Clayton Kershaw and the Others
I understand why the BBWAA announces finalists for awards before it announces the winners — they’re trying to build some suspense, some anticipation, and, okay, that’s fine, even if the finalists don’t tell us much we couldn’t predict. For example, we’ve long known the AL Cy Young was going to come down to Felix Hernandez vs. Corey Kluber. All the non-contenders have been officially eliminated, which does nothing. It’s even funnier when you get a race that isn’t a race at all, like Jose Abreu and the AL Rookie of the Year. Or, say, Clayton Kershaw and the NL Cy Young. According to the BBWAA, Kershaw is a finalist, along with Johnny Cueto and Adam Wainwright. The ESPN forecast is somewhat torn between Felix and Kluber, in one league. In the other league, it’s less torn:
Unanimous, in other words. And this is the way things are likely to play out. Clayton Kershaw ought to get every single first-place vote, because of the gap between himself and the others. You can eyeball the gap if you want, but we can also put some numbers to the idea. What would it have taken for Kershaw to come away resembling Cueto or Wainwright?
It’s already been determined, unofficially, that Chris Sale isn’t winning the AL Cy Young because he spent time on the DL. He wasn’t sufficiently better than Felix or Kluber to make up for the missing playing time. That’s not an issue with Kershaw, who was also hurt, and who finished with fewer innings than Cueto and Wainwright. Giving the award to Kershaw will be slightly uncharacteristic, in that the BBWAA has historically preferred durability, but the difference between his performance and the other performances is so large that there’s not even really a question.
Here’s a table! You understand it.
| Name | IP | ERA- | FIP- | wOBA |
|---|---|---|---|---|
| Clayton Kershaw | 198.3 | 50 | 51 | 0.233 |
| Johnny Cueto | 243.7 | 61 | 88 | 0.259 |
| Adam Wainwright | 227.0 | 66 | 79 | 0.259 |
By ERA-, you’re looking at a gap of 11 points between Kershaw and Cueto. By FIP-, it’s 28 points, between Kershaw and Wainwright. By wOBA allowed, it’s 26 points between Kershaw and the other two guys. So, consider that Kershaw threw fewer innings than his competition. What would it take to reduce Kershaw’s numbers to the other numbers, over equivalent innings?
Let’s compare Kershaw and Cueto first. Cueto threw 45.1 more innings than Kershaw did. So let’s give Kershaw a hypothetical 45.1 more innings. What would his numbers need to look like over that stretch to equal Cueto’s?
We start with a 109 ERA-, more than double Kershaw’s actual ERA-. If Kershaw posted a 109 ERA- over 45.1 more innings, he would’ve equaled Cueto. That works out to a 4.37 ERA. That’s not completely nuts, but we’re also not finished. How about FIP-? Now you’re looking at a 250 FIP- over 45.1 more innings, working out to a 9.82 FIP. As for wOBA allowed, you’d be asking for Kershaw to allow a .352 wOBA over 212 more batters. That’s asking Kershaw to allow the average hitter to hit like 2014 Adrian Gonzalez.
Okay! Now Kershaw and Wainwright. Wainwright threw 28.2 more innings than Kershaw did, so, let’s give Kershaw a hypothetical 28.2 more innings. What would his numbers need to look like over that stretch to equal Wainwright’s?
We start with a 177 ERA-, more than triple Kershaw’s actual ERA-. That works out to a 6.59 ERA. And as for FIP-? You’re looking at a 273 FIP-, working out to a 10.28 FIP. For wOBA allowed, you’d be asking for Kershaw to allow a .389 wOBA over 149 more batters. That’s asking Kershaw to allow the average hitter to hit like 2014 Edwin Encarnacion.
That’s how big the gap was between Clayton Kershaw and the next-best starting pitchers in his league, according to the BBWAA. In order to make up the difference, given Kershaw’s lower innings total, you’d be looking for him to pitch somewhere between very poorly and impossibly poorly. By the fielding-independent metrics, the difference between Kershaw and Cueto or Wainwright was a few dozen innings of pitching like a position player. That’s why Kershaw should win the 2014 NL Cy Young with every first-place vote possible.
What’s funny is that, a year ago, Kershaw won the NL Cy Young, with all but one first-place vote. Kershaw did that a year ago, and then he got better. He lopped a chunk off his walk rate. He increased his strikeouts by 25%. He generated more grounders. Kershaw last season posted a 2.88 xFIP, and then he dropped that this season to a 2.08 xFIP, and Kershaw, historically, has pitched better than his xFIP. There are people trying to manufacture a debate about Clayton Kershaw vs. Madison Bumgarner. No. It’s insane. That’s insane. A handful of October appearances doesn’t erase what Kershaw has been able to accomplish, in 2014 and beyond 2014.
Consider the season that just happened. Going all the way back to 1900 for some reason, there are 8,489 qualified pitcher seasons. Kershaw’s 2014 just ranked 31st by ERA-, 10th by FIP-, and 10th by a blend of ERA- and FIP-. In other words, by ERA-, Kershaw ranked in the top 0.4%. By FIP-, he ranked in the top 0.12%, and the same could be said of his blended stats. Names of the pitchers who had seasons that beat Kershaw’s blended ERA/FIP:
Make no mistake: what Kershaw did in 2014 was of historical significance. It was a season worthy of a Pedro comp, even given the short-term injury.
And for purposes of evaluating Kershaw within his own time, let’s look at all the pitchers who’ve thrown at least 750 innings over the past decade. So now we’re looking at Kershaw’s career, instead of just his 2014 season. It’s a sample of 146 pitchers. By ERA-, Kershaw ranks first. By FIP-, Kershaw ranks first. By a blended ERA- and FIP-, Kershaw ranks first. He’s eight points ahead of Felix Hernandez. He’s moved 13 points ahead of Justin Verlander. Clayton Kershaw’s the best pitcher of his day, and he’s coming off a single season for the ages. He’s 26.
What the world didn’t need, probably, was another Clayton-Kershaw-appreciation post. But consider that, a year ago, Kershaw looked like quite possibly the best pitcher of his day. Consider that, a year ago, Kershaw was coming off probably the best season of his career. Consider that Clayton Kershaw found a way to improve on Clayton Kershaw. God help the National League if that happens again.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.
So fricking great in the season, so bad in the playoffs. Who saw that coming?
Here’s how I like to think about it, given the passage of time:
Kershaw, season: 32% strikeouts
Kershaw, playoffs: 37% strikeouts
Also:
Kershaw, season: 3.13 xFIP
Kershaw, postseason: 3.21 xFIP
does kershaw have a higher than expected hr/fb ratio in the post-season?
1 disastrous inning and 1 bad inning erased 2 dominant 6 inning starts. If the Dodgers had another reliable reliever outside of Kenley I think they may have removed Kershaw earlier in at least one of the games. What a bummer.
I apologize for my prior post. What I meant to say was that even if you combine the regular season and post-season statistics, Kershaw was still significantly better. And nobody with at least half of the brain power of Johnny Gomes can argue that Bumgarner is equal or better than Kershaw.
Why did you have to throw out the Pedro comp?
Pedro:
Yr/FIP-/ERA-
1999/30/42
2000/46/35
His 2000 ERA was lower (marginally) in the middle of the steroid era. Kershaw is great but he hasn’t sniffed 99/00 Pedro.
Right, he didn’t match peak-peak Pedro, but he perfectly matched regular-peak Pedro, in 2002-2003.
Oh, Pedro Martinez. The definition of “absurdly amazing”
Fair enough
The sad thing is, Pedro was a dodger before they let him go… I still have two Pedro signed balls from when he was a dodger
I think the Pedro comp is natural because looking up insane 2014 Kershaw stats leads to a lot of result of “best ever aside from Pedro”
I just wanted to say. I always find it ominous that Walter Johnson made 666 starts.
The really stupid part of that Kershaw or Bumgarner article is they specifically ruled out any consideration of their contracts which is the only thing that might make the question even debatable in the first place. Not to mention that if you’re going to do that why even poll GMs vs. scouts or other FO personnel? What specific insight are they supposed to have once you rule out the whole maximizing performance within a budget equation that’s the core of their job? That could have at least been interesting to test the limits of how much smaller market teams would be willing to invest in any single player even if he’s the most dominant starting pitcher of his generation and still just 26.
Because that’s the point. You may find this hard to believe, but the most recent postseason has sparked a lot of people to legitimately suggest that a team would rather have Bumgarner than Kershaw, just straight up, no contract consideration. ESPN is either in that boat, or they’re at least responding to the fan expression.
Keep an eye on this year’s mlb.com GIBBY’s, the fan-voted yearly awards. Watch Bumgarner win or threaten to win the “Best starting pitcher” category, which is expressly covering only the regular season.
As far as I’ve seen, ESPN (or Jayson Stark to be more specific) is the entire boat on this “debate”. I haven’t seen it posted as a serious question anywhere else.
I highly doubt anyone will see this, but I found myself revisiting this post, and so now we can see how this prediction of mine turned out. Recall that the GIBBY results are determined entirely by fan vote.
Select GIBBY results to follow, name followed by percentage of votes received:
Most Valuable Major Leaguer (regular season only)
1st: Clayton Kershaw, 35%
2nd: Madison Bumgarner, 34.4%
Starting Pitcher of the Year (regular season only)
1st: Clayton Kershaw, 42.5%
2nd: Madison Bumgarner, 38.2%
3rd: Corey Kluber, 7.1%
So either people did not understand that these categories only apply for the regular season, or the postseason performance of each has really impacted how people view each pitcher. Either way, it also shows how people inflate the importance of the postseason, as if the fact that we care more about the results somehow means the results are more meaningful from a player evaluation perspective.
All this is to say that my original point is legitimate. Among casual fans, “Bumgarner or Kershaw” in a vacuum is a legitimate question.
What would Felix or Kluber or Sale have looked like in the NL?
Kershaw had an all-time season.
But could you add another strikeout/9 to the 3 from the AL, assuming they pitched in the NL? (Ervin Santana went from 7 k/9 to 8+ k/9)
How much lower would their bb/9 have been?
Shave 10 runs off Felix and he matches Kershaw’s ERA.
Course, 10 runs is a lot – but add a few innings and remove a few runs and he’s there. At bats vs pitchers should have allowed more innings, strikeouts and fewer runs.
Sale already matched Kershaw’s k/9, essentially.
Add 60 at bats vs pitchers, what happens?
Can we assume 40% k rate? 50%?
50% would be another 30 strikeouts, putting sale at 12.33 k/9 (vs kershaw’s 10.8).
60 at bats vs pitchers would surely mean fewer walks for Sale as well.
Sale had 39 walks or 4 per lineup spot over the season on average (maybe more for designated hitters?).
Remove 5 walks and Sale is at 34 (1.76/9) vs kershaw’s 31 (1.41/9)
Kershaw’s all-time season doesn’t seem that far ahead of the AL Cy Young contenders.
So, you’re saying if you arbitrarily add a huge number of strikeouts to every pitcher in the AL and arbitrarily get rid of a huge number of hits, walks and runs, all of them continue to be not as good as Clayton Kershaw? Fascinating case you’ve made.
not so arbitrary.
I took Sales total batters faced, divided by 9 to get the average at bats per lineup slot, which was somewhere in the 70’s.
I reduced that to 60 to account for pinch hitters.
Sale had 30%+ k% vs non-pitchers.
Why would 40% be a stretch?
Would 50% be a stretch against pitchers?
I don’t know.
NL pitchers struck out 36.4% of the time while batting.
And walked 3% of the time.
40% for Sale against pitchers is not a stretch by any means.
50% might be.
One mistake I made is simply adding the 30 k’s to Sale’s total.
I didn’t subtract strikeouts from Sale vs a lineup slot before adding the additional 30 back in.
Sale was @30% vs non-pitchers, and assuming 40% vs pitchers would only add 6 strikeouts more than he already had.
12 strikeouts more at 50% vs pitchers.
12 more strikeouts = 11.38 k/9
Kershaw vs pitchers in 2014
in 41 at bats (not plate appearances – I didn’t look for sacs, or walks and these are starting pitchers)
3 Hits
16 non-strikeout outs
22 strikeouts
K rate is based on batters faced, though.
I think its safe to say kershaw was over 40% k rate vs. pitchers.
This “analysis” is nonsense.
Oh look, someone else did the research already.
“In moving to the NL, a pitcher’s strikeout rate (K/9) would be expected to rise more than a half-point. That’s huge. Walk rate is less significant at just an 0.05 difference.”
http://www.hardballtimes.com/the-statistical-impact-of-switching-leagues-for-pitchers/
“Overall, a pitcher moving from the AL to the NL would see his ERA drop by about 0.41 points”
That would put Sale at 1.76 ERA, and a k/9 of 11.33, which matches my revised estimate above.
Do you like eating crow?
The username is very fitting for your argument.
You can’t just add pitchers. The AL lineup, by putting their worst defender at DH, opens a spot on the field which may just as easily be filled by a glove-first guy. You can assume Sale would have more strikeouts if he faced pitchers, but you don’t know what strikeouts he got from facing #9 hitters who couldn’t crack an NL lineup.
Plus, it’s all moot because there are park/league-adjusted stats for this and Kershaw blows everyone away in those, as Jeff showed above.
Plus kershaw struck out over 30% of the non-pitchers he faced this year – better than Sale.
Sale was at 30.4% (non pitchers) vs Kershaw’s 31.9% (with pitchers).
Do you doubt that the NL is an easier place to pitch than the AL?
Pitchers struck out 36% of the time.
Non-pitchers were around 20%.
That makes a difference.
FIP isn’t league adjusted.
The AL scored 500 runs more than the NL.
K% was 1% more in the NL than the AL.
Yes but his 3 in 4 years, and barely losing that one year puts him miles and miles ahead of all the others.
I think you could make a reasonable argument that by missing innings Kershaw A)Forced the Dodger to throw a below replacement level Maholm B)IP are undervalued by WAR because it doesn’t factor in saving reliever.
I’m not saying that Kershaw wasn’t more dominant or impressive. But on strict “what happened” value, I think you could argue for Cueto. Though Kershaw gets point for pitching to AJ Ellis’s miserable framing
The composition of the rest of the team’s roster should have no basis whatsoever on individual awards. What someone else did or did not do has zero impact on the amount of value Kershaw added while on the field.
And you’re saying that ~45 IP of “saving reliever” (which isn’t even saving a reliever, because another SP is still taking Kershaw’s place) is worth the colossal difference in value between Kershaw and Cueto. The extra innings are already factored into Cueto’s value. Kershaw did more in less IP. He was amazingly dominant and blew everyone else out of the water. That you have to justify Cueto = Kershaw through bizarre, and arbitrary reasoning is pretty telling.
If a starter goes, say, 43 fewer innings because he’s getting pulled 4 or so outs sooner on average, that costs the bullpen which has to cover those extra outs.
If a starter goes on the DL then you call up another starter or long reliever, and the extra outs are largely covered by 0 WAR pitchers with only slightly more strain on the pen.
Kershaw averaged 7.1 innings per appearance. The relievers needed to get 5 outs on average when he started. That’s not a problem. When he was on the DL he was on the DL, some bullpen strain, but not much, not enough to make up for the extra coverage he gave the rest of the time.
It should be pointed out that Maholm averaged 5.1 innings per start when he filled in for Kershaw (over seven starts), so the impact on the pen was roughly 14 innings, total, spread out over a 162-game season. If you take those 14 innings and add them to Kershaw’s with, say, a replacement level of offensive prevention, Kershaw still comes out far, far above Cueto, which is exactly the sort of scenario this article is discussing.
You can do something similar for Troy Tulowitzki vs. Ian Desmond for the NL Silver Slugger at shortstop. I calculated that Tulo would’ve had to go 0-for-113 in order to get his wRC+ down to Desmond’s. Of course since Tulo only had 375 PAs while Desmond had 648, that means you also have to add 162 PAs of regular old Desmond, but still. Or as another commenter at Bucs Dugout said, you could’ve had 400 PAs of Tulo and 200 PAs of Michael Martinez (0 wRC+) and it would’ve come out better than the silver slugger winner.
Of course Desmond’s Silver Slugger was a really terrible award; Hanley Ramirez and Jhonny Peralta completely blew him away in wRC+, but Desmond had more RBIs. Fortunately no one cares about the Silver Slugger, which is why I’ve had to leave this comment on a post about Clayton Kershaw.
Same thing for Adrian Gonzalez over Rizzo. It seems like Cy Young voting has gotten a lot better lately, but the silver slugger is still totally out there.
Yeah, that’s bad. RBIs RBIs RBIs.
Hope Harvey is back to make a competition of it next year.
The 1-0 gem in 2013 was what mlb needed
Why not show the stats including postseason and regular season play? Here are the ERAs:
Kershaw: 2.14
Bumgarner: 2.60
And some other reference points:
Wainwright: 2.59
Cueto (no postseason to include): 2.25
Yep, Kershaw’s still ahead. Eyeballing it, the same thing would be true if you used FIP or xFIP as well.
I never will understand why analytical baseball people throw out the postseason altogether. Postseason statistics are data just like regular season is. Larger samples are generally better than small ones. Why not just include all the games these guys played?
I was thinking a similar thing. Do the projection systems include any postseason data? They don’t need to be weighted highly, but they are real innings that happened.
In considering Cy Young voting it’s appropriate to throw out postseason stats, because it’s a regular season award and the voting takes place before the postseason.
Sure, but the article is also about the supposed “debate” over who is better, Kershaw or Bumgarner. Surely the last few games these guys pitched are relevant data to answer that question.
“There are people trying to manufacture a debate about Clayton Kershaw vs. Madison Bumgarner. No. It’s insane. That’s insane. A handful of October appearances doesn’t erase what Kershaw has been able to accomplish, in 2014 and beyond 2014.”
This quote gives me hope. I’ve seen 4 out of 5 online comments about Kershaw in the last week bringing up what a choke artist he is and how much better Bumgarner is and repeat. Obviously it’s my fault for the places I’ve been reading these (Twitter and Facebook), but facts just go over these people’s heads.
This isn’t to say Bumgarner isn’t a great pitcher. He is really great. What he did in the playoffs and WS was phenomenal. That doesn’t change the fact that Kershaw has done far more than he has every year since Bumgarner has been in the league. Going from 2011 till 2014 (I’m not putting 2010 because it was Bumgarner’s rookie season, and while it was good, he greatly improved after it) here are their stats:
Bumgarner – 60-43 129 GS 831.2 IP 8.66 K/9 2.16 BB/9 0.77 HR/9 3.08 ERA 3.07 FIP 3.21 xFIP
Kershaw – 72-26 126 GS 895.1 IP 9.53 K/9 2.01 BB/9 0.51 HR/9 2.11 ERA 2.41 FIP 2.78 xFIP
Bumgarner’s numbers are really good, but he’s tied for 15th in fWAR at 14.6, while Kershaw leads in that time at 25.8.
Both pitchers are great, and they are magical to watch going head to head, with both guys dominating the other team like you can’t believe. I’m just glad to see rational thought comes into play when talking about who is better.