Statistical Metrics II – Pitching

Building on last week's discussion on the various statistical metrics we use to describe and value the production of batters, it's about time to look at the other side of the game: pitching.

Because there are so many variables to consider when valuing one pitcher versus another, pitching metrics are always a bit touchy. Most metrics will love traditional aces, what with their high strikeout potential, but others will differ on how good groundballers like Jason Marquis and Aaron Cook are. I hope we can shed a little light on what all the numbers and acronyms mean once again, and show those of you unfamiliar how we use them properly.

To begin, we'll kick it old school. Wins.

Now Wins are kind of a fun way to look back at a season and say "this guy did a good job for his team" after it's all over. The strength of the pitching win is that they can be a good way to reward a pitcher for an excellent performance. The downside is that they are an "ends justify the means" kind of stat to keep. For example, Ubaldo Jimenez pitched 8IP, struck out 9, gave up 2ER, and got the Win. However, the day before, Jorge De La Rosa pitched 6IP, gave up 6ER, and also got the Win.

Wins definitely correlate over a few seasons with the best pitchers, as the best tend to not allow many runs and pitch late into games. The issue with wins is that they tend to over reward pitchers who can pitch a lot of innings. My favorite example is Livan Hernandez, who tends to finish seasons with double digits in the W column, but is a noteworthy hit machine and a generally awful pitcher. Hernandez tends to get rewarded because he can hang on a long time and give his team a chance to crack into the opposition's bullpen.

Wins also zero predictive value from season to season. At best, they are a performance stat.

We can debate wins and losses all day, but that's a very brief reasoning on why I look past wins and losses as a measure of a pitcher's value.


Join us after the jump as we step into a pile of other pitching metrics.

ERA is the next logical metric to look at, and a very decent one to use in general. The upside of ERA is obviously that it gives you a good idea of how many runs a pitcher will allow in general. The downside is that it can over reward pitchers with an excellent defense or a majestic pitcher's park. ERA can tend to reflect the team's overall run preventing ability, with an obvious glut of the responsibility on the pitcher. Now, before you start typing your response on how pitching to your defense is a skill in itself, we'll talk about ERA discrepancies later on down the line.

An average ERA is 4.50, typically, while the MLB average this season is 4.35. Sub 3.00 and you're an all-star. Sub 2.50 and you're elite. Above 5 and you're a back end starter. Above 6.00 and you really shouldn't be pitching in the majors.

We all get ERA, and it's basically the easiest to just cite and not be too far off of base. You're not gonna see too many 2.00 ERA pitchers who are claimed to actually suck and 6.00 ERA pitchers claimed to actually be studs. Nobody should harp on anyone for citing ERA. But there are better metrics to view as we move forward.

To dial the metrics up a bit, there's also WHIP to take into consideration. WHIP is obviously just Walks + Hits / Innings Pitched. That's it. It's just a measure of the traffic a pitcher allows. And obviously, more traffic = bad. Average WHIP is 1.391. Whip below 1.2 is all-star level. WHIP below 1.1 is elite. WHIP below 1.00 is usually reserved for Mariano Rivera.

Before we touch on the more SABR slanted metrics, let's take a look at component stats. These are where the skills of a pitcher are really quantified.

K/9, BB/9, and HR/9 are the basics of the component stats, and they're complemented by GB% and FB%. Additionally, K/BB also works well with them, as it can provide context for a high BB/9 or a low K/9.

Now in terms of where these numbers fall in terms of where the league pitches is as follows: 6.9 K9, 3.5 BB9, 1.1 HR9, 2.01 K/BB. K/BB might be the more important thing to look at with a lot of pitchers, as a low K/9 complemented by a low BB/9 provides a solid K/BB, and you can rationalize a high BB/9 if your pitcher also carries a high K/9.

So rather than trying to put levels on K/9 and BB/9, we'll talk about K/BB instead. As mentioned above, 2.00 is about average. 1.50 is passable. 1.00 and you're probably not MLB caliber. 3.00 and you're a pretty dominating pitcher. 4.00 and up and you're downright elite.

Now the one component rate I don't lump up there is the HR/9 component. Because HRs in general are such a park-affected part of baseball, it's hard to just look at a HR/9 and say "good" or "bad". You kind of need to look at it based on your team and the environment you play in. As Rockies fans, I view a 1.00 as the line for someone really worth pursuing - in terms of starters at least. 2.00 is right out. 1.50 is pretty undesirable. A HR/9 of .70 or below is a solid pitcher.

So all these component numbers lead to a couple of things: firstly, you can use them to paint a picture of the pitcher in question, and secondly, they lead into the first of 2 SABR style metrics I'm going to review here.

FIP, for those SABR types following, is the logical followup to the component statistics. FIP is basically combining your K/9, BB/9, and HR/9 into one nice metric which mirrors ERA directly. An average FIP is the same as league average ERA, by definition. Now what FIP does is that it takes the balls in play out of the situation entirely (as a Home Run isn't really in play, fielders can't do anything to affect them) and rates a pitcher just based on what he does to the batter. Counting Rocks covered FIP a little while back, so you can get some more information on it there.

To really look at FIP, you need to compare it to ERA. For example, Matt Cain is sporting a 2.43 ERA right now - which is obviously excellent. However, Cain is backing that 2.43 ERA up with a 3.91 FIP. This isn't to say that Cain is pitching at a 3.91 ERA level, but it suggests that his strikeouts, walks, and homers are good enough on their own to put up a 3.91 ERA....does that make any sense? Don't look at FIP too literally, because for guys who pitch to contact a lot, it misses a major portion of their game.

How you need to look at FIP is really in the difference between FIP and ERA. In Cain's case, it would suggest that Cain is reaping the benefits of a good defense, a pitcher's park, and/or some good luck. Am I trying to say Cain sucks? Absolutely not. But what's interesting about Cain is that his past 4 seasons of FIP are (starting with 2009 and rolling backwards): 3.91, 3.91, 3.78, 3.96. Cain is pretty much the same pitcher this year as he has been the past 3 years, but this season, things are clicking better for him. Counting Rocks ran an analysis of Matt Herges in this same fashion, if you want to see a bit more of how we use components to evaluate a pitcher.

There is one extension to FIP, and that is xFIP, as presented by The Hardball Times. For all intents and purposes, xFIP is the same as FIP, but it normalizes the HR component of FIP to put everyone on a bit more of a level playing field. It should also be noted that xFIP has a pretty decent predictive value from year to year.

Moving forward, we have one final metric to discuss, and then I'll set you all loose to go all SABR on everyone out there.

tRA, again, covered by Counting Rocks, is yet another "looks like ERA" metric that has the same idea as FIP, but it actually takes regular balls in play into account in how it is scored. Basically, every event that happens in a game is worth a certain number of runs and a certain number of outs. What tRA does is sums up the number of "expected" runs and "expected" outs and then calculates it on a scale of 27 outs, or 9 innings. So basically, a ground ball is worth, say .90 outs, and a single is worth .75 runs. tRA is essentially the pitching equivalent of wOBA.

Another way to look a tRA is that it is a "report card" of how a pitcher did. When you see a pitcher who had an alright game, got a lot of ground balls, but also had a lot of balls squirt through and dying quails to no-man's land and such, but really pitched a better game than the final box score suggests, tRA will give them credit for the good things they did while appropriately penalizing them for the bad things they do. tRA, as compared to ERA, can do the same kind of thing that comparing ERA and FIP can do, but it will also take the hits and such into play.

For example, I have a guy who strikes out 4.5 guys per 9 innings, walks 2.5, and gives up .80 HR. FIP is gonna say this guy is pretty good. But what FIP misses is that the guy has a WHIP of about 1.70 because he's incredibly hittable in the meantime, gives up a pile of singles. tRA is gonna punish him for all those singles.

So to summarize all the fun we've had today, let's go ahead and check out a chart real quick:

 

Elite

All Star

Average

Back-End

Adam Eaton

ERA/FIP/tRA

2.50

3.25

4.50

5.25

6.00+

WHIP

1.00

1.10

1.40

1.60

1.75

K/BB

4.00+

3.00

2.00

1.50

1.00

 

So that's it for this week, fair RowBots, I hope you have a good grasp of the advanced metrics now so when the statheads get on a roll, you're not left in the dark!

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Purple Row

You must be a member of Purple Row to participate.

We have our own Community Guidelines at Purple Row. You should read them.

Join Purple Row

You must be a member of Purple Row to participate.

We have our own Community Guidelines at Purple Row. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9351_tracker