I compiled this data to assist with another analysis and figured that a small segment of this site’s readership would have interest in the results.
To stratify minor league left-handed pitchers (LHPs) based on their 2012 performance against left-handed batters (LHBs).
Step 1. Compile and crosscheck the 2012 vs LHB stats for each minor league LHP who faced at least 40 LHB at or above the High Rookie level (Appalachian League or Pioneer League) during the 2012 minor league regular season plus playoffs (using a combination of minorleaguecentral stats and game logs, FanGraphs game logs, and MiLB play-by-play recaps and boxscores).
Step 2. Compute 4 stats for each LHP using only their vs LHB data (accumulated at the High Rookie level and up):
- K% = K / PA
- BB&HBP% = (BB+HBP) / PA
- BABB = H / (AB+SF-K)
- LD&OFFB% = (number of linedrives and outfield flyballs) / (number of linedrives, outfield flyballs, nonbunted infield flyballs, and nonbunted groundballs)
Note that BABB (batting average on batted balls) is BABIP without the HR being subtracted from the numerator and denominator. BABB is being used rather than batting average against (BAA) since K% and BAA are weakly to moderately correlated (a high K% tends to drive down BAA given the extra outs recorded via the K). Think of BABB as a measure of the pitcher’s skill and luck at avoiding hits on batted balls. Think of LD&OFFB% as a measure of the pitcher’s skill at avoiding the batted ball types that are most dangerous in terms of extra-base-hit potential: linedrives (LD) and outfield flyballs (OFFB), each of which have about a 20% chance of being an extra-base hit. These 4 variables have very limited to zero correlation with one another, so being good (or bad) in one metric doesn’t affect how the player rates in any other one.
Step 3. Compute each LHP’s league average for each of the 4 stats. As an example, Pitcher A’s league average K% was the total number of strikeouts that LHB had against LHP in that league divided by the total number of PA had by LHB versus LHP. If a pitcher appeared in more than one league, their league average was weighted in proportion to the number of LHB that the pitcher faced in each league (if Pitcher B faced 38 LHB in League 1 and 62 LHB in League 2, then his League Average K% = 0.38 x League 1 Average K% + 0.62 x League 2 Average K%).
Step 4. Compute each LHP’s +/- Value for each metric based on how much better (+) or worse (-) they were versus their league average.
Step 5. Rank each +/- Value of each LHP versus the entire sample of qualifying pitchers based on how many standard deviations (SD) each +/- Value was better/worse than the average +/- Value for the metric under analysis using a 0 to 100 scale where 0 is 5 SD worse than average, 10 is 4 SD worse than average, …, 40 is 1 SD worse than average, 50 is average, 60 is 1 SD better than average, 70 is 2 SD better than average, … 100 is 5 SD better than average. In contrast to the traditional 20-to-80 scouting scale upon which this scheme is loosely based, values below 20 and above 80 were allowed and whole number scores that don’t end in 0 or 5 were left as is rather than rounded to the nearest 0 or 5 value.
Step 6. Compute a composite score to rank each LHP within the study population based on their overall performance across all 4 metrics. This formula was used:
The constants within the parenthesized innards of the formula were determined based on the frequency of each event (K, BB or HBP, H, LD or OFFB) when a LHP faced a LHB over the entire minor league season. Thus, each LHP’s Composite Score was computed by giving 21.8% weight to their K Score, 10.4% weight to their BB+HBP Score, 28.7% weight to their BABB score, and 39.1% weight to their LD+OFFB Score. The remainder of the equation simply scales the results so that 10 points of Composite Score amounts to 1 SD. So as before, a LHP with a 60 Composite Score is 1 SD better than an average LHP is against LHB, with that average LHP rating at 50 in Composite Score. Note that the 28.7% BABB component is the only fielding- and field-dependent part of Composite Score.
Top 10 LHP vs LHB by K Score
22nd-round 2012 draftee Long led easily here with his K% going up rather than down in jumping from Short-season A to High A ball in mid-August. Yes, that’s that Oliver Perez in third place; he made 33 appearances out of the Mariners’ bullpen after late June. Combined these 10 pitchers threw almost exclusively in relief, with only 3 starts made between them.
Top 10 LHP vs LHB by BB&HBP Score
With the exception of Ortiz, none of these pitched above High A. The ten also reflect a more even mix between starters and relievers versus the K Score Top 10.
Top 10 LHP vs LHB by BABB Score
Long topped the K Score list as well as this one making him the LHP against whom LHB had the most difficulty putting a ball into play and the one against whom they had the most difficulty getting a hit when they did manage to put one into play. The polar extremes of Testa's BABB Score and BB&HBP Score are fascinating; he was the most generous minor league LHP at walking/hitting LHB while also being the 2nd stingiest in terms of allowing them a hit.
Top 10 LHP vs LHB by LD&OFFB Score
Septimo and Ramos cracked this Top 10 as well as the BABB Top 10.
Top 30 LHP vs LHB by Composite Score
Septimo ranked #1 in Composite Score of the 547 LHP evaluated and logged 14+ innings over a few stints with the White Sox from June onward. Septimo was claimed off waivers by the Sox in 2011 from the Diamondbacks, who had converted him from an outfielder in 2008. Septimo has typically posted atrocious control stats as a pitcher (18% career BB%) but managed to be roughly average at that skill in 2012; average MLB control could be enough for him to stick in the bigs as a late-inning reliever or lefty-batter specialist given his strikeout and favorable batted ball type tendencies. Video and Pitch F/X data has the fidgety southpaw throwing a 4-seam fastball that averages 92 to 93 mph and a good slider from a three-quarterish armslot. Nieves, a 2007 7th-round pick out of the Puerto Rico Baseball Academy (high school), was released by the Brewers in March and signed with the Tigers in May. Though he had spent 2011 pitching well against LHB and terribly against RHB in High A, the Tigers wound up keeping him at their short-season A-ball team for all of 2012. MiLBTV footage (8/12 @ Aberdeen, 9/4 @ Staten Island) shows Nieves throwing lefty bats a fastball and sweeping slider from barely north of sidearm and the first base end of the rubber. YouTube video from 2009 shows him coming from a nearly three-quarter armslot, so there’s been some delivery changes made within the 3-year interval with the intent of thwarting LHB. Now 23, Nieves was re-signed by the Tigers after 2012 only to be gobbled up by the Blue Jays with the final, and their third, selection of the minor league phase of the Rule V draft; he probably seems best cast to begin 2013 in a lefty specialist role perhaps at High A. Keep an eye out for Long showing up in Houston during 2013. Three consecutive Long appearances are archived on MiLBTV (8/7 @ Aberdeen, 8/11 & 8/12 @ San Jose); besides his calling card of a sweeping sidearm slider he has some funk factor in his delivery plus follow-through and will occasionally throw pitches over the top against RHB to disrupt their focus. Septimo, Hottovy, Ramos, Cedeno, Paterson, Romero, Outman, Layne, Perez, Downs, and Werner all pitched in the majors in 2012 with the remainder save for Ring yet to toe the rubber as an MLBer. Chime in on any of the rest if you’ve seen them in 2012 or earlier.
Bottom 30 LHP vs LHB by Composite Score
Maloney, who saw MLB time with the Twins in 2012 and Reds before then, took bottom honors ranking in 547th place. While on the one hand MLB.com post-2012 #91 prospect Daniel Norris was just 19 in his 2012 season and barely faced enough LHB to qualify for this study, this clearly isn’t the sort of list that Blue Jays’ fans want to see his name on after 2013 and beyond (he didn’t crack John’s post-2012 Top 50 Pitching Prospects list). The Jason Lane is the former MLB outfielder; the Diamondbacks released him after he posted the AAA numbers shown and he ultimately wound up being a good starting pitcher in an independent league rotation that later featured his 2005 World Series teammate, Roger Clemens.
All Study Pitchers vs LHB
Data for each of the 547 LHP evaluated can be found in this Google spreadsheet. Download away and look up your favorite minor league lefties to see how they fared per this system.
This is only one season’s worth of data per pitcher; more seasons of data would allow for stronger inferences to be made. No accounting for park effects was made. Starters, relievers, and in-betweeners were included – consider that relievers generally enjoy a few percentage points advantage in K% versus starters. Note that vs RHB stats were not analyzed, and those would be particularly important to look at in the case of a potential MLB starting pitcher. Some well-performing pitchers were limited as to how much better than league average they could be since the league average to which their performance was compared was lower relative to other leagues’ averages on the same metric. In the real world a 50 score in AAA beats a 50 score in High Rookie ball, but here they are equivalent.
What is probably being measured here
In many cases, this analysis figures to indirectly be assessing the quality of the pitcher’s breaking ball(s) versus same-sided bats during 2012. A lefty reliever who doesn’t distinguish himself in a positive way versus other minor league lefties in terms of his performance against lefty bats at the middle to upper levels of the minors probably doesn’t have much of a future as a prospective major leaguer. Similarly, a lefty starter who doesn’t fare well versus lefty bats won’t figure to have relief work as a fall-back plan if the starting route doesn’t pan out. That said, the goal here was more to identify statistical over- and under-achievers more so than project how the pitchers may fare as major leaguers; the former is much easier than the latter.