FanPost

AstroGraphs: The Statistical Evolution of Norris, Lyles, and Harrell from Draftees to Would-Be 2013 Astros


As of the initial publication date of this FanPost (February 7, 2013), Bud Norris, Jordan Lyles, and Lucas Harrell each remain Astros and thus would stand to be the proverbial frontrunners for beginning 2013 in Houston's starting rotation. Since they are also the candidates who are most familiar to Astros fans, let’s take a look at who they are now and how they got here from a statistical perspective.

Data Examined

These will be the 6 stats presented:

1. BB%: this is simply the number of walks allowed divided by the plate appearances against

2. K%: same as BB%, except with strikeouts as the numerator

3. BABIP: this is batting average against on balls in play (BIP), tweaked slightly in that home runs are considered balls in play while bunts are not (sacrifice flies are also treated as at-bats in the computation).

4. SLGBIP: this is slugging percentage against on balls in play, which amounts to total bases divided by balls in play, again inclusive of home runs and exclusive of bunts

5. Line Drive or Outfield Flyball per BIP: this is the percentage of balls in play that are classified by the official online game recaps of mlb.com and milb.com as line drives or outfield flyballs, again with home runs but not bunts counted as balls in play. Note that some of those line drives and flyballs are hits, some are outs, and some are neither (e.g., reached on error).

6. STBNL: a stat to be named later.

The player data presented was compiled from the MLB Advanced Media online repository or from the Retrosheet archives that reference the same. 2006 and 2005 minor league data was compiled from the minorleaguesplits archival Retrosheet-themed league data files. Major league averages for the stats were computed using Baseball Reference, unless specified otherwise.

Non-batted Ball Data

BB%

A few explanations are in order before we start to examine the data … Note that as you trace from the left start to right end of the pitcher’s plotted line below, you are moving chronologically; appreciate that after Lyles’ MLB debut in 2011, he then zigs back to AAA to open 2012 before returning for his second crack at MLB. Note also in the case of Harrell that I am only including MLB seasons for which the pitcher faced 200 batters, so while 2012 was technically Harrell’s 3rd MLB season given his 2 prior MLB cups o' coffee for the purposes of these analyses 2012 was Harrell’s first in the bigs. In the cases where the pitcher appeared at more than one minor league level, I pooled all of the data for that season and then plotted the data over the x-axis at the appropriate spot to reflect the percentage of batters faced at the higher and lower level; thus each marker symbol represents a single year of minor league data or a single year of MLB data (minor and major league data accumulated in the same year were not pooled).
Bbharrelllylesnorrisjpg_zps2ebe010f_medium

Lyles leads the trio handily here, with his seasonal BB% hovering between 6% and 7% over most of his pro career. Harrell was quite wild in his first 2 full-season minor league campaigns but has been better since sitting out 2007 with shoulder surgery, generally keeping his seasonal BB% between 8% and 11% and finishing the 2012 MLB season at approximately 9.5%. Norris’ BB% was showing an ominous gradual uptick over his pro career heading into his 2011 season, but since he seems to have righted that ship by posting consecutive 9% seasons. As a frame of reference, consider that 2012 National League starting pitchers collectively had a 7.4% BB%.

K%

This is clearly the domain of Norris, whose seasonal K% dropped gradually from 28% at Tri-City to 22% in his debut MLB season. Since then the K% has consistently remained around 22%, which stands to be about 3 percentage points better than average for a National League starter over that interval.

Kharrelllylesnorrisjpg_zps92efd2f9_medium

Conversely, Lyles’ K% has dropped precipitously over his pro career, a 12 percentage point difference from his short-season debut to his MLB data which now rates 3 percentage points below the collective average for NL starting pitchers. Having published elsewhere on SBNation that a AA starter who went through the California League enjoys a statistically significant 3% boost in his K% versus a AA starter who graduated from another High A league certainly makes me wonder if Lyles’ bypassing of the California League explains in part why his K% deviates from the similar track that the 3-year-older-at the-time Norris seemed to be following when each effectively skipped that level (Norris jumped the statistically-inconsequential Carolina League). Harrell’s K% numbers were much lower versus the other two’s at the lower levels of the minors but his data had remained very consistent about 15% as he climbed the minor league ladder until 2011 when he had a 5% spike in his K% at AAA that seemed for the most part to carry over into his debut season as a major league starter. While on the one hand Harrell’s 17% 2012 K% was 2 percentage points subpar for an NL starter, on the other he showed a huge increase in K% during the second half posting a 20.6% K% from July forward.

Batted Ball Data

BABIP

Lyles’ minor league BABIP values have been amazingly consistent, staying within a tight range that spans from a low of .342 (twice) to .358 as his high. In his 2 MLB campaigns, Lyles again rates similarly with .345 and .337 values. Norris, conversely, has been as low as .293 and as high as .375 as a minor leaguer, though his 4 major league seasons vary much less from .336 to .360 with the two better seasons being the most recent ones.

Babipharrelllylesnorrisjpg_zpsfc7a6ccd_medium

Getting to the elephant in the room, notice that the line that traces Harrell’s data, though clearly volatile in the first 3 seasons, doesn’t overlap either Norris’ or Lyles’ line at any point. Among his 7 BABIP data points on the plot, Harrell owns 6 of the 7 lowest BABIP values of the 3 pitchers with only Norris’ short-season 2006 debut preventing Harrell from sweeping the top 6 spots. As a pitcher who surrenders relatively few homers, Harrell stands to rate even better versus his peers when BABIP is calculated inclusive of homers, as I did here. Assuming that I did the computation right from Baseball Reference’s data, NL pitchers (inclusive of relievers) would have had a .325 BABIP using the scheme that I employ, which is .015 above the .310 figure that Harrell posted in his first go as a full-time major league starter. The minor league data that preceded it suggests that Harrell should better his new league’s seasonal value for pitchers many more times than not. While Lyles’ data paints him as consistent at BABIP, it also seems to portray him as a pitcher who has probably been subpar at the stat across most of his professional career.

SLGBIP

Harrell struggled so much with his AAA batted ball outcomes in 2010 that Norris’ 2009 AAA SLGBIP value actually beat Harrell’s 2010 SLGBIP. Otherwise each of Harrell’s 7 seasonal SLGBIP values beats every other SLGBIP value on the plot including all of Lyles’ data. Stated differently, in the best SLGBIP "season" of Lyles’ seven (2012 AAA), he failed to better the worst seasonal SLGBIP posted by Harrell as a pro. Likewise, in his best SLGBIP season as a pro, Norris managed to beat only Harrell’s worst seasonal SLGBIP as a pro. Computing the value as I did here, 2012 NL pitchers posted a .516 SLGBIP which is a quite hefty .071 above Harrell’s miniscule .445 2012 value. Note that both of Lyles’ MLB SLGBIP values and all 4 of Norris’ fall above the 2012 NL pitchers’ .516 value.

Slgbipharrelllylesnorrisjpg_zpsb947739a_medium

Line Drive or Outfield Flyball per BIP

The logical first hunch would be that Harrell’s extreme success on the SLGBIP metric is owed to his groundballing tendencies, which I prefer to view as his avoidance of line drives and outfield flyballs since those 2 batted ball types account for about 93% of extra-base hits. I turned to Statcorner for a comparable league reference on this stat owing to inconsistencies in how the more well-known sites handle flyballs caught by infielders and bunts, and NL starting pitchers allowed a line drive or outfield flyball on 47.7% of nonbunted balls in play. Harrell’s 2012 37.8% rate is 10 percentage points under that and is in keeping with most of the minor league data that preceded it.

Ldoffbharrelllylesnorrisjpg_zps86999fc1_medium

Lyles' 2012 AAA and 2012 MLB data shows a marked improvement versus the same data from 2011. That amounts to a 10% drop at the major league level and a 17% drop at AAA, so Lyles clearly improved his ability to avoid more dangerous contact in 2012 with his 42% MLB figure beating the figure for NL starters by nearly 6 percentage points. Norris’ plot shows a gradual rise that would be expected over the minor league portion and eventually stabilize in the majors, but unfortunately the trend has continued and he’s now venturing into much worse than league average waters. In fact, in all 4 of his MLB seasons, Norris has been over the 2012 NL pitcher league mark of 47.7% and he was 5 percentage points above that reference in 2012. Circling back to the opening sentence of this subsection, the plot below of each pitcher’s seasonal data confirms the SLGBIP-related benefits of avoiding line drives and outfield flyballs.

Slgbipvsldoffbharrelllylesnorrisjpg_zpsa2cf5e75_medium

Line Drive or Flyball Hit to "Pull-Field" Third of Outfield per BIP

This is a stat that you really can’t look up anywhere, but it often does a bit better job of explaining pitchers’ SLGBIP values than simply looking at their flyball and line drive rates. The reason is that the overwhelming majority of home runs are hit to the pull-field third (about 62%) and about 8 percentage points more doubles are hit to the pull-field than the opposite-field third (triples tend to be symmetrically distributed with a bias towards the centerfield third). And going to this additional level of detail helps better separate out the groundball specialists who do and don’t often get pulled in the air and likewise the flyball artists who do and don’t. So let’s look now at how the trio has evolved in terms of this stat.
Pulledofldoroffbharrelllylesnorrisjpg_zps4f723021_medium

Beyond that Lucas Harrell is elite at avoiding flyballs and line drives, he’s still better in so much as he has always been hard to pull in the air and that tendency carried over strongly into his debut season as a major league starter. The good velocity and movement on his fastballs and that he seldom hangs his breaking breaking stuff and other off-speed offerings surely factors into this. I would speculate that the average minor/major league righthanded starter probably has roughly 13% of his batted balls go as liners or flies to the pull-field third of the outfield based on similar evaluations done on other pitchers, with the best rating at about 6% and worst at around 22%. For Harrell to be posting a 7.4% figure in his debut season as a major league starter puts him in fairly exclusive company, though just how exclusive is hard to pinpoint given that the stat can’t be accessed for most pitchers without a great deal of data compilation. Lyles’ frequency of pulled outfield flies or line drives in the majors did not improve much versus his 2011 figure, so while he cut down considerably on the number of line drives and outfield flies versus 2011 he still has some work to do in terms of limiting his rate of getting pulled in the air. While Norris’ frequency of a line drive or outfield fly rose a bit in 2012, he did manage to drop his percentage of outfield liners and flies hit to the pull-field third of the outfield by 3% versus his 2010 and 2011 values though he remains subpar at this skill by major league standards. This final plot illustrates how well this stat predicts SLGBIP.

Slgbipvspulledofldoroffbharrelllylesnorrisjpg2_zps4e8e0ab0_medium

Summary

Bud Norris

Norris is the type of pitcher who must rely on his premium non-batted ball skill, K%, and the plus slider which he uses to realize or set up those whiffs to experience success as a major league starter. His batted ball data tends to be subpar across the board (BABIP, SLGBIP, LD+OFFB, etc.) making his subpar BB% data more consequential than it would be otherwise. Particularly concerning is that his percentage of line drives plus outfield flyballs has risen in each of the last 2 seasons after rating just a bit worse than average in his first 2 big league campaigns. So how does what you just read affect your thoughts on whether or when to trade Norris?

Jordan Lyles

Lyles’ premium skill would seem to be control, as his BB% data have typically been better than his peers and there is some potential to improve further with experience. His seasonal K% values have dropped considerably as a pro and I would question his capacity to improve on his major league data to date barring some unanticipated velocity gain. As such, Lyles would stand to rate well below the average major league righthanded starter at the strikeout. Lyles’ batted ball data is a bit of a mixed bag at this point. His track record would lead us to expect that he’ll be a pitcher who consistently rates about .010 worse than league average at BABIP. His SLGBIP data is more variable over his career and is thus harder to get a read on. That Lyles showed a huge improvement in his capacity to avoid line drives and outfield flyballs at 2 levels of competition during 2012 gives us some hope that he’ll be able to get his SLGBIP values down closer to or better than league averages looking forward, especially if he can limit the prevalence of batters pulling him in the air.

Lucas Harrell

There were times last season when I wondered if Harrell’s 2012 stats were simply a mirage cast by his catching clubs that couldn’t hit or weren’t hitting. Having looked at his historical batted ball data in such detail eliminates that concern as it sure looks like what we saw then was just the major league rendition of the comprehensive body of minor league work that preceded it. If the aforementioned second half spike in K% (to over 20%, which is essentially a bit better than league average) can be nearly sustained or perhaps even improved upon, then Harrell could well be looking at a breakout season as far as the national scene goes in 2013. Even without that lofty of a K%, Harrell still projects to be an excellent starter who will be extremely underpaid for the next several seasons. That his batted ball data is so impressive really diminishes most concerns related to his BB% being 2 to 3 percentage points over league average given the limited danger he incurs when batters do manage to put the ball in play. And beyond the fact that Harrell projects to be elite by big league standards at inducing groundballs and therein avoiding line drives and outfield flies, he also stands to be among the best handful of major league starters at preventing batters from pulling the ball in the air which further limits the trauma that they can impart on him. Do any of you Astros fans really want to trade this guy?

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join The Crawfish Boxes

You must be a member of The Crawfish Boxes to participate.

We have our own Community Guidelines at The Crawfish Boxes. You should read them.

Join The Crawfish Boxes

You must be a member of The Crawfish Boxes to participate.

We have our own Community Guidelines at The Crawfish Boxes. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9351_tracker