Last Friday, Carlos Martinez faced the Mets. In that start, he gave up seven hits, five runs (all earned), and two dingers in five innings. His start prior to that was against the Nationals. In that start, he also gave up seven hits, five runs (all earned) and two dingers. It was his fourth start giving up exactly five earned runs this season, and he’s gone exactly five innings in all four. He’s given up two home runs in three of those four starts, and seven hits in three of those four starts. Not the kind of consistency we want.
In fact, these four starts have led to rumbling about Carlos Martinez inconsistency, immaturity, and hair. Those starts have led some, including myself, to question whether Carlos Martinez is an ace or just a number one (if there’s even a difference).
Traditionally, the ace is viewed as a stopper. Someone the team can always count on to end a skid or win the important games. If Martinez is inconsistent, maybe he’s just a great number two pitcher and not the head of a staff at all.
However, I think it’s fair to wonder whether Carlos Martinez is actually more inconsistent than his peers. I dug into the numbers to find out, using game log data available on Baseball Musings (via RetroSheet) and season stats available on FanGraphs.
I defined Carlos Martinez peer set as the fifty best starting pitchers by FIP (Fielding-Independent Pitching) with at least 200 innings pitched since 2015. Next, I calculated all of their game ERAs, FIPs, WHIPs, innings pitched per game, and quality start percentage. Five solid, results-based metrics to evaluate pitcher consistency without relying on granular pitch-by-pitch data.
*I compiled the peer set data on July 7th and included all games up until July 6th. Additionally, I included Carlos Martinez latest start, but did not update the data set to include other games from July 7th thru 9th.
First, let’s look at quality start percentage. To earn a quality start, a pitcher must go at least six innings and give up three runs or less. Those requirements might not seem like much, but from 2007 to 2016, the St. Louis Cardinals had a .772 winning percentage when their starter made a quality start.
Since becoming a full-time starting pitcher in 2015, Martinez has made a quality start in 52 of his 77 appearances. His 67.5% quality start percentage ranks 16th best among the qualified top fifty pitchers, sandwiched between Jake Arrieta and Gerrit Cole. His total of 52 quality starts made is tied for eighth most with Dallas Keuchel.
Next up, innings pitched per start. Carlos Martinez has averaged 6 1/3 innings per start since 2015. In this sample, that puts him in 20th place, just ahead of Chris Archer and right behind Jon Lester. Pretty good company.
Another way to look at innings per start is to take the standard deviation of innings pitched across all starts. A player who tosses nine innings one game and only five the next will have a higher deviation than a pitcher who goes six and eight. Through this lens, Martinez ranks only 33rd, but his 1.32 innings standard deviation is near the sample average of 1.27. Basically, he’s not noticeably more inconsistent than similar pitchers, at least in terms of innings per start.
Looking at WHIP, Carlos Martinez ranks below average in his peer set. His 1.28 WHIP is slightly higher than the peer set average of 1.22. Consistency-wise, though, Martinez WHIP standard deviation of 0.58 ranks 12th lowest (or 12th most consistent) in the set. So, while he allows just barely more baserunners than his peer average, he’s more steady across all starts.
Moving to the standard deviation of FIP, Carlos Martinez looks even better. His 1.83 FIP standard deviation ranks 8th lowest (or 8th most consistent). By this metric, Martinez is more consistent than Noah Syndergaard, Jake Arrieta, and Yu Darvish, to name a few. He’s also more consistent than Max Scherzer, Clayton Kershaw, and Chris Sale, to name a few more.
Finally, regarding the standard deviation of ERA, Carlos Martinez again ranks 8th lowest at 3.32 runs. If you’re worried that standard deviations aren’t the best way to measure consistency, maybe knowing that Adam Wainwright’s WHIP, FIP, and ERA standard deviations all rank in the highest ten (most inconsistent) will quell those concerns.
Interestingly, the ERA and FIP deviations are higher than I expected. The lowest ERA standard deviation belongs to Clayton Kershaw at 2.31 runs, and Madison Bumgarner is the only other pitcher with a deviation less than three runs. FIP, unsurprisingly, shows less deviation than ERA. James Paxton owns the lowest FIP standard deviation at 1.53 runs, and seventeen more pitchers are below 2.00 runs. Still, even just two runs of FIP is the difference between Brandon Finnegan and Jake Arrieta.
So what does that tell us? Essentially, that pitching is a high variance profession, even for starters. It’s unrealistic to expect even the best pitchers to give up three or fewer runs game after game. They have bad games. Sale, Scherzer, and Kershaw have each had starts this year where they allowed five or more runs.
Bottom line, Carlos Martinez is not an inconsistent pitcher. He’s a top thirty pitcher who, no matter how you measure it, rates out as equally or more consistent than his peers in the MLB’s top fifty starters.