Jump to content

Providing independent coverage of the Minnesota Twins.
Subscribe to Twins Daily Email

The Forums

Article: Game Thread: Twins @ Rays, 4/20@6:10pm CT

Minnesota Twins Talk Today, 07:43 AM
Things will get back to normal tonight when the Twins open up a 3-game series in Tampa. By normal, I mean just a normal series in a norma...

Article: Puerto Rican Dream

Minnesota Twins Talk Today, 07:25 AM
They say you should write your dreams down right when you wake up in order to remember them clearly. After returning from a dreamlike day...

5th Starter

Minnesota Twins Talk Today, 07:23 AM
Was it Mejia and not Hughes in town to start today? He was scheduled for Rochester who went with a bullpen game today instead.

Nick Gordon - 2018

Adopt A Prospect 2018 Today, 06:27 AM
  Nick Gordon is a 22-year old lefty-hitting shortstop drafted fifth overall in the first round of the 2014 June amateur draft out...

Article: Twins Minor League Report (4/19): Jimmy Kerrigan...

Twins Minor League Talk Today, 06:06 AM
The Twins were off on Thursday and so were half the Minor League affiliates. The Cedar Rapids Kernels were rained out for the sixth strai...

Photo

Free Agent Starters and Rotation Candidates: By The Numbers (Part I)

Posted by Greg Logan , 28 January 2018 · 1,168 views

Free Agent Starters and Rotation Candidates: By The Numbers (Part I) As we near the two week mark before pitchers and catchers report in Fort Myers, we're starting to see signs that this painfully frigid free agent market may be finally thawing. The Brewers staked their claim on a crowded NL Central with two big outfield acquisitions last week, and rumors abound that Yu Darvish is closing in on a decision that might open the free agent starter floodgates.

While we wait to hear whether Darvish picks the Twins or sends the front office scrambling for Plan B, let's take a look at how the top four free agent starters – Darvish, Jake Arrieta, Lance Lynn and Alex Cobb – stack up against the existing Twins rotation candidates by the numbers. Today we'll start with rate stats, and I'll follow up with a "Part II" that takes a deeper look at the major WAR and projection models.

Posted Image

Posted Image
Posted Image

I imagine every Twins Daily reader is familiar with the drawbacks of both ERA and FIP. ERA holds the pitcher completely responsible for every ball in play, ignoring defense, ballpark factors and dumb luck. FIP clears the pitcher of any responsibility on balls in play that don't leave the ballpark, ignoring quality of contact on those balls in play.

Enter this author's new favorite pitching stat: Statcast's new xwOBA metric. You can find a detailed description at the previous link, but you could say that xwOBA takes FIP to the next level: maintaining the pitcher's responsibility for strikeouts and walks while also giving pitchers due credit (or penalty) for their quality of contact beyond just home runs.

Ervin Santana is a great case study here. Erv's 2017 ERA was stellar, but his FIP suggests that it was helped out quite a bit by some combination of defense, ballpark factors and luck. xwOBA helps us cut through the noise here by showing that Santana's overall production (.292 vs. .320 lg avg) was more in line with his ERA (3.28 vs 4.49) than his FIP (roughly lg avg).

Yu Darvish's numbers, especially in the second half, tell a similar story. xwOBA suggests that he was far and away the best of the available free agent starters in spite of his inflated ERA, and that his second half was terrific despite some of the traditional results suggesting he faded. Perhaps this, and his strong early starts in the playoffs, suggest that his World Series collapse was in fact the result of pitch tipping rather than an overall fade in production.

Arrieta's second half may be even more interesting. His ERA drops substantially in the second half, and while his FIP suggests that it may have been more luck than production, his xwOBA reinforces that he was indeed generating significantly better quality of contact to go with his continued strong strikeout and control rates.

With Lance Lynn and Alex Cobb we can say with some certainty that either would have slotted well ahead of any Twins pitchers not named Santana or Berrios in 2017, but let's take a closer look at the numbers. Nick recently applauded Lynn on Twitter for his consistency and the numbers back him up here, particularly looking at xwOBA which suggests that an inflated 2017 FIP may not be that concerning. With Cobb, many have pointed to his hot second half as a positive, but his FIP remained essentially the same and his xwOBA actually regressed in the second half, suggesting his second half surge may have had a fair amount of luck attached to it.

What else jumps out at you in these numbers? Mejia's second half? May's strong 2015 xwOBA as a starter?

  • Oldgoat_MN, markos and Tom Froemming like this



Is there any systematic evidence that long term, certain pitchers generate worse contact than others, other than in terms of the crude categofies of GB, LD and FB? In other words, is there any systematic evidence that xwOBA ads anything to FIP (or SIERA which I think already takes into account GB%) other than short-term randomness? I'm open to the possibility, but I haven't seen it.Certainly certain hitters generate better launch angles and exit velocities on FB/LD, but that doesn't necessarily mean that certain pitchers do.

 

That being said, this is an interesting analysis.

    • Greg Logan likes this

 

Is there any systematic evidence that long term, certain pitchers generate worse contact than others, other than in terms of the crude categofies of GB, LD and FB? In other words, is there any systematic evidence that xwOBA ads anything to FIP (or SIERA which I think already takes into account GB%) other than short-term randomness? I'm open to the possibility, but I haven't seen it.Certainly certain hitters generate better launch angles and exit velocities on FB/LD, but that doesn't necessarily mean that certain pitchers do.

 

That being said, this is an interesting analysis.

https://www.fangraph...-pitcher-skill/

 

This suggests that xwOBA is not any more predictive than FIP. 

 

All of which is to say, it’s hard to look at how a pitcher did against contact in one year and project for the next. We could be dealing with a change in talent level and we could be dealing with some luck or the individual sample sizes just might not be big enough.

 

 

    • Greg Logan likes this

 

https://www.fangraph...-pitcher-skill/

 

This suggests that xwOBA is not any more predictive than FIP. 

I keep waiting to see a really in-depth analysis of the "stickiness" of contact quality. Craig notes there that there are some sample size concerns, and I'd agree. We still only have three seasons of Statcast data to work with, so hopefully as we get more data we'll be able to see a bigger n on an analysis like Craig's. I'm curious why he doesn't look at 2015-to-2016 in addition to 2016-to-2017 since the data's there.

Regardless of the projectability, I think xwoBA shows us a much more objective view of a pitcher's performance independent of defense than FIP does, and certainly more than ERA does. 

Thanks for sharing the post. It is nice to have someone compile all the data. Personally, I agree with your statement that Lynn and Cobb probably slot better behind Santana/Berrios rather than in front of them. Also, I'm pretty bullish on Mejia, so it is nice to see number that back that belief up.

Photo
ToddlerHarmon
Feb 05 2018 05:55 AM
What strikes me is that Arrieta is only disappointing compared to the high bar he set. These numbers show him to be the cream of this list alongside Yu