an ‘elite’ pitcher, then the 80% rate, while a little high, isn’t astronomical. Pitchers such as Cliff Lee, Halladay, Kershaw, Lincecum, Weaver, Lester, etc are typically around the 76% – 79% ratio. Above league average but then again, these are above average pitchers.

 

#2 Pitch Type

Last season Ricky threw around 60% combined FB + Cutters. This season the rate has jumped to 70%. His average velocity has also creeped up over 1 MPH. Whether this ties into FIP is debatable. However it does shed some light on Ricky becoming a different pitcher in 2011.

# 3 LD% (line drive percentage)

Previous LD% were right around league average at  18-19%. This season Romero has one of baseballs best LD% rates at 14.1%. In turn, his FlyBall% has risen, while his GroundBall rate has remained constant. Essentially he’s replaced Linedrives with flyballs, a worthy trade-off for any pitcher. Romero has allowed 9 more HR than last season but I would argue that the lower LD% equates to lesser contact, which in turn helps aid #1 on our list, LOB%.

 

#4 BABIP

As expected after looking at the LD% drop, Romero’s 2011 BABIP has shrunk over .40 basis points, down to .244 vs .289 last season.

 

 WTF?

 

Has this exercise answered anything? Well yes and no. The Jury is still out on whether the 2011 version of Ricky Romero is improved (as ERA and, to be frank, regular Jays fans would suggest), has declined (as indicated by WAR or FIP), or has remained constant. However, one thing is for sure. Citing specific sabermetrical stats in a manner such as “he’s got a ‘insert number here’ WAR” is far from the entire story. It’s really a chapter at best.

has written for mopupduty.com since 2006. Follow Matthias on Twitter, Facebook and Google +

  • gabriel

    It seems like the two ways pitchers can outperform their advanced metrics like FIP, xFIP, SIERA is to have a below-average HR/FB rate, like Matt Cain; or to have a below-average BABIP. Romero has shown no sign of being able to depress his HR rate, but his BABIPs over his three season are .333, .289, .244, which is a pretty great progression, and hopefully he can prove that he does have this particular skill in the coming years.

    • http://mopupduty.com Matthias

      Can a pitcher have a high level FIP without a below avg HR/FB? I’m sure there are a few instances here and there but for the most part they go hand in hand.

      • gabriel

        Most of the top pitchers have HR/FB rates fluctuating around average. Halladay’s career rate is pretty much average, for instance. There’s some skill involved as Matt Cain demonstrates, but it’s not common.

  • http://www.chriskol.com Chris K

    I haven’t played ball in over a decade and I was certainly no pitcher. Not trying to pick a fight or anything but I’m really interested in knowing the answer to the following:

    #1 With regard to a strand rate, is that really an important stat? It seems like one of those things like RBI that shouldn’t really change the situation. I mean, you’ve still got to get the guy at the plate out and unless you can prove significantly that you can induce double plays, wouldn’t the stat be meaningless?

    #2 Does this actually matter in the grand scheme?

    #3 Does the 5% LD drop really negate the HR increase?

    #4 I’m not really sure I have anything to say on this point.

    Final: I don’t know…it seems to me like the sabermetrics tell a better story on his abilities and development than points you’ve posted.

    Looking forward to your response and keep up the great work on the blog!

    • http://mopupduty.com/index.php/about-mop-up-duty/authors_matthias_koster/ Matthias Koster

      No problems Chris, thanks for the comments.

      “Final: I don’t know…it seems to me like the sabermetrics tell a better story on his abilities and development than points you’ve posted.”

      If one looked purely at the Saber aspect (FIP, WAR), they would conclude that Romero has allowed more HR, and thus is having a poorer season than 2010 and leave it at that.

      That’s too simplistic. The other points showcase part of the story (at least that was what I was aiming for)

      1. Stranding 80% of runners on base vs 70% has a huge effect on bottom line runs scored. From a Saber perspective, where runs are treated as parts of equations, LOB doen’t matter. From an ERA and actual runs scored perspective, it does. LOB will fluctuate but as I showed in the article, if Romero can move up to ‘stud’ status, his LOB and thus real runs scored should be lower than previous seasons.

      2. Do pitch types matter in the grand scheme of things? It shows that Ricky is changing his approach to a degree. This is important data IMO and something missed by simply looking at FIP.

      3. From an expected hits degree, lowering linedrives is huge. Batting average on linedrives is about .740 across the MLB. Cutting down the % of LD and replacing them with FB is a plus, even if a few more HR are leaving the yard. Could his change in pitch type (point 3) be one of the reasons for the LD decline?

      ’2.9 WAR’ doesn’t explain any of these above points.

  • Stan

    I certainly agree with you that people tend to use stats without putting them in the proper context all the time. However, I wouldn’t worry about Ricky’s fWAR too much since it’s based entirely on FIP as you’ve said. The problem with FIP as FanGraphs says itself (posted from FanGraphs glossary below)is that it does a better job of predicting the future than measuring the present. Therefore to me, fWAR is largely irrelevant for measuring performance for the current season. To me Baseball Reference’s WAR is a far better indicator of how a pitcher performs. They have Ricky ranked tied for 5th with Shields in the AL and tied for 8th overall.

    - FIP does a better job of predicting the future than of measuring the present, meaning there can be a lot of fluctuation in smaller samples. It’s not a great choice when trying to describe how a pitcher performed during a single game.

    - In case you’re curious, here’s the formula for FIP: ((13*HR)+(3*(BB+HBP-IBB))-(2*K))/IP + constant

    The constant is solely to bring FIP onto an ERA scale, and is generally around 3.20 (but it can be derived by finding the league-average FIP and subtracting that result from league-average ERA).

    • http://mopupduty.com matthias

      Great comment Stan, thanks for posting it.

      Baseball Prospectus also gives Romero proper 2011 WAR credit, as his ’11 score of 3.0 is greater than his 2010 WAR of 2.9.

      Readers (and especially other bloggers) need to shop around for their stats. The usual Fan Graphs cut & paste model leaves much to be desired.