Clayton Kershaw, Felix Hernandez and Appreciating Greatness

How do you measure a truly great starting pitching season? Having an ERA that starts with a “1” generally qualifies, but there’s some obvious issues with that. First, that’s happened 82 times in the last 100 seasons, making it notable but perhaps not unthinkable. Second, obviously, are the flaws inherent to ERA, most importantly that it’s not adjusted for ballpark or league. Pedro Martinez (2000), Sandy Koufax (1964) and Carl Mays (1917) all had an ERA of 1.74. Clearly, none of them were facing the same kind of offenses.

You could, if you wanted, go by WAR. Steve Carlton‘s 1972 and Martinez’ 1999 make sense atop the list, but convincing people that Bert Blyleven‘s 1973 was the third-best season ever or that Bob Gibson‘s legendary 1968 was merely his third-best season seem like tougher sells. Besides, since that’s a counting stat rather than a rate stat, it means no modern-day pitcher will ever be able to come close, because it seems pretty safe to say that we aren’t seeing a starting pitcher top 320 innings again, as both Carlton and Blyleven did.

FIP? That’s better, though still imperfect. Martinez, again, and 1984 Dwight Gooden top the leaderboards there, followed by a pair of guys essentially playing a different sport, Walter Johnson and Pete Alexander in 1915. (In 1915-16, Johnson pitched 706.1 innings; he allowed one home run. It was, as were 24 of the 97 total dingers he allowed, an inside-the-park job.) FIP also assumes some league-average inputs, and if we want “best-ever” perhaps we don’t want to assume any kind of average; like WAR, you’d also have a tough time winning a bar argument with something that you need to explain formulas for.

Enough setup, then. To the point, now. Clearly, there’s many different ways to do this, and no obvious, unassailable answer. You could make an argument for probably a dozen different years as the “best” starting pitching season of all time. What I’m doing, today, is to break it down into the most important things a pitcher can do that are more or less entirely within his control:

  • Get strikeouts
  • Limit walks
  • Limit homers
  • Get ground balls

If you’re good at any of those things, you might have a big league career. Be good at two, and you might have a productive big league career. Be good at all of them, to an exceptional degree, and you’re a superstar.

Let’s go back to the qualified pitcher seasons since 1915, of which there are 7,464, and set some requirements for those four items. There’s 149 seasons (0.019%) where a starter has struck out at least 25% of the hitters he faced. Of course, the sport is striking out so much more often now than it used to, making raw whiff numbers a bit skewed towards the present-day pitchers, something of the inverse of how WAR is going to work better for the older guys who pitched more often. (To use a favored example, Gibson’s career 19.4 K% is identical to luminaries Chris Capuano, Shaun Marcum and Mike Trombley.) So I asked our pal Jeff Zimmerman to help me figure out how many times a K% was at least eight percent higher than the league average for that season. The answer, as it turns out, is 202 times. It’s slightly more often than using straight K%, but less often than I think I expected.

669 times, or 0.089%, a pitcher has walked fewer than five percent of the hitters facing them. Only 18 times — 0.002% — has a pitcher topped 25% in strikeouts and kept walks below five percent, and there’s some great seasons in there. Martinez’ 1999 and 2000, of course. Matt Harvey, last year. Curt Schilling shows up three times. Cliff Lee twice. Only one of the 18 is from before 1999, which is Koufax’ 1963. We can do better than that. We don’t just want good, or very good. We want truly elite, and that means a fantastic ability to prevent home runs. Four of those seasons had more than one home run per nine innings. Ten of them were above 0.50/9. Still very good, and yet now out of the mix.

We’re left with four seasons. Now we’re going to split that in half. An out is an out, and so a a grounder may not be specifically more valuable than a fly out or a pop out. But it is exceptionally difficult to hit a ground ball over the fence, and so ground ball pitchers are often seen as more desirable than fly ball types; you can pitch a ground ball pitcher anywhere, but the fly guys look a lot different based on what park they’re in. Let’s limit our pitchers to only seasons where the ground ball rate has been at least 50%; this, unfortunately, goes back only to 2002, when batted ball data was first available. 218 pitchers have done that.

Now, let’s combine it all, which leaves us with K rate above 25%, BB rate below 5%, and ground ball rate above 50%. We’re left with two guys. They’re both doing it right now. You probably already know who they are:

Name K/9 BB/9 K/BB HR/9 K% BB% ERA- FIP- xFIP- ERA FIP xFIP GB%
Clayton Kershaw 10.76 1.25 8.58 0.40 31.80% 3.70% 50 49 54 1.78 1.75 2.01 55.00%
Felix Hernandez 9.68 1.60 6.06 0.35 28.30% 4.70% 51 55 62 1.95 2.07 2.38 56.10%

It’s probably not earth-shattering to point out that Clayton Kershaw and Felix Hernandez are the best pitchers in baseball. They’re all but guaranteed to win the Cy Young awards assuming health; Kershaw could have been headed towards a streak of four in a row if not for the infatuation with R.A. Dickey‘s knuckler and story in 2012. Neither one is even 29 yet, and we’re talking about them both as though they’re obvious Hall of Famers someday.

What this is, really, is just showing another way that they’re the best pitchers in baseball. They’re two of a very, very small group — 0.002% in the last 100 seasons — with the ability to get whiffs at the pace they’ve done while keeping walks to a minimum. They’re two of only four to limit homers below 0.5/9 in that group, which is not unrelated to the fact that they keep the ball on the ground so well. Due to the limitations of batted ball data, we can’t say with complete certainty that no one has done this before, but we can make an educated guess. We can say it’s insanely, unbelievably good.

If you prefer this in a visual format, here’s a graph of every qualified starting pitcher season since 2002, all 1,136 of them. Obviously, a huge majority of pitchers are clustered in the middle, getting grounders between 35 and 50 percent of the time, and with K%-BB% between 5 and 15%. If you’re at the bottom left, you’re probably in trouble. (No, that’s not a mistake — there are four guys who managed to qualify with negative K%-BB% marks.) If you’re one of the two dots in the extreme upper right, well… you’re doing it right:

gb-kbb

The point, again, is not simply to say, “hey, Kershaw and Hernandez are good.” You know that. What might somehow be not fully appreciated is just how good, and at how many different things. Perhaps, because offense is so low right now, it’s more difficult to stand out as a pitcher. Or perhaps offense is so low because of how great pitchers like Kershaw, Hernandez, Chris Sale, Stephen Strasburg, Yu Darvish, etc. all are. Or maybe, and possibly most likely, we’re seeing two of the best to ever pitch, , we’re seeing them at their best, and even within the context of the great things they’ve done before, they’re outdoing themselves. They’re doing things we’ve just about never seen before, ever. We should be thankful every single time they take the mound. This isn’t just “best pitcher in baseball” performance; this is “best you might see in your entire lifetime” kind of stuff.



Print This Post



Mike Petriello used to write here, and now he does not. Find him at @mike_petriello or MLB.com.


Sort by:   newest | oldest | most voted
tehzachatak
Member
tehzachatak
2 years 18 days ago

Who is the point out to the far right beneath Kershaw? (~27.5% K%-BB%, ~40% GB). That must be a pretty remarkable pitcher as well. Darvish?

tehzachatak
Member
tehzachatak
2 years 18 days ago

Never mind, this chart is 2002-2014. Looks like.. ’02 Schilling?

a eskpert
Guest
a eskpert
2 years 18 days ago

It has to be.

Feeding the Abscess
Guest
Feeding the Abscess
2 years 18 days ago

Chris Sale

The Corey Kluber Society Will Rise Again
Guest
The Corey Kluber Society Will Rise Again
2 years 18 days ago

I can’t but point out that Kluber, this season, rounds off to those 25/5/50 marks (27.2% K rate, 5.2% walk rate, 49.7% GB). Of course Kershaw and King Felix don’t just hit those marks, they blow through them; they’re definitely the two best pitchers in baseball. But we could easily have a third member of the club by season’s end.

The Corey Kluber Society Will Rise Again
Guest
The Corey Kluber Society Will Rise Again
2 years 18 days ago

Oh, and #KeepNotGraphs

Nate
Guest
Nate
2 years 18 days ago

In a parallel universe Felix doesn’t exist and Kluber, Sale, Darvish, (and to a lesser, injury-prone, extent, Tanaka) are all competing for the Cy Young.

AA
Guest
2 years 18 days ago

I’m lucky enough to have strong memories of the careers of Roger Clemens, Greg Maddux, Pedro Martinez and now Kershaw and Felix. Soak it up.

cmg8462
Member
Member
cmg8462
2 years 18 days ago

Matt Harvey is the point below Felix @ 48% grounders — His numbers were really really good (K-27.7% BB-4.5% GB-48% HR/9-0.35 — As a Mets fan I really hope he can be even half of what he was in 2013 coming back from TJS.

bucs_lose_again
Guest
bucs_lose_again
2 years 18 days ago

Fantastic read.

Dustin
Guest
Dustin
2 years 18 days ago

Corey Kluber has a non-zero chance of making this list. ZIPs projects he’ll throw 50 additional IP this season. That would put him at 222 IP for the season. His K% is currently at 27.2%, BB% is at 5.2%, GB% at 49.7% and HR/9 at .52. Now for some extrapolating.

Kluber is facing exactly 4 batters per IP(4.008), so we’re looking at 200 additional batters faced, for a season total of 888.

To hit a 25% K rate, Kluber needs 222 K’s. That’s 35 K’s in 50 IP. He’s likely in the clear there.

To get below a 5% BB rate, Kluber can only walk 44 batters. He’s at 36 currently. 8 BB’s over 50 IP is a 1.44 BB/9. It’s a tough number to hold to, but Kluber has only walked 7 in his last 54 IP.

To get to a sub .50 HR/9 in 222 IP, Kluber needs to allow 12 or fewer HR. He’s already at 10. This will be hard, but not impossible. He’s only allowed 2 in his last 68 IP.

Calculating the GB rate is a bit more difficult, but if Kluber had generated 2 additional GB’s at any point this season, he’d be above 50%.

It’s likely not going to happen, but if he keeps pitching the way he has, Kluber’s 2014 is going to come dangerously close to making this leaderboard of the greatest performances of modern times.

The Corey Kluber Society Will Rise Again
Guest
The Corey Kluber Society Will Rise Again
2 years 18 days ago

Arbitrary endpoint alert, but if you throw out Kluber’s April starts and take his season from May onward, he has exactly a 5% walk rate and 50% GB rate. And his grounder rate’s been generally improving throughout the season. I’d give him better than even odds of ending up above 50%. As you suggest, walks are probably gonna be the limiting factor here.

maqman
Guest
maqman
2 years 18 days ago

Beauty is in the eye of the beholder. In over 60 years of watching MLB games the best I’ve seen in order were Koufax, Pedro, Nolan Ryan, Randy Johnson, Kershaw, Felix, Ron Guidry and Fernando Valenzuela.

tz
Guest
tz
2 years 18 days ago

For those too young to remember Guidry, he was a small power-pitching lefty whose track record was a bit like a later-blooming Johan Santana:

http://www.fangraphs.com/graphsw.aspx?players=1005124,1007124,755

I never saw Koufax, but if he was anything like peak Guidry he must have had filthy stuff.

mt low rider
Guest
mt low rider
2 years 17 days ago

I was pretty young but Koufax had the nastiest curveball perhaps ever and threw gas as well. IIRC he had more than a few years over 350 K’s. Then he dominated my Twins in the ’65 WS. Too bad his career was so short.

Bip
Member
Member
Bip
2 years 17 days ago

I’m sure Fernando was exciting to watch, but he seems like the odd man out in that group.

Rick
Guest
Rick
2 years 18 days ago

By what method do you compare Kershaw’s NL season with Felix’s AL season?

The Foils
Member
The Foils
2 years 18 days ago

pluses and minuses

Kim
Guest
Kim
2 years 18 days ago

Yeah, imagine if Felix didn’t have to face a DH and got a pitcher nearly every start.

Sabremetrics
Guest
Sabremetrics
2 years 18 days ago

Then imagine if there were some way to adjust for park and league factors. What would we call these amazing stats?!?!?

Nick
Guest
Nick
2 years 17 days ago

Yeah then imagine if he had the distraction of having to bat… hmm, I’d say it about equals out.

It’s time to get rid of the DH, get rid of the pitcher batting, and go to 26 man rosters with 8 man lineups. That would be great.

Bip
Member
Member
Bip
2 years 17 days ago

I think people assume the DH adjustment is more dramatic than it really is. It’s not like NL lineup – pitcher + DH = AL lineup. Most DH’s could find jobs in the NL as first basemen. Sometimes, to change an NL lineup to an AL one, you might move the 1B to DH, LF to 1B, and then maybe put a defensive-specialist into the empty OF spot. This lineup doesn’t hit much better than the NL version, but it actually defends better.

So, overall we would expect AL lineups to hit better than NL ones, but the difference is not going to be as dramatic as swapping the pitcher for a typical AL DH. And, in fact, AL defenses might even be better, because AL teams get to hide their worst defender, possibly opening a spot for a D-first player, which could be an advantage for Felix (though maybe not, given the Mariners’ tendency to collect DH’s.)

rustydude
Member
rustydude
2 years 18 days ago

Finally, a piece about Kershaw that also points out how outstanding Felix’s season is, as well. Two best pitchers for the last 6 years, as well.

The Humber Games
Guest
The Humber Games
2 years 18 days ago

As you said, the game has changed. Maybe the fact that you came up with two pitchers from this year is more a sign that our preferences for pitchers are specific to this era? For example, maybe fly balls hurt pitchers less in a deadball era, so the definition of excellent peripherals was different then than now. Maybe pitching to contact rather than strikouts was a premium talent in an era when contact did less damage. It would be interesting to see if skill set correlation to WAR was different 50 years ago, and which skill sets were producing the best results and if that is different from now. Just a thought.

olethros
Guest
olethros
2 years 18 days ago

No batted ball data prior to 2002, and I very much doubt that many baseball folks thought in terms of peripherals, or rate stats, until relatively recently. Hell, many baseball folks are actively, vociferously resistant to those concepts now.

Alan
Guest
Alan
2 years 17 days ago

Frankly, I don’t really understand the opening when the author seems to reject methodologies because they don’t properly account for era or ballpark…

…and then chooses criteria that are highly biased to the current era. K Rates are very high right now, so it’s not a shock that two modern players came up.

Mike Green
Guest
Mike Green
2 years 18 days ago

Fernandez is having easily the best year of his career. Whether he will sustain it is another question. If you look back at his career to date, he has been both great and very durable. Nonetheless, he is noticeably behind Tom Seaver and Roger Clemens on the key performance measure for a player with about 2000 innings- ERA+. If you are looking at peak performance, he’s obviously a long way from Pedro or for that matter Randy Johnson or Lefty Grove.

Kershaw is a whole other story. He’s pretty much at the same performance level as Seaver and Clemens, but how durable he will be over time is another question. Again, from a peak perspective, Kershaw is pretty clearly looking up at Pedro. And it isn’t close.

The Ancient Mariner
Guest
The Ancient Mariner
2 years 18 days ago

Which Fernandez are you talking about?

#KeepNotGraphs
Guest
#KeepNotGraphs
2 years 18 days ago

Helix

The Ancient Mariner
Guest
The Ancient Mariner
2 years 18 days ago
Bip
Member
Member
Bip
2 years 17 days ago

That’s pretty cool. I’d like to see the same graph in RA/9 WAR, as Kershaw has 7 more RA/9 WAR than fWAR.

andyk
Guest
andyk
2 years 18 days ago

How much does the DH change the strikeout rate for Hernandez? Does it change when he pitches in NL parks? Curious to know if Kershaw’s K rate changes in AL parks too. Guess I’ll go try to look that up now.

TKDC
Guest
TKDC
2 years 18 days ago

I wonder how many guys had those benchmarks through 136 (or even 180) innings in the last 12 years? I’m not taking anything away from what these guys have done, but the season isn’t over.

Dougie G
Guest
Dougie G
2 years 18 days ago

Shout out to my boy Pedro for getting jobbed off this list because of lack of data. Meets the first three criteria and never had a GB% higher than 42% even after his best years.

SABRphreak
Guest
SABRphreak
2 years 18 days ago

If this article is about truly great starting pitching seasons in our lifetime, then we need to start with Pedro Martinez’s 1999 and 2000 seasons. Perhaps we can do that with Kershaw’s 2014 after the season (although fairly unlikely). Kershaw and King Felix may be the best this season, but not “best in our lifetime” sort of stuff.

Here are some stats:

ERA+
1. Pedro Martinez, 2000 Red Sox: 291
2. Greg Maddux, 1994 Braves: 271
3. Greg Maddux, 1995 Braves: 262
4. Bob Gibson, 1968 Cardinals: 258
5. Pedro Martinez, 1999 Red Sox: 243

FanGraphs WAR (since 1974)
1. Pedro Martinez, 1999 Red Sox: 12.1
2. Roger Clemens, 1997 Blue Jays: 11.1
3. Randy Johnson, 2001 D-backs: 10.7
4. Pedro Martinez, 2000 Red Sox: 10.1
5. Roger Clemens, 1988 Red Sox: 10.0

It’s also fun to recall the 1999 All-Star game where Pedro pitched two innings, struck out five hitters and didn’t allow a single hitter to reach base. 19 of his 28 pitches were thrown for strikes. The feat is even more impressive when you consider the players he struck out — Barry Larkin, Larry Walker, Sammy Sosa, Mark McGwire and Jeff Bagwell — the same year that McGwire and Sosa hit more than 60 HRs each. If it wasn’t for an error by Roberto Alomar, Martinez would have pitched a perfect two innings.

BTW, here’s Pedro’s opponent batting line for 2000: .167/.213/.259. Are you not impressed?

tz
Guest
tz
2 years 18 days ago

Using Baseball References neutralized pitching tool, Pedro would have had a 1.95 ERA and a 0.81 WHIP in 2000.

Oops, that was what his numbers would have been playing for Colorado for 2000.

Let’s try again: 1.33 ERA and 0.66 WHIP, using the historic MLB run environment

http://www.baseball-reference.com/players/m/martipe02-pitch.shtml

(For fun, if Pedro was pitching for the 1968 Cardinals: 1.00 ERA, 0.57 WHIP)

BenRevereDoesSteroids
Guest
BenRevereDoesSteroids
2 years 18 days ago

And that was in a more hitter friendly park than Felix/Kershaw, with the MLB seeing more offense as a whole than it is today! And he still put up those ridiculous numbers! 1.39 FIP in 200+ inning says it all.

james wilson
Guest
james wilson
2 years 17 days ago

Best pitcher on the needle ever.

Shawnuel
Guest
Shawnuel
2 years 18 days ago

Since “Fernandez” was invoked, I’d like to mention that Jose Fernandez, had he remained healthy, could easily have been a part of this article the way his season was going.

ivdown
Guest
ivdown
2 years 17 days ago

Didn’t he have like 8 starts?

Hurtlocker
Guest
Hurtlocker
2 years 18 days ago

Wow, just wow. I keep forgetting that there wasn’t math more than 20 years ago. Walter Johnson was great because he was so much better than his his peers. So was lefty Grove, so was Sandy Koufax. You can spin this mathmatically any way you want, but to claim modern math makes these guys the best ever is just wrong.

Bip
Member
Member
Bip
2 years 17 days ago

It’s excluding players before 2002 because we don’t have batted ball data before that…

Kazinski
Guest
Kazinski
2 years 17 days ago

It seems awfully arbitrary to put in GB% as that data only goes back to 2002, now you are only talking about season’s in the recent past, which kind of undercuts the argument of all time greatness.

I took a stab at coming up with a list that goes back at least to 1960, and takes offensive levels of different eras into account. The filter I came up with was WHIP < 1, Qualified innings, and OPS+ 50, there were 15 players total, including Chris Sale this year and Jose Fernandez last year.

As a little further comparison between Felix and Kershaw here are their slash lines against non-pitchers this year, which explains why Felix’s overall numbers don’t match Kershaw’s but his OPS+ is significantly better:

Kershaw .209/.242/.307
Hernandez .191/.231/.274

Kazinski
Guest
Kazinski
2 years 17 days ago

I guess my table got edited. Here is a link:

http://bbref.com/pi/shareit/Hhvjm

Vinny
Guest
Vinny
2 years 17 days ago

Looks like you forgot to move the decimal points on the pitcher seasons.
e.g.
149/7464=0.019=1.9%
669/7464=0.089=8.9%

Great article though…

mt low rider
Guest
mt low rider
2 years 17 days ago

In any case Felix and Kershaw are must see TV this year. I’m a Seattle fan but I certainly can’t make the case that Felix has been better than Kershaw other more starts and IP. They have both been dominant. Felix has 12 losses and no decisions this year and has given up 18 earned runs in those games. That’s mind blowing. Felix has pitched his whole career knowing if he gave up 3 runs he would most likely get an L. Kershaw at least has a competent offense behind him. Hard to quantify that but common sense says Felix should have 20 or 30 more wins with a league average offense. 2 great pitchers, I hope they both stay healthy.

Mark
Guest
Mark
2 years 17 days ago

I don’t get calling out Dickey for winning the Cy in 2012. Since he’s thrown the knuckleball his ERA has always been better than his FIP, and it makes sense to use RA-9 WAR instead of FIP based WAR for him. He had 6.2 that year, and it’s not an unreasonable choice.

Kris Gardham
Guest
Kris Gardham
2 years 16 days ago

This is kinda neat, and I’d actually like fangraphs to do more stuff like this. What you’re doing is basically breaking down the components of FIP and graphing/charting them individually.

I’m not sure who’s the visualization guy for fangraphs, but I think it’d be awesome to create one of those fancy infographics showing where the top players WAR/FIP/etc comes from.

I love graphs and stuff, and I’d be interested to see a) the best players in graphs and b) the absolute average players in graphs because there’s so many different ways to get to 0.

Ellis
Guest
Ellis
2 years 15 days ago

I didn’t read through all the comments yet, so I don’t know if somebody has asked this, or if it has been quantified, but here is my question:

Is there any way to separate NL vs AL, specifically relating to Felix vs. Kershaw? How much does facing a pitcher/pinch hitter more often compare to facing a DH? Is there any way to measure this and alter the stats based on league differences? How different might Kershaw and Felix’s stats be if they switched places?

With the higher frequency of inter-league games, I’m sure it’s less of a difference now, but there must be a difference.

wpDiscuz