Dillon Gee, A Year of Fortunate Events?

Gee rehabbed back at his first pro stop in 2014

Photo credit: Kathleen Malone-Van Dyke

Dillon Gee was an exceptional pitcher from the 2^nd half of 2013 through the start of the 2014 season, producing an ERA- that would have ranked top 10 in the majors among qualified pitchers in 2013-14, but has been below average by ERA- outside of that stretch, so which Gee should we expect now that he appears likely to take Zack Wheeler’s spot in the rotation?

            Whenever I think about Dillon Gee’s career, I always think back to those SNY commercials from when he was tearing it up in 2007 during his professional debut with the Brooklyn Cyclones – “Fans! Come out to get a glimpse of the future Mets, watch Dillon and Dylan pitch for your Brooklyn Cyclones!” The other Dylan was Dylan Owen, a righty born about 6 weeks after Gee and drafted one round before him. And Dylan Owen would have the better season with Brooklyn that year, posting a better ERA, FIP, K%, HR/9, and a similar BB%. But Dylan Owen would flame out at the highest level of the minors in 2013, while Dillon Gee has compiled a 40-34 record with a 3.91 ERA over 106 games (103 starts) and 639.2 IP for the Mets since his first call-up in September 2010. As a side note, there was another current Mets player on that 2007 Cyclones team – Lucas Duda would have a strong pro debut at age 21 too.

Although his W-L record has been good in the majors (.541 winning percentage), and he’s provided the Mets with a valuable total of innings in multiple seasons, other metrics suggest he has been a below average pitcher – for example, he has a career 108 ERA-/117 FIP-/109 xFIP-.  In that sense, his success has been similar to Jason Marquis so far – Marquis has a career record of 121-114 (.515), with 1921 IP in 368 games (309 starts), and below average 108 ERA-/115 FIP-/108 xFIP- rates. Throughout their careers (until recently for Marquis, who is 8 years older) they have generally been what teams expect from the back of the rotation – they will win the team a few games with random great starts, lose the team a few games with random terrible starts, but mostly just keep things close enough that the team has a chance to win it, and they have a chance to get the W. Marquis has had a few stretches where things broke well for him and the results were much better than both his peripherals and career rates, but when you play the game as long as he has, you see both sides of that coin, and Marquis has also had some stretches where things didn’t break his way and he was among the worst starting pitchers in the league. There is no guarantee how Dillon Gee’s career will turn out, but there is a good chance that, just like Marquis, he’ll continue to have his good stretches and his bad stretches.

With the Mets starting rotation in a near constant state of flux after John Maine and Oliver Perez bombed at the start of the 2010 season, I was among those rooting for the Mets to give Gee his first chance before September – he was a younger option (24), having a strong season in Buffalo (ended up leading the league in strikeouts), and, at the very least, more interesting than most of the arms being given starts (guys like Pat Misch, Hisanori Takahashi, Raul Valdes, and Fernando Nieve). But Gee wasn’t supposed to be good when he got to the Mets, not according to the scouts who said he merely had good control of three below average pitches and a consistently high-rated changeup – it was ranked the best changeup in the system by Baseball America 3 times. While his 5 starts that September were a little shaky (12.5 K%: 11 BB%), he had a little luck (.225 BABIP and 80.7 LOB%), which led to a strong 2.18 ERA, and a 2-2 record. Then, despite below average metrics in 2011 (120 ERA-/125 FIP-/116 xFIP-), Gee would finish the season with a 13-6 record over 27 starts, thanks in large part to a strong first half. As expected, his K% stayed below average in the majors during 2011, but, unexpectedly, his BB% did too (10.1%). After a rough start to his 2012 season, Gee had turned things around heading into the all-star break, but he’d have to shut it down for the rest of the season with a clot issue in his shoulder. At 6-7, his record wouldn’t look as nice in 2012, but the metrics liked him more, as he finished with a 110 ERA-/100 FIP-/91 xFIP-, thanks to an improved 21 K%: 6.3 BB% (and better run prevention). Which brings us to 2013, which started terribly for Dillon, but turned around after a brilliant start against the Yankees on May 30, during which he struck out 12, walked none, allowing only a solo homerun and 3 other hits.

When Gee came to the mound on the night of May 30^th, he had a 6.34 ERA/4.92 FIP – aka, 179 ERA-/138 FIP- – over 10 starts (49.2 IP, 69 H, 8 HR, 15.9 K%: 7.3 BB%), and the Mets had Zack Wheeler on the way, so he was pitching for his spot in the rotation. Gee had been frustrated with the results to that point, saying the following after a loss to the Cardinals two weeks earlier.

“I don’t know what to say, I didn’t feel all that bad tonight. I’m a ground ball guy and I think seven ground balls got through. I don’t know what else to do.”

With a 46.2% groundball rate through the Yankees start, he was pretty average in that department, but he wasn’t wrong to be feeling a little unlucky through that stage of the season, as his .365 BABIP allowed and 68.5 LOB% were below major league average and his career rates, which you can see in Table 1 below.

Table 1 – Dillon Gee’s career stats in 3 groups: i) before that start against the Yankees, ii) over the next calendar year, and iii) after he returned from the DL in 2014.

Time	IP	BABIP	HR/9	LOB%	ERA-	FIP-	XFIP-	K%	BB%	LD%	GB%	FB%
9/7/10-5/25/13	353	.290	1.02	71.0	119	118	109	17.2	8.6	18.9	48.1	33.1
5/30/13-5/10/14	202	.259	0.98	83.3	77	109	108	16.9	5.5	19.3	41.1	39.5
7/9/14-9/25/14	84.2	.294	1.28	72.1	138	132	112	17	7.8	17.6	46.1	36.3

             First, how a few of Gee’s rate stats from the 2^nd group would have ranked among 79 qualified pitchers across 2013-14 had these been Gee’s only stats: t-9^th in ERA-, 3^rd in BABIP, 1^st in LOB% (Yu Darvish would’ve been 2^nd at 81.4%, one of 4 pitchers who actually had a 80+ LOB% during 2013-14), t-19^th in BB%, t-62^nd in K%, t-59^th in FIP-, t-61^st in xFIP-.

            Comparing the 2^nd group with the groups before and after, the most noticeable differences that would lead to such a big improvement in ERA- are the decrease in BABIP and BB%, and the increase in LOB%. The decrease in BABIP against can be explained partially by the combination of Gee allowing more fly balls without giving up more homeruns, and the Mets using the shift more over the past two years – except Gee’s BABIP returned to his previous career average after returning from the DL last season, even though the Mets would have been shifting just as often. The decrease in BB% seemed natural for Gee, who had generally posted elite BB-rates in the minors, including a 5.9 BB% in his final full minor league season of 2010, and had taken a big step forward in that area in 2012 before his injury. Both of these factors could have helped increase his LOB%, but Gee went from average to elite and back by LOB% standards across the 3 groups, and that was the biggest driving force in his great ERA improvement. If you substitute in his career LOB% during that stretch, then his ERA jumps up to about 3.55, which is slightly better than the 3.62 ERA that earned him a 102 ERA- in 2013. I looked at all of this last year, and considering how much better Gee’s LOB% was during group 2 than his career rate, and that it was two percentage points better than the next closest pitcher across 2013-14, it seemed likely to me that Gee was in for a drop in performance soon. I had actually started to write something about Gee’s statistical outliers at one point last year, but he was injured and due to be out for a while, and then I got swamped behind minor league coverage and pushed it aside.

When I was writing this piece last year, I started out by looking at Brooks Baseball to see if he had any major changes in approach or pitch results during this stretch – another reason I pushed it aside is because I had too much data from Brooks, so I am just going to summarize the main points here. Against lefties, Gee basically kept the same approach. Against righties, Gee started throwing his slider nearly twice as often (increased from 14 to 26%), especially when ahead, and entirely at the expense of his changeup (dropped from 17 to 5%). His slider had been his best pitch by average and ISO, so it makes sense that he’d want to throw it more. What’s interesting is that Gee’s slider usage against righties returned to 15% when he got back from injury, although his changeup usage only rose to 11%. Perhaps Gee also looked at his Brooks profile page and saw that hitters had been teeing off some on his slider during his hot stretch, while his changeup and curve had been getting much improved results. While his slider had been a reliable groundball pitch in the past (49.5% GB rate in group 1), it has allowed more line drive and fly balls since that fateful start against the Yankees, and has had a sub 40% groundball rate since. His sinker also got less groundballs during the streak, but it bounced back in that area after his return from the DL last year. Finally, the Brooks zone profiles are useful because they are like hot zones for pitch locations. Looking at Gee’s zone profiles, his hot zones are on the outer half to both lefties and righties, but he was in the outer third of the strike zone more frequently during his hot stretch than before and after, when he threw more pitches out of the zone.

In review, Gee walked fewer, and benefitted from a lower BABIP and higher LOB%, which helped lead to a huge improvement in ERA- for a calendar year. Gee may be capable of replicating his improved BB% moving forward, but it’s unlikely the other rates return to what Gee did during his hot streak. Players are constantly adjusting their approach to the league, but Gee didn’t make drastic changes to his approach during his hot stretch, he just started throwing one pitch to righties more at the expense of another. He did benefit from improved results from two pitches (changeup and curveball) per Brooks, but saw the results from his slider and sinker go the other way. Before and after his hot stretch, Gee produced like a 5^th starter by ERA-, FIP-, and xFIP-, while during his hot streak he was a #1 by ERA-, and a #4 by FIP- and xFIP-. FIP and xFIP are better indicators of future performance than ERA, and both have stayed in the 4^th/5^th starter range in three of Dillon Gee’s four full seasons (2012 was the lone exception), which is fair to expect from Dillon Gee moving forward. Although he is perfectly acceptable at the backend of most MLB rotations, he is no Noah Syndergaard or Steven Matz, so even though I’ve been a Gee fan who’s been following from the beginning, I can’t wait for the Mets to replace him in the rotation with the younger and (hopefully) better models.

What Are Gamma-Rays? http://t.co/2GeI8Dlys8 vía @LiveScience useful both in health field as in astronomy pic.twitter.com/oJwCg4yZB6
— Beyond The Matter (@BeyondMatter) March 21, 2015