Dillon Gee and the expectations of a fifth starter

"He kept us in the ball game" is a sentence I'm sure I've heard Ned Yost speak to the Kansas City media over a thousand times. That's a dangerous sentence because it implies two things - one, the pitcher did well and second, that he should get credit for what his offense does behind him.

"He kept us in the game" is usually only uttered when a pitcher gives up 4-5 runs. You don't hear it often when a team loses 1-0 because it's the offenses fault that those bums couldn't score. However when it happens in a 5-4 win, the manager credits the pitcher for not having a disastrous outing.

On Tuesday in Minnesota, Dillon Gee allowed three runs in six innings. That qualifies as a "quality start" (which is another bad metric - if it is even a metric) but his performance put the Royals down early, making another comeback necessary. A quality start at a minimum is six innings and three earned runs. That doesn't sound like much but it's a 4.50 ERA (going only by the two given stats). For what it's worth, the league average ERA this year is 4.22, so it would be a below average outing in that context.

That ERA also incorporates not only what the pitcher did himself, but what the defense did behind him.  If we look at Fielding Independent Pitching (FIP), which strips out the defense and gives us an idea of what the pitcher, Gee had an 8.98 FIP thanks to allowing three home runs and racking up only two strikeouts. By Game Score, a metric devised by Bill James to weigh individual pitching performances, he scored a 36, his fourth worst outing of the year by that metric.

Despite the FIP-based views, Ned Yost was happy with his starters performance and how Gee...you guessed it...kept them in the game.

Ned Yost on Dillon Gee's performance: "The adjustment he made was huge to keep us right there, in that ballgame. pic.twitter.com/7xMHlvyAwP

— FOX Sports KC (@FSKansasCity) September 7, 2016

A few days back Hunter Samuels of BP Kansas City wrote about the Royals "fifth starter" Dillon Gee.

The Royals signed Gee to a minor-league contract this winter, hoping he could provide some roughly average innings in the rotation and/or bullpen, as needed. The weren't expecting him to dominate. They weren't expecting him to be an All-Star. They were expecting average, and average is what they've gotten.

Gee has a 4.37 ERA in 105 innings this season. That comes out to an ERA+ of 101. Just above average. But the averageness doesn't stop there.

I agree with Hunter in the expectations of what the Royals were hoping for with Gee. They weren't expecting him to dominate, and yeah, maybe they were expecting him to be average. However, he has not been above average. He has not even been average.

The two flaws with his argument are (a) using Gee's full season innings pitched, including his time as a reliever; and (b) using ERA+. Gee has spent some time in the bullpen and it's of course important to delineate time spent pitching for a few outs vs time spent pitching for several innings. Gee as a starter this year has a 5.17 ERA and 5.23 FIP. He's been worth 0.1 fWAR. That's basically replacement level. Among qualified starters a 5.17 ERA would be fourth worst in baseball and his 5.23 FIP would be sixth worst. Now Hunter wrote his piece prior to Gee's most recent start Tuesday against the Twins, so admittedly I'm working with more up to date information, but it is not as if Gee had really been very good to begin with. Prior to Tuesday's start Gee was worth 0.2 fWAR.

Maybe call it a crusade against ERA but I am one of the more vocal opponents of the stat, including ERA+. It is truly an awful stat and even worse ERA+ doesn't even do what it says it does (if you want more of an explanation on this read these two articles).

If you want to use league average scalers then ERA- and FIP- are the ones to use. Gee has a 118 ERA- (18% worse than league average) which would be tied for seventh worst in baseball with Hector Santiago. By FIP- (which is a much better descriptor of a pitchers performance/talent than ERA-) Gee's 124 FIP- is fifth worst in the majors.

This is just a minor nitpick of course and the main emphasis should be on how Gee has pitched only as a starter when looking in context of him being a fifth starter. Gee as a pitcher this year has not been good, and he certainly has not been average. There is just no debate there whether you are using ERA, FIP, fWAR, bWAR, or any valuation metric.

On the Twitter, former Royals Review writer and all around good guy Clark Fosler and I got into a debate on Gee and what makes a fifth starter, in which he proposed this question to me:

@Shauncore @MichaelEHuck if the premise is what constitutes a solid 5th starter, then what statistical grouping would you use instead?

— Clark Fosler (@cfosroyalsauth) September 7, 2016

I honestly wasn't sure how to answer this. My first thought was "okay let's try to come up with some quick and dirty method." I didn't have anything at the time but thinking about it now, let's try to find some scale.

That's pretty simplistic but that could work maybe? Let's put it to the test with some names.

Of course I cherry picked some names. Guys like Zack Greinke are performing more like a #4 starter this year than an ace like he has in the past while Rick Porcello probably isn't  a top 10 pitcher in baseball. However I think the scale works loosely when it comes to expectations. You don't expect your fifth starter to be excellent but I think ~1 win from him is a reasonable expectations. It's not a perfect scale I'll admit and its virtue is in its simplicity.

My response to Clark though was that I didn't think there was a statistical group necessarily and I based that opinion off the idea of relativity. The fifth starter on the Dodgers going into the season was supposed to be one of Alex Wood or Brett Anderson (before Anderson unsurprisingly got injured for the 1,000th time in his career) with possibly Zach Lee waiting in the wings. ZiPS projected Wood, Anderson, and Lee to be worth 2.6, 1.9, and 2.2 WAR respectively. That's more of a number four starter (and a league average pitcher) than a proverbial fifth starter. Because the Dodgers rotation was (key word being was) so stacked, good pitchers (guys who would be at the top of the staff for the Royals and other teams) were pushed down into a fringe role or the bullpen.

The fifth starter on the Royals was supposed to be some combination of Chris Young, Kris Medlen, and/or Dillon Gee. I guess you could say Medlen was the clear cut number four so just focusing on Young/Gee, ZiPS projected 0.4 and 1.0 WAR respectively (though Young's projection was as a dual RP/SP swingman type). Needless to say the fifth starter role for the Royals was expected to be much less than one for the Dodgers talent wise.

That projection of Gee also lines up perfectly with the starter scale, and if Gee would have been worth one win this year then he would have been a fine fifth starter...but he hasn't been. Gee has been replacement level and unless you are expecting him to be worth one win for the rest of the year (which is roughly what Mike Trout is projected to do) then he's not going to meet his projection. Had anybody said at the beginning of the year that Gee was a decent fifth starter, sure, that's an argument you could back up. However it's clear now that Gee doesn't fit that role, unless your expectations are much lower than mine (which is fine of course).