Last week, I scratched the surface of the competition for the Astros' starting third base job by looking at traditional hitting stats and nontraditional defensive stats. A four-player Spring Training tryout at a position is fairly common, but this situation is unusual because not one of the players is the favorite, all have the possibility of starting the year on the 25-man Major League roster, and each candidate has legitimate questions about his ability or experience level.
- Brett Wallace has played little third since college, but has the highest prospect pedigree of the four.
- Chris Johnson finished near the top of the Rookie of the Year voting in 2010, but his performance backslid so drastically in 2011 that he lost his position.
- Jimmy Paredes played well in 2011, but his stats were boosted by unsustainable luck. He has far less professional experience than his competition.
- Matt Downs appears to have the skills needed, but has spent his career as a backup or platoon player.
Based on the big-picture information I explored last week, I chose Chris Johnson as the best option to start at 3rd in 2012. This week, I looked more closely at their past offensive outputs using the Pitch F/X tool. I particularly looked for what percentage of the pitches they faced resulted in a positive outcome. In this post, I've defined a Positive Outcome as the number of times a player reaches base on contact and/or knocks in at least one run.
Versus LHP/RHP Splits:
I first examined the most common split used in baseball to assess a hitter's ability by separating the pitches they faced from Right Handed Pitchers from pitches thrown by Left Handers. Note that in most cases in this post, I disregard the results for Paredes because the dataset is too small to draw any reasonable conclusions.
The data for Left Handed pitches shows that while Brett Wallace has the highest contact rate against them, he has the least success in creating a positive outcome. Johnson and Downs have a slightly lower rate of making contact when swinging against lefties, but are able to turn that contact into production with more success than their competitors. In very few opportunities, Paredes has shown such a low contact rate that his luck driven positive outcome rate can not be expected to continue.
In a larger sample size versus Right Handed Pitchers, Johnson and Paredes displayed the most success turning pitches into a positive outcome. In stark contrast, Downs retains the same contact rate as against LHPs, but the expectation of a positive outcome is lower than the other competitors. Thus, if Brad Mills wished to maximize his team's ability to create runs, by this data Downs should be a platoon hitter only.
Explanation of Positive Outcome Percentage:
This would be as good a time as any to explain the significance of these percentages. On the charts above, there does not seem to be much difference between a hitter who creates a positive outcome on 5% of the pitches he faces versus 9%. Both numbers are less than ten, and therefore the difference is insignificant, right?
No. The difference could mean as much as 30 to 50 RBI for that player over the course of a season, and therefore that many more runs scored for the team. (Below, PA=Plate Appearances, P=Pitches Seen by the Hitter. Runs Plated Per Positive Outcome is the percentage of times the player bats in at least one run. The Multiple RBI Factor is an approximation of the number of RBIs a player gets per RBI-generating pitch. In other words, I'm trying to account for pitches that yield two or more RBI.)
By extrapolating the data above, the possible improvement by selecting Johnson to face RHPs instead of Downs (for example) could net 10-ish more runs over the course of a season. That is a huge difference, particularly if one realizes that the same exercise can be applied to all positions on the field. Utilizing this type of data could make setting the daily lineup more challenging than it might appear on the surface.
Versus Pitch Type:
The fun thing about baseball for spreadsheet jockeys is that with the proper data, one can dig down to the seed level to squeeze out every drop of citrus goodness from the grapefruit that is a player's tendencies. The metaphoric grapefruit that I've chosen to examine below is the candidates' ability to hit pitches that are on the opposite extremes of the pitch-type scale. One the left end, pitches are relatively slow and move drastically along the azimuthal plane (look it up). On the right end are pitches that leave the pitcher's hand with triple-digit speed.
The first fruit under inspection is the "fast" fastball, which I'm defining as a pitch that leaves the pitcher's hand at a speed above 94 mph (or 252,672 furlongs per fortnight, if you are keeping track at home). The reason for the 94 mph cut-off is that any Triple-A hitter worth his salt should be able to hit a 90 mph fastball if it's thrown down the gullet. But conventional wisdom, sabermetrics, and our own eyeballs tell us that some hitters are unable to catch up with the really fast pitches, either because of bat speed or coordination issues.
The "Fast Fastball" chart above further illustrates why Matt Downs should be a part-timer (a finding that surprised me because conventional numbers paint him in a favorable hue. Now I learn that some of his advanced stats more resemble Rothko's 1968 "Untitled".) The chart also might help explain Wallace's struggles in the majors. Though the number of high-octane fastballs is dwarfed by the overall number of pitches thrown to a batter, Wallace's low positive outcome percentages in all three categories implicate his bat speed and/or coordination as possible areas of improvement. This is disappointing for a prospect with such near-universal high regard for his "hit tool".
The opposite side of the pitch-selection spectrum is the changeup, which can be 30 mph slower than the fastest of fastballs. Drastic changes of ball speed can befuddle even the best hitters.
Here we learn that Downs may just be a "slowball" hitter. He appears to be frustratingly inept against the fastest pitches, but skilled at adjusting to and mashing the slower ones. If one were to start a league where pitchers all threw between 70 and 85 mph, Downs might be your first draft pick.
Sadly, Wallace's numbers are not much improved over the fastball chart. He seems unable to catch up to power pitches, and also has trouble adjusting to changes of speed. Alas.
It's interesting to note that in both of the charts just shown, Johnson's Positive Outcome per Pitch percentage (say that ten times fast!) is a rock steady 8%. He can hit the power pitches, but he's also adjusting to changes in velocity. Superstar or scrub, that is a skill that is difficult to teach, and somebody should pat him on the back.
Finally, the breaking balls. Most major league pitchers will challenge young hitters with a slew of pitches that drop away from the plate or bore in on the batter because in the minor leagues, those hitters are rarely exposed to quality curve balls or sliders such as they see in the majors. Often it is failure to adjust to these knee-hemorrhaging pitches that prevents a top prospect from succeeding in The Show.
Here, we finally see a bit of improvement from Wallace, who raises his positive outcome to 7%. This chart also shows that Downs struggles with breaking balls more than his compatriots, while Paredes ties for the lead at 8% in a small data sample. Once again, Johnson holds steady at 8%. Fastballs, Slow balls, curvy balls...this guy seems to be able to adjust to them all. Interestingly, Johnson is able to maintain that high percentage despite a ghastly 59% contact rate. As bad as that contact rate is, he is somehow able to compensate and make a higher percentage of his contacts count than his competitors can. If Johnson improved his contact rate on breaking pitches, he could be come a fearsome hitter against them.
Finally, a quick glance at all pitches seen by the candidates in the major leagues:
Surprise of all surprises, Johnson's overall Positive Outcome per pitch comes in first, at 8%.On a positive outcome per pitch basis, Johnson takes the prize, weighing in at 8%. Looking back, Johnson has an 8% in every category of this study except for "vs LHP", where he still posts a competition-highest 7%. Wallace and Downs' 6% scores are dragged down by struggles against fast fastballs and breaking pitches.
It would be remiss to not mention that a significant component of offense is not included in the "Positive Outcome Per Pitch" stat: Base on Balls. So here are the candidate's Walk Rates for the major leagues:
- Wallace: 8.2%
- Johnson: 4.8%
- Paredes: 5.0%
- Downs: 8.2%
Wallace and Downs tie for first at 8.2%, which is a fine number but nothing special. Johnson and Paredes have some work to do if they wish to raise their BB% out of "bad" territory.
Based on the data above, I'm sticking with last week's pick that Chris Johnson should be the everyday starter at third base for the Astros in 2012. Johnson's consistency in being able to hit pitches of all shapes, sizes, creeds, races, religions, and galaxies of origin is impressive by any standard, and one hopes that his lousy 2011 was no more than a case of the sophomore jinx. If so, a good showing for the next couple years could make him a valuable trade chip when his later arbitration years draw nigh.
Paredes should be sent to Triple-A, in hopes that he can continue work on his hitting to show that he deserves the title of "3rd Baseman of the Future" and replace Johnson at a later date. He's young; he has time. Downs should continuein his role of sometimes-starter and situational pinch hitter, since breaking down his success by pitch type exposes his weaknesses against power fastballs and breaking balls. But what to do with Wallace? It would be a shame to waste a one-time top prospect as a bench hitter. It would be a crime to give him a starting job in the majors based on what he's shown so far. Sadly, it might serve no purpose to send him to the minors. Hopefully, Wallace can figure some things out and overcome his struggles against pitches at the extremes of the spectrum.
Post Script: For those of you wondering how Johnson's astronomical BABIP in 2010 affected his numbers, I churned 2010 and 2011 separately and here are the results: 8.8% Pos. Outcome per Pitch in 2010 and 7.0% in 2011. So I think 7.5% to 8.0% for future expectations could be right on the money. That's a possible 90-RBI Third Basemen, which every Astros fan should be thrilled with.