Originally Posted by the 'stache
Was he?
I think WAR is good for comparing players within the same era. I think it is far less useful when comparing players from different eras.
Let's look at the last five years of both pitchers. 1962 to 1966 for Koufax, and 2011 to 2015 for Kershaw (which is not yet completed, but it's close enough).
One thing we have to remember is that a pitcher's wins, or more specifically, his win percentage, is a component of WAR. A pitcher can pitch great, and yet if his team doesn't score runs in support of his efforts, he might not get a win. Earlier I referenced what a shame it is that Kershaw doesn't win more. It's obviously not because he doesn't pitch well. I would imagine it's a lack of run support, in part. Let's compare some numbers.
You said that Kershaw's best WAR is 7.8. Koufax has three seasons better from a WAR standpoint: in 1963 (10.7), 1966 (10.3), and 1965 (8.1). In those three seasons, Koufax won 25, 27 and 26 games. In the five year period, Koufax was 111-34, a .766 win pct. He averaged more wins per season (22) than Clayton Kershaw's best win total in any of his five years (21 in both 2011 and 2014). Kershaw, in his near five years, is 82-32, a .719 win pct, and an average of 16 wins a season. Kershaw has started 151 games to Koufax's 176.
When individual metrics are compared side by side, I don't see anything else that would explain the huge difference in WAR.
ERA:
Koufax 1.95
Kershaw 2.14
Strikeouts per 9 IP:
Koufax 9.4
Kershaw 9.8
Walks per 9 IP:
Koufax 2.1
Kershaw 1.9
Strikeout to BB ratio:
Koufax 4.57
Kershaw 5.04
WHIP:
Koufax 0.926
Kershaw 0.942
Hits per 9 IP:
Koufax 6.3
Kershaw 6.5
ERA +:
Koufax 167
Kershaw 171
Home runs allowed per 9 IP:
Koufax 0.6
Kershaw 0.5
Their numbers are really, really close, aren't they? They give up about the same number of hits and walks per game, Kershaw strikes out batters slightly more often often. Koufax has a better ERA, but relative to the rest of the league at the time they played, Kershaw's ERA + is slightly better.
The difference in WAR? Koufax pitched at a time when complete games were far more common (100 for him, only 18 for Kershaw). That leads to a huge discrepancy in innings pitched:
Koufax 1,377
Kershaw 1,072.
By the time this season is done, Koufax will still have thrown about 250 more innings, which breaks down to about 50 per year. So, while Koufax was allowed to stay in, and pitch a whole game, Kershaw gets pulled out, and his it is the responsibility of his bullpen to preserve the lead, and ultimately get Kershaw the win. Poor bullpen performance means ultimately fewer wins for the pitcher, affecting the starter's winning percentage, and WAR.
Look at Kershaw's 2013 season. He started 33 games, had a 1.83 ERA (best in the NL), and won only 16 games. Look at some of the dominant starts he had where he didn't get a win:
May 3rd vs San Francisco, 7 IP, 1 earned run allowed. No decision. Dodgers lose 2-1.
May 8th vs Arizona, 7 IP, 1 earned run allowed. No decision. Dodgers lose 3-2.
June 10th vs Arizona, 7 IP, 1 earned run allowed. No decision. Dodgers lose 5-4.
June 15th vs Pittsburgh, 7 IP, 1 earned run allowed. No decision. Dodgers win 5-3.
July 31st vs New York Yankees, 8 IP, 0 earned runs allowed. No decision. Dodgers lose 3-0.
September 8th vs Cincinnati, 7 IP, 2 earned runs allowed. No decision. Dodgers lose 2-3.
6 starts, 43 innings pitched, 6 earned runs allowed. Kershaw had a 1.26 ERA across these six starts, and got nothing to show for it. The Dodgers scored 11 runs. Kershaw, on average, got 1.83 runs support per game. The Dodger bullpen? They gave up 12 runs in 14 innings.
Then there are some of the games he lost:
August 6th vs St. Louis, 6 IP, 2 earned runs allowed. Dodgers lose 5-1.
August 27th vs Chicago Cubs, 7 IP, 1 earned run allowed. Dodgers lose 3-2.
September 13th vs San Francisco, 7 IP, 2 earned runs allowed. Dodgers lose 4-2.
Clayton Kershaw should have won 20 games easily in 2013, if not more. But his bullpen was awful in the games he started, and he didn't get any run support in many of his games.
Compare Kershaw's 2013 season to Sandy Koufax's 1963 season. Koufax won 25 games, and again, Kershaw won 16.
Kershaw had 33 starts. Koufax 40. Kershaw's ERA was 1.83. Koufax's ERA was 1.88.
The Dodgers scored 125 runs, or 3.79 RPG, for Kershaw's 33 starts.
The Dodgers scored 172 runs, or 4.30 RPG, for Koufax's 40 starts.
The Dodgers scored 2 or fewer runs in 17 of Kershaw's 33 starts (51.5%)
The Dodgers scored 2 or fewer runs in 12 of Koufax's 40 starts (30.0%)
What am I driving at? We're becoming conditioned to look at WAR as the be-all, end-all metric for comparing players (not you, specifically, Peter, I'm speaking in the abstract). If we accept this, Sandy Koufax was a much better pitcher than Clayton Kershaw is now. After all, Koufax's best WAR seasons of 1963, 10.7, is much better than Kershaw's best season of 7.8 in 2013, a difference of nearly 3 wins. But the truth is, the discrepancy in their WAR figures can be accounted for by the very fact that Koufax won more, and pitched more games/innings. He also got a half more run support per game, had fewer instances where the offense scored 2 or fewer runs in his starts, and didn't have to suffer at the hands of an inept bullpen. Now, Koufax should have a higher WAR. Yes, starters back in the 1960s did throw more innings per start, and Koufax held his level of excellence through higher pitch counts. But was he better than Clayton Kershaw on an inning by inning basis? No. The numbers do not support this assertion.
|