|
| Johan leads the NL in ERA, but has he been the best pitcher?
|
With Johan Santana coming back to the pack lately, the NL currently has nine starters with ERA between 2.00 and 2.70.
1. Johan Santana - 2.00 ERA in 72 IP
2. Wandy Rodriguez - 2.26 ERA in 67 2/3 IP
3. Matt Cain - 2.31 ERA in 66 1/3 IP
4. Dan Haren - 2.42 ERA in 78 IP
5. Johnny Cueto - 2.43 ERA in 74 IP
6. Jair Jurrjens - 2.59 ERA in 66 IP
7. Chad Billingsley - 2.59 ERA in 80 IP
8. Zach Duke - 2.62 ERA in 79 IP
9. Josh Johnson - 2.66 ERA in 74 1/3 IP
No one else is under 3.00 and there aren't any relievers on pace for 60 saves, so let's label the above as the nine legitimate candidates.
First off, I like looking at unearned runs and defense. Santana and Rodriguez have both given up seven unearned runs apiece, while no one else on the list had allowed more than two. Haren is the only one in the group yet to give up an earned run.
On to the gloves. Duke is being helped out by the game's second most efficient defense to date. Billingsley's defense is third, while Cueto's is ranked fifth. On the other side, the Mets are 19th, the Astros 25th, the Giants 18th and the Diamondbacks 22nd.
With additional factors, such as bullpen support and ballpark, being weighed, VORP still has an identical top nine NL starters, but ranks them in a vastly different order.
1. Dan Haren - 29.0
2. Johnny Cueto - 26.4
3. Chad Billingsley - 25.4
4. Zach Duke - 24.6
5. Matt Cain - 24.3
6. Josh Johnson - 23.1
7. Johan Santana - 22.4
8. Jair Jurrjens - 21.8
9. Wandy Rodriguez - 19.1
I'm not a big fan of VORP, but I think it comes up with a very good list here. Santana was the NL's best pitcher for six weeks and he'd probably be my Cy Young pick for the rest of the season, but he doesn't deserve the top spot right now.
First third NL Cy Young
1. Haren
2. Billingsley
3. Cueto


It's insane that Tim Lincecum doesn't show up anywhere here. That fact should put to rest any lingering doubts about VORP's utter inadequacy and irrelevance. Nobody in the NL has pitched better than Lincecum (1.94 FIP, 4.33 K/BB, .25 HR/9), but his topline numbers have been skewed by bad luck (.369 BABIP).
That's because VORP is about what you did do, not what you could have done with average luck.
BABIP is nice to keep track of as it's a decent predictor of future performance since it will likely regression back to the mean, but it still counts.
If you're not a big fan of VORP, then why don't you use some other context-independent stats, such as FIP or xFIP? I guess I don't see the point of the second half of the article.
I think FIP shows us more how a pitcher might have pitched, rather than how he did. The truth is that those runs did score against Lincecum, and while he shouldn't be penalized for bad fielding, the bad luck has mattered when it's come to wins and losses.
Still, it is worth pointing out.
FIP
1. Lincecum - 1.97
2. Wandy - 2.52
3. Javier Vazquez - 2.61
4. Johan - 2.71
5. Billingsley - 2.81
6. Josh Johnson - 2.97
7. Haren - 3.15
I suppose there is something to be said for this line of thinking, but lots of research has shown that pitchers have very little control over the things Lincecum has done worse than the others listed above. We shouldn't give pitchers credit for what might have happened, true. We also shouldn't give them credit for playing in front of good fielders or for simply getting lucky. The gap between Lincecum and the field based on controllable skills is substantial. He has pitched better than anyone else in the NL, and it's not even close.
That may have been true a few years ago. But further research is starting to show that pitchers can affect some of those factors a bit, like BABIP. For example, extreme groundball pitchers like Brian Bannister, Brandon Webb, and Chien Ming Wang can keep their BABIP lower because the balls in play that they are likely to give up are more likely to be infield balls that are either fielder easier than line drives or when they do fall for hits, they fall as singles rather than doubles or triples. Plus the balls in play are more likely to end up as double plays as well.
Tim Lincecum is a pretty decent flyball pitcher, and even last year his BABIP was an above average .310. On the other hand ignoring their current injury-riddled seasons, Chien Ming Wang and Brandon Webb have never posted BABIP as high as that .310.
Of course, this isn't to take anything away from how great a pitcher Tim Lincecum is. He's easily a Top 5 pitcher and even his "down" stats this season are pretty damn good. However, regardless of luck or defense, those were the hands he was dealt and the outcomes are the outcomes. He's great, just not first 1/3 of the season CY Young great ;-)
That's why I said "very little control" instead of "no control." And you've got the other part backwards. Groundball pitchers have higher BABIPs, because balls hit on the ground are more likely to find holes. Fly ball pitchers allow fewer hits on balls in play, but have more of them go for extra bases (either over outfielder's heads or into the gaps). I think this makes intuitive sense and is borne out by research.
Your larger point is a valid one and deserves addressing. I know that evaluating players based on the elements of performance with predictive value while ignoring more the traditional stats is hard to stomach, but I think it's the correct approach. As a fan and armchair analyst, I'm interested in knowing who's doing the most to give himself and his team a chance to win. In this case, the answer is clearly Tim Lincecum.
I don't think I've got it backwards, I just checked baseball reference and the ground ball pitchers, the sinkerballer ones anyway, almost all had a BABIP a good .10 under league average.
Buy anyway, where the real interesting stuff is is looking at BABIP for the three different hit types, ground ball, line drive and fly ball. The BABIP for each of those types is more or less consistent over a large enough sample. So you can figure out what a pitcher's or hitter's overall BABIP should regress to based on their ratio of ground balls to line drives to fly balls.
As for traditional stats vs modern/predictive stats. I like the modern ones better, and I love to use the to figure out how good a player is, how they relate to other players, and how they should perform in the future. But when it comes down to awards, I believe they should be based on actual performance and outcomes, not what should have happened or what could happen. That's how the actually wins and loses work. The World Series Champion isn't always the actual best team in the league, but they are still the World Series champion. The winningest team isn't even always the best team, but they were the winningest. Albert Pujols is the best hitter in baseball, but that doesn't mean I'm handing him the MVP every year. Some years luck and outcomes might not go his way while they do go another players way. I like to select my award winners are what actually happened.
Again, I think your larger point is a totally fair one. But, again, I must disagree with you about the BABIP stuff. Look at the data you referenced for the three different hit types and you'll see what I mean. Fly balls have an expect BABIP about 100 points lower than grounders. Obviously, liners are the surest way to get a hit, so a pitcher who prevents them (be it by inducing lots of grounders or however) will have a lower BABIP. When you compare the average fly ball pitcher to the average ground ball pitcher, though, the former will have a lower BABIP.
Hopefully this resolves your dispute with Paul over pitcher BABIP and groundball/flyball rates: "Groundball rate is somewhat correlated with BABIP (positively) only because flyballs that are not homeruns are more likely to be turned into outs than groundballs. The correlation is not all that high (.11), and is barely far enough from zero to assume there is a real correlation." (Taken from Matt Swartz at Baseball Prospectus, ) So Paul is more or less right (not that your other points aren't valid or interesting).
Hey, isn't Matt Swartz just a BP Idol contestant at this point ;-) That was the only article that I voted on this week though, it was pretty damn good.
As for the BABIP stuff, I guess I fell to the small sample size monster of just looking at a few examples of extreme ground ball pitchers.
Happens to the best of us, Adam. I appreciate the back-and-forth.
Why does everyone seem to leave Chris Carpenter out of the Cy Young race?
Because he's been injured and has pitched all of 38 innings so far. If we are going that route, we may as well talk about Jonathan Broxton and a half dozen other relievers as well.
Yeah, this is strictly for the first third of the season. There's still a chance that Carpenter could factor into the race before year's end, but he's already spotted the top pitchers 40 innings.