In the first of this 3-part series of FanPosts, I plotted how several key batted/nonbatted ball stats have changed over the course of the professional careers of Bud Norris, Jordan Lyles, and Lucas Harrell. In the second, I graphed how those same stats have changed for John Ely, Philip Humber, and Alex White using the data of Norris, Lyles, and Harrell as frames of reference. So now let us look at 3 more righthanders who have limited to no experience in a big league rotation but would stand to be decent bets to enter the Houston Astros rotation at some point during 2013: Jarred Cosart, Brad Peacock, and Kyle Weiland.
These 5 stats were computed for each minor league season (if > 150 batters faced) and major league season (if > 200 batters faced) of the player:
- BB%: walks divided by plate appearances against
- K%: strikeouts divided by plate appearances against
- SLG on batted balls: this is slugging percentage against on non-bunted batted balls. It is being used as a measure of the loudness of contact against the pitcher.
- LD + OFFB %: this is the percentage of non-bunted batted balls that are classified by the official online game recaps of mlb.com and milb.com as line drives or outfield flyballs. Note that some of those line drives and flyballs are hits, some are outs, a few are both, and a few are neither (e.g., reached on error). This stat is being used as a measure of risk for louder contact, given that 93% of extrabase hits are of these batted ball types.
- OFLD+OFFB to Pull-Field % : this is the percentage of non-bunted batted balls that are classified as outfield line drives or outfield flyballs hit to the batter’s pull-field third of the outfield. This stat is being used as a second measure of risk for louder contact, given that about 66% of home runs and over 40% of line drive or flyball doubles are hit to the pull-field third of the outfield.
The player data presented was compiled from the MLB Advanced Media online folders or from the Retrosheet archival files that reference the same. League averages and such were computed using Baseball Reference, unless specified otherwise. As in the previous posts, players who spent time at 2 minor league levels in a season have that data plotted over the point of the x-axis that reflects the relative percentage of batters faced at each. And again the lines connecting the markers move in chronological order.
Non-batted Ball Data
After very good A-ball debuts, Cosart and Weiland have each seen their control waver posting BB% figures in the 9% to 10% zone in each season since, or roughly 2 to 3 percentage points above the 7.4% mark that 2012 MLB starters posted. Peacock has had 3 minor league seasons within 1 percentage point of that 2012 MLB starters’ BB% reference datum, with his first and most recent BB% marks closer to 10% and 11%, respectively. Lyles’ consistently good performance in the BB% category stands out relative to the others, casting him in the role of rotation control artist.
Much like Lyles before him, Cosart’s K% took a big tumble of almost 10% in the season after Low A ball, and the small spike upward in the 2012 AA-biased season has him checking in now at 18% which would be a hair lower than Lyles’ mark for his AA-biased season. Weiland also showed a near 10% decline in jumping directly from short-season A to High A ball but has seen his K% numbers rise by a few percentage points in the 2010 AA and 2011 AAA seasons since, to the point that Weiland’s AA and AAA K% data are virtually equivalent to that which Norris posted when he was at those 2 levels. Peacock’s progression provides a stark example of why one shouldn’t overreact to a high school draftee posting low K% figures at the lowest rungs of the minors. Since having K%s in the neighborhood of 16% to 17% in the 2008 and 2009 A-ball seasons, Peacock has since posted marks of 25%, 30%, and 22%. That 30% number at mostly AA in 2011 would be the highest in any season by any of the 9 righties reviewed so far in this series. After beating Norris’ AA 24% K% handily that season, Peacock essentially posted an identical K% in 2012 to the 22% one that Norris posted in AAA at 2009. Though neither Peacock nor Weiland have received a lengthy turn within an MLB rotation, each looks to have some potential to be an above average starter at the strikeout and perhaps even a well-above average one as Norris has been in his 4 seasons an Astro. Cosart’s 2012 K% was almost exactly on the number posted by MLB starters in 2012.
Batted Ball Data
SLG on Batted Balls
Cosart has clearly been both remarkably consistent and remarkably good at stifling power in his 3 full pro seasons; his three values of SLG on Batted Balls vary negligibly from a low of .422 to .429 as a high, and would each rate almost .100 points under the .516 figure that MLB pitchers put up in 2012. He’s also incredibly similar to how Harrell rated at the SLG on Batted Balls stat at equivalent levels of the minors. Weiland’s also been very good at the stat, bettering the 2012 MLB Starters’ number by quite a bit in 3 of his 4 seasons and equaling it in the fourth. Peacock has been more erratic, twice rating much better than that mark and twice rating much worse including last season when his SLG on Batted Balls was .613 for Oakland’s AAA affiliate. Peacock’s plot looks very similar to Norris’.
LD + OFFB %
The next graph shows one reason why Cosart’s SLG on Batted Balls data has been so low, namely that he’s been exceptional at avoiding line drives and outfield flyballs. After a modicum of a spike in 2011, Cosart’s LD+OFFB% fell all the way below 32% in 2012, which netted him the top spot among AA-biased starters beating their mean by almost 3 standard deviations. Cosart is clearly living in the same exlcusive neighborhood as Harrell was and is at this stat, perhaps even in a nicer residence. Weiland’s first two seasons of pro data is not too dissimilar from Cosart’s, but since then Weiland's LD+OFFB% has gone up a bit more though he still has the look of a pitcher who could beat the rate of MLB starters as a group. Peacock’s percentage has continued to rise with each successive season, with his most recent 54% PCL figure being a full 7% above the 47% number of 2012’s MLB starters (per StatCorner).
Fitting a regression line to a plot of the 6 pitchers’ SLG on Batted Balls versus their LD+OFFB% shows that while the latter generally does a respectable job of predicting the former, Peacock’s most recent 3 seasons of slugging data haven’t particularly lined up well with his line drive and flyball rates and especially so in 2010 and 2011.
OFLD + OFFB to Pull-Field %
This stat would be the other explanation as to why Cosart has done well at limiting slugging as he’s been under 6% in 2 of his 3 full pro seasons. That basically means that in those 2 seasons you would generally have had to witness 16+ batted balls against him just to see one hitter pull a flyball or line drive to their pull-third of the outfield, and that would typically amount to one such event per start given the number of batted balls per outing. Here again Cosart seems be the equal of Harrell and potentially better. While Weiland was similar to Cosart early on at LD+OFFB%, here there’s plenty of separation between the two at each level of competition. Peacock rates a little worse than Weiland does at each level and this is mostly an offshoot of their disparate LD+OFFB% data, as each has had a similar percentage of their outfield line drives and outfield flyballs going to the pull-field third of the outfield at each minor league level (30% to 36%, typically, as shown below). As a frame of reference, Weiland’s 13.5% OFLD +OFFB to Pull-Field % from 2011 would probably rate a percentage point or so worse than average for MLB starting pitchers. Cosart has been much tougher to pull aerially than the other two, only allowing 18% to 23% of his outfield liners and flies to be hit to the pull-field in the last 3 minor league campaigns.
Cosart’s batted ball stats to date may well have already earned him of the unofficial title of the minor leagues’ best starting pitcher at inducing weak contact, especially given the advanced level of competition that the most recent of it was compiled against. Cosart looks to be the spitting image of Harrell in terms of the stats evaluated here and he conceivably could be still better at the batted ball ones. And that would be something in so much as Harrell was among the best handful of major leaguer starters at those sorts of numbers in 2012. And an MLB average K% or perhaps even a slightly higher one is certainly within the realm of possibility given Cosart’s 94 to 98 mph fastball velocity and the potential for more intensive coaching directed at finishing off hitters in 2-strike count situations. That Cosart now rates a few percentage points worse than what would be average for a MLB starter in the walk rate department would not figure to be a major area of vulnerability given how weak the contact versus him projects to be. Financial considerations, among other factors, should keep Cosart out of the Astros rotation for the first month or so of 2013, but he seems destined to force his way into the rotation by the All-Star break if not sooner.
Peacock seems very similar to Bud Norris the minor leaguer on the metrics evaluated here and as such it wouldn't be unreasonable to project a Norris-like outcome for the early years of Peacock's major league career. The area where the two would seem to differ the most may be in the rate of line drives and outfield flyballs, as Peacock has surrendered those batted ball types about seven percentage points more frequently in the last 2 seasons than Norris did when he toiled in AA and AAA. As is true with Bud, the adverse consequences of Peacock's probable higher-than-typical line drive plus flyball rates as a major league starter could be exacerbated by worse-than-average walk rates and mitigated by better-than-average strikeout rates.
Given the shoulder pain that disabled him just three starts into his first full-time gig in a major league rotation last April and the dominoes that fell after an infection developed in the joint as a complication of treating the injury, Weiland doesn’t figure to show up in the major league or even a minor league rotation for at least a few months. He still qualifies as a rookie by official MLB standards and thus remains a prospect in the technical sense. Heading into his abbreviated 2012 season, Weiland had shown some potential to be better than MLB average at batted ball outcomes and strikeouts and as such it would be premature to dismiss his future chances of sticking in the majors as a starter without seeing more of a healthy him on the mound.
Since this post effectively concludes the series looking at the upper echelon of righthanded candidates for the 2013 rotation, which of the following seven (beyond Harrell) has the best shot to be one of the Astros top 2 performing starters over the 3 seasons spanning from 2014 to 2016 (exclude 2013): Humber, Lyles, White, Ely, Weiland, Peacock, or Cosart? Cast your vote in the formal poll appended below.
Which one of the following is most likely to be among the Astros' top 2 starters statistically from 2014 to 2016?
Philip Humber (0 votes)
Jordan Lyles (37 votes)
Alex White (8 votes)
John Ely (1 vote)
Kyle Weiland (1 vote)
Brad Peacock (3 votes)
Jarred Cosart (62 votes)
112 total votes