Posts Tagged ‘Sabremetrics’

The American League Cy Young race was seen by many as a battleground between modern sabremetrics and old fashioned statistical analysis. As a result, the “controversial” selection of Felix Hernandez was heralded in some corners as the dawning of an age of enlightenment, while in others it was viewed as a turn to the “dark side”.

There’s no point in trying to settle that debate because both sides seem firmly entrenched in their positions. What is more interesting, however, is whether or not Hernandez’ selection represented much of a change in the thought process used by the BBWAA voters to elect the Cy Young.

As Tyler Kepner noted in the New York Times’ Bats Blog, one really didn’t need to delve too deeply in advanced metrics in order to appreciate Hernandez’ accomplishments in 2010. Even though he was a pedestrian 13-12, the Mariners’ ace led the league in ERA and placed second in strikeouts (only one behind the leader), two statistics that have factored into historical voting almost as much as wins.

Without a doubt, wins have always played a role in selecting the Cy Young. In the American League, 26 of the 45 winners since 1967 (the first year a separate award was given in each league) led in wins, while 39 came within 10% of the league-leading total. After removing the four relief pitchers who won the award, the percentage increases to 63% and 95%, respectively. In the National League, 64% of non-reliever Cy Youngs finished first in wins, while 79% finished within 10% of the best total. Combined, the correlation between the Cy Young and win total is stronger than any other statistic.

Cy Youngs Who Have Not Finished Within 10% of League Leader in Wins

Note: Relievers excluded from consideration.
Source: Baseball-reference.com

When you consider that the last two AL Cy Youngs, Zack Greinke (84% of top win total) and Hernandez (62% of top win total), are the league’s only award winners to not finish within 10% of the leading win total, it does seem as if there has been a philosophical change among the voters. After taking a deeper look, however, it becomes clear that Cy Young voters have never lived by wins alone.

Cumulative Rankings of Cy Young Award Winners in Three Traditional Statistical Categories





  Leader Within 10% Leader Within 10% Leader Within 10%
AL 63% 95% 44% 76% 22% 41%
NL 64% 79% 33% 85% 44% 59%

Note: Relievers excluded from calculation of percentages; the American League had two award winners in 1969.
Source: Baseball-reference.com

As illustrated in the chart above, both ERA and strikeouts have a strong relationship to winning the Cy Young. In fact, in the National League, a higher percentage of winners have come within 10% of leading in ERA than wins. Nonetheless, the correlation to wins is still significantly stronger, but a few caveats are in order. Because ERA is a rate statistic, there is not only more room for variance, but qualified pitchers with much fewer innings pitched are not disadvantaged as they would be with cumulative totals. Also, although strikeouts, like wins, are cumulative, the league leading total is often 10-15x more than the best win mark. Again, this introduces more variance among the leaders. After taking those two qualifications into account, it certainly seems as if voters have always given very careful consideration to both ERA and strikeouts.

Relative Performance of Cy Young Winners in Wins, Strikeouts and ERA

  Leader in all 3 categories Leader in at least 2 categories Leader in at least 1 category
AL 10% 37% 83%
NL 10% 38% 92%
  Top 10% in all 3 categories Top 10% in at least 2 categories Top 10% in at least 1 category
AL 29% 73% 100%
NL 28% 82% 100%

Note: Relievers excluded from calculation of percentages; the American League had two award winners in 1969.
Source: Baseball-reference.com

In the National League, only three Cy Young starters have failed to lead at least one of the three categories under consideration in this analysis, while seven have come up short in the AL. However, every award winner has at least finished among the top 10% in one category, with a large percentage achieving the feat in at least two. It should be noted, however, that among the 18 pitchers who only ranked in the top 10% of one category, 16 were listed among the leaders in wins (including 14 who actually led their league). Only Greinke in 2009 (led the league in ERA; 84% of wins leader; and shade below 90% of strikeout leader) and Rick Sutcliffe in 1984 (traded to NL on June13; only finished among top-10% in ERA, but went 16-1 for Cubs) were able to buck that trend. So, although it seems as if voters have always considered more than just wins, they have often allowed a leading total in that category to obscure other relative deficiencies.

An evolution in the criteria that beat writers use to vote on awards like the Cy Young certainly seems to be underway. However, it would be stretch to suggest that this gradual tidal shift is really a sea change. As startling as Felix Hernandez’ win total may be (at 62% of the leading total, it is the lowest among all non-relievers since 1967), he still led the league in ERA and strikeouts (he was actually one behind Jered Weaver, but based on rounded percentage was within 100% of the leading total), which has usually been good enough to win the Cy Young.

Pitchers Who Led Their League in Strikeouts and ERA, but Didn’t Win the Cy Young Award

Year ERA/K Leader Wins WAR Cy Young Wins WAR
1970 Tom Seaver 18 6 Bob Gibson 23 8.7
1971 Tom Seaver 20 9.2 Fergie Jenkins 24 9.2
1979 J.R. Richard 18 6 Bruce Sutter 6 4.6
1987 Nolan Ryan 8 5.5 Steve Bedrosian 5 2.6
2002 Pedro Martinez 20 5.7 Barry Zito 23 6.5

Source: Baseball-reference.com

Of the five times (out of 21 cases) that the electorate turned away from the ERA/strikeout leader, it either awarded a reliever or the league leader in wins. Of the latter, however, each pitcher also happened to have a higher WAR: Jenkins and Gibson were the league leaders in their Cy Young season, while Zito finished a close third behind Roy Halladay (6.9) and Tim Hudson (6.6). Interestingly, the correlation to WAR doesn’t stop there. The baseball-reference.com blog ran a comparison of Cy Young winners to WAR and found that 46 of the 80 non-relievers (58%) actually led the league in the website’s calculation of the metric, which isn’t far from the 64% correlation to the stodgy old wins category. This really shouldn’t come as much of a surprise, however, considering our previous discussion. As foreign and advanced as WAR may seem, common place metrics like strikeouts and those that influence ERA play a significant role in many sabremetric constructs.

So, let the philosophical battle rage on. Hernandez’ Cy Young victory doesn’t belong to either side.

Read Full Post »