Kyle Gibson, Chase Anderson, and ERA vs DIPS

[Laika]

Kyle Gibson, Chase Anderson, and FIP Dogmatism

 

Long gone is the initial, romantic promise of defense independent pitching statistics (DIPS); the notion that we can cleanly separate pitcher skill from the messy results on the field is dead.

 

Sure, FIP correlates better to future ERA than ERA does, and every subsequent evolution of DIPS (xFIP, SIERA, DRA, and whatever comes next) purports to offer additional accuracy or applications, but by now we have seen enough pitchers for which DIPS is more misleading than helpful that it makes sense within any individual pitcher analysis to at least turn our minds to whether or not the truth lies within the DIPS.

 

Blue Jays fans will always fondly remember Marco Estrada’s prime, when his unique pitching style lead to such an inordinate number of easy flyouts and made it really seem that he was controlling batter contact quality in a way that FIP could not appreciate. Ask Baseball Reference, which uses runs surrendered in its pitcher WAR calculation, how many wins Estrada was worth in 2015 and 2016 and it will say 7.5. Ask Fangraphs the same question and the answer is 4.5.

 

The price of a win in baseball has been estimated at close to $10M; the disagreement over Marco Estrada’s true talent in 2015 and 2016 was a $30M question.

 

Kyle Gibson was another $30M question. He’s a Ranger now for that amount of guaranteed money and if the baseball media is real news Toronto was interested and could probably have had him for a little bit more than that.

 

Gibson is an interesting projection. His career ERA is 0.23 runs higher than his FIP. He was not all that good by any measure in 2016 and 2017, with ERAs over 5.00 and FIPs of 4.70 and 4.85. Then in 2018 a slight uptick in velocity helped him drop the ERA all the way down to 3.62 and the FIP sank with it to 4.13. 2019 was a puzzler: his strikeout rate and velocity stayed up, his FIP stayed near his 2018 mark, but his ERA shot back up to nearly 5.00 again.

 

The 2020 Steamer projection for Gibson is 3.0 fWAR. Ask Fangraphs how good he has been in the last two years and it says 2.6 wins in each season. Ask about his career and it says 13 WAR.

 

Baseball Reference thinks Gibson has been much worse - a 9.6 WAR pitcher in his career, and it thinks three of his six full season have been so bad that he was not worth even one WAR. With 1087 career innings pitched, at what point do we start looking at ERA and not FIP? Someone knows this answer. Do we take the increased strikeout rate as a new beginning and throw everything before 2018 out the window?

 

The pitcher that the Blue Jays did land this offseason has essentially the opposite trend. Chase Anderson, a flyball pitcher, has a 3.94 ERA in his career but a 4.54 FIP. Ask Fangraphs and he’s been worth 7.5 WAR in his 857 career innings; ask Baseball Reference and the answer is 9.9 wins.

 

The third big baseball site, Baseball Prospectus, has the most modern and advanced ERA estimator in its newfangled DRA. BP says Chase Anderson was worth 1.4 WARP in 2019 with a 4.84 DRA, while Gibson was worth 0.3 WARP with a chunky 5.60 DRA. In 2018 DRA thinks Gibson was worth 2.5 WARP with a 4.21 skill level and Anderson was worth negative 0.4 WARP with a 5.52 DRA (Chase Anderson had a 3.93 ERA in 158 innings that year so this disparity is immense).

 

Steamer thinks Chase Anderson will have an ERA and FIP of 5.49 in 2020 – a seemingly absurd position for the stupid computer to take, given Anderson’s career marks.

 

All of this is supposed to confuse you. Pitcher value analysis and projection using publicly available statistics and ERA estimators is more confusing than ever.

 

Don’t make the mistake of eyeballing Kyle Gibson’s FIP or his Steamer projection and thinking that his talent level is easily ascertainable. The teams aren’t doing that because they certainly have more sophisticated information at their disposal. The popularity of Fangraphs (which is great) and their decision to use FIP as the single input for pitching WAR should not give us tunnel vision.

 

As simple-minded members of the public we are forced to concede that we can really have no idea how good players like this are; we have no idea how Anderson or Gibson will perform in 2020 because we don’t even have much of a clue how good they actually were in 2019!

 

Kyle Gibson’s realistic outcome range goes from a productive and resurgent veteran who will fit in great with Lance Lynn and Mike Minor to the frustratingly ineffective and hittable Kyle from 2016, 2017, and 2019. It’s fair to find it a bit more likely that he will fall towards the positive end of that range given the increase in velocity and strikeouts (keep in mind however that the strikeout rate league-wide has gone up in this timeframe.)

 

Chase Anderson’s realistic outcome range is just as wide. Based on runs prevented Chase has had the best single season and the best career between the two of them and it’s not even close. He has exhibited a positive velocity trend too.

 

The Anderson acquisition inspired little fanfare while Kyle Gibson was a common free agent target for Blue Jays fans. I’ll be watching these two enigmatic pitchers closely in 2020.

 

This article was meant to meander and reach no conclusion, just like my mind when I try to figure out how good these two will be.

[Spanky99]

4:20!

[BTS]

This was both a good read and a shameless play at securing another poster of the year award.

[Laika]

Personally, I would take both. Unless we plan to spend the 90M in available payroll on non-Gibson pitchers.

 

But also, you have to take into account the difference in roles last year. Anderson was in the NL and also protected heavily by MIL. He only had 15 innings of third time through. Gibson had 35 IP of that variety and a 9.42 ERA! Also Anderson buffed his stats a bit with 10 pure relief innings.

 

As far as contact quality goes, it appears Anderson has the advantage with a lower exit velocity against and xwoba. Anderson also had a better barrel %. But again, those numbers are affected a bit by his usage probably.

 

Then there are other variables like pitch mix. Whereas Gibson has a s*** fastball, he has a great slider that he could use more. Anderson has a better fastball but he's lost the curve he had a couple years ago. He says he's working on it so you have to take that into account. Perhaps Pete Walker can work his magic /s.

 

And yet there's even more subterfuge to navigate considering Gibson had E. coli last year and lost like 20 lbs. And faded hard down the stretch.

Anderson has the best year among both pitchers, that being 2017. His hard hit% was top 5% in the league back then. Again, with the curveball. So he has a higher demonstrated ceiling.

 

E. Coli in the spring but then ulcerative colitis later in the year. The latter can be a chronic issue and teams might actually factor it into his projected long term value!

[EZe]

E. Coli in the spring but then ulcerative colitis later in the year. The latter can be a chronic issue and teams might actually factor it into his projected long term value!

 

I have a family member that was recently diagnosed with Crohn's similar to ulcerative colitis. They've been bed ridden, in and out of hospital and downright debilitated. As you indicated its chronic and likely comes back in flare ups. I can't imagine a high level athlete playing through that. Also potentially the E. coli was misdiagnosed and an original flare up. The contract was lower than I expected, this factored in.

Edited November 30, 2019 by EZe

[TheHurl]

What about TIPS what does it predict?

 

TIPS = 6.5*O-Looking(PitchF/x)% – 9.75*SwStr% – 4.8*Foul% + C

[KevinGregg]

What about TIPS what does it predict?

 

TIPS = 6.5*O-Looking(PitchF/x)% – 9.75*SwStr% – 4.8*Foul% + C

 

What's C? Just the normal fip constant?

[Terminator]

Good post. Anderson does look like a FIP beater and it's hard to believe Steamer's projections for Gibson when they are predicting a career year at age 32. They are probably closer in talent than most think. I think both players and their respective transactions were good moves though.

[Key22]

Do the statistics also account for quality of batters/teams faced?

 

Back in the ole days you could pretty much look at WHIP and say hey the guy is probably pretty good if his WHIP is around 1 to 1.2. Pretty crappy if it is 1.4

[Laika]

Do the statistics also account for quality of batters/teams faced?

 

Back in the ole days you could pretty much look at WHIP and say hey the guy is probably pretty good if his WHIP is around 1 to 1.2. Pretty crappy if it is 1.4

 

- ERA and FIP do not adjust for those things

- ERA and FIP minus (ERA- or FIP-) do adjust for league and park factors. These are scaled to 100, so 90 is good, 125, is awful, etc. So, for example if a 4.40 ERA in the AL East equals a 4.00 in the NL East, two pitchers with those ERAs will have the same ERA-.

 

Chase Anderson career

ERA- 94; FIP- 110

 

Kyle Gibson

106; 102

 

- fWAR and bWAR do adjust for league and park (but differently, I think).

- The new DRA on BP adjusts for a lot of things, I think.

 

WHIP is interesting historically because it's kind of like the first talent descriptor outside of ERA. People used to realize that ERA was wildly random and they thought WHIP was more stable and therefore more indicative of true talent or future ERA. It's pretty crude though and there are better tools to use than WHIP in any instance. It's pretty mean to groundball pitchers who can give up lots of singles and sometimes walks but can still be good, get a lot of double plays and avoid home runs, etc.

[FrozenRopes]

- ERA and FIP do not adjust for those things

- ERA and FIP minus (ERA- or FIP-) do adjust for league and park factors. These are scaled to 100, so 90 is good, 125, is awful, etc. So, for example if a 4.40 ERA in the AL East equals a 4.00 in the NL East, two pitchers with those ERAs will have the same ERA-.

 

Chase Anderson career

ERA- 94; FIP- 110

 

Kyle Gibson

106; 102

 

- fWAR and bWAR do adjust for league and park (but differently, I think).

- The new DRA on BP adjusts for a lot of things, I think.

 

WHIP is interesting historically because it's kind of like the first talent descriptor outside of ERA. People used to realize that ERA was wildly random and they thought WHIP was more stable and therefore more indicative of true talent or future ERA. It's pretty crude though and there are better tools to use than WHIP in any instance. It's pretty mean to groundball pitchers who can give up lots of singles and sometimes walks but can still be good, get a lot of double plays and avoid home runs, etc.

 

Who do you work for?

[TheHurl]

Who do you work for?

 

BJMB and Radio Scouts

[TheHurl]

What's C? Just the normal fip constant?

 

Sorry didn't see this. C is a constant that changes from year to year to adjust to the ERA scale (to make an average TIPS = average ERA). For 2013 this constant was 2.57.

 

This was of course Chris Carruthers ERA predictor from 6 years ago.

 

http://www.breakingblue.ca/2013/11/21/tips-a-new-era-estimator/