Since I have a bunch of time on my hands in this month, being that I'm off from college, I've been compiling a leaderboard of what I call "True ERA." The formula for that is pretty simple:
True ERA = FIP + (xFIP - FIP) + (.92tRA - FIP)
How did I come up with this, you ask?
Well, FIP is widely regarded as a sort of "true ERA" since it takes out all luck. xFIP (Wow, I just started a sentence with a lowercase letter, and it was correct) is FIP with a HR/FB adjustment, so we can say that (xFIP - FIP) is simply a "homer adjustment."
So, if a pitcher's xFIP is six but his FIP is five, we can say that he should have actually allowed an extra run per nine innings because his HR/FB rate should have been higher.
tRA adjusts for batted ball splits. It is on a "Run Average" scale, as opposed to an ERA one, so since approximately 8 percent of runs are unearned, to get tRA on an ERA scale, we have to multiply it by 0.92. Any "tRA" figures you see in my articles have this .92 coefficient.
Anyway, we can call (.92tRA - FIP) a "batted ball adjustment."
So, practically, if a 4.50 FIP pitcher had a 4.20 xFIP and 4.10 tRA, his True ERA is actually 4.50 + (4.20 - 4.50) - (4.10 - 4.50), which comes out to 3.70—lower than any of the three metrics individually.
This is useful for two reasons. First of all, it puts the pitcher's true ability into context, like ERA is traditionally said to. Second, it can be used to determine how lucky a pitcher is. Doing so is easy:
Luck = ERA - True ERA
As I said, I'm currently compiling True ERA and Luck data, and I have to enter the tRAs manually (for some reason, tRA isn't sortable on Fangraphs), so while I'm entering the data over the next few days, I'll be looking at some of the anomalies I find.
The first one I've seen who has a huge Luck number is Giants righty reliever Brandon Medders.
Last year, Medders went 5-1 with a 3.01 ERA in 68 and two third innings, certainly a solid middle relief season on the surface.
FIP takes issue with Medders unimpressive 1.81 K/BB ratio, and says he's only acceptable, not above-average, with a 4.07 mark. An elevated strand rate (80.4 percent) and a just-slightly-lucky BABIP of .296 knocks Medders down a run.
However, xFIP takes issue with the righty's low 7.7 percent HR/FB rate, likely a function of pitching half his games in San Francisco's spacious park, and many road games in San Diego and Dodger Stadium. Medders' xFIP is 4.49, which, for a middle reliever, is rather poor.
tRA delivers the big whammy on Medders' 2009; batters made hard contact very often against the former Arizona reliever. A whopping 26.9 percent of batted balls off Medders were line drives, which puts his xBABIP (Expected BABIP) much higher than the .300 range that BABIPs typically fall.
Medders' BABIP should actually be in the mid-to-upper-.300s, which makes his .296 figure, while it's in the typical "normal" range, look very lucky.
Medders' tRA is 5.10.
Using my True ERA formula on Medders, we get:
(4.49 - 4.07) + (5.10 - 4.07) + 4.07 = .42 + 1.03 + 4.07 = 5.52
That's 2.51 runs of luck over Medders' 3.01 ERA—by far the highest number in the 34 entries I've looked at thus far.
While ERA says Medders did a nice job, True ERA says he should be headed back to the minors in 2010.
Medders' career True ERA of 5.33 is actually 1.97 runs above his career 3.36 ERA. With a career K/BB ratio of just 1.70 and only an average homer rate, coupled with a high line-drive rate of 21.7 percent, I'd expect some serious regression from the right-hander in 2010.
Luckily, the Giants have an extremely deep pitching staff, so if Medders does indeed struggle, it shouldn't hurt the team too much.