Last week in this space, we attempted to quantify the impact of strikeout and walk rates on hitters’ offensive production. This time around, let’s do the same with starting pitcher performance, and then a dig a little deeper into pitchers’ batted-ball profiles in an attempt to assess the true-talent level suggested by the underlying data.
Our control group will be 12 starting pitchers who have changed clubs this offseason and logged a substantial innings load in 2013. First, let’s separate these pitchers’ Ks and BBs from the batted balls they allowed to isolate the relative contributions of each to overall pitcher performance.
|K %||BB %||BIP AVG||BIP SLG||BIP RUN||BIP R:100||TOT AVG||TOT OBP||TOT SLG||TOT RUN||TOT R:100||ACT ERA||ERA:100|
The table above lists each pitcher’s K and BB percentage, the AVG and SLG they allowed on all batted balls, and the total AVG/OBP/SLG they allowed to all batters, including the K and BB information. (HBP are not included in overall OBP, and SH and SF are included as outs for purposes of this exercise.) The BIP and total run values are calculated and scaled to MLB-average ERA as follows: the square of ((1.7 * Pitcher OBP + Pitcher SLG)/ (1.7 * MLB OBP + MLB SLG)) * MLB Avg. ERA. The estimated run values excluding and including the K and BB data are scaled to 100 in the sixth and 11th columns above. For comparative purposes, pitchers’ actual ERA and relative actual ERA scaled to MLB average are listed in the two rightmost columns above. None of the above are adjusted for park factors.
The impact of K and BB rate extremes is readily apparent. High K/Low BB guy Dan Haren‘s calculated relative ERA plummets from 119 to 102 once K/BB data is included, while low K/average BB guy Jordan Lyles‘ calculated relative ERA rises from 108 to 117. Comparing the actual ERA vs. the calculated ERA identifies some pitchers who benefited from or were the victims of good or bad sequencing, among other factors, in 2013. Bartolo Colon‘s 3.26 calculated ERA was much higher than his 2.65 actual ERA, while Lyles (4.52 vs. 5.59) and Edinson Volquez (4.70 vs. 5.71) both had actual ERAs over one full run higher than their combined K/BB and batted-ball profiles would suggest. Again, team defense, ballpark, luck, etc., has not yet been taken into consideration. Now let’s dig a little deeper and look at the batted-ball breakdowns for this group of pitchers.
|POP %||REL POP #||PCT POP||FLY %||REL FLY #||PCT FLY||LD %||REL LD #||PCT LD||GB %||REL GB #||PCT GB|
Each pitcher’s balls in play allowed are broken down by type above, and are expressed relative to MLB average (scaled to 100), and as a percentile rank among 2013 regular major league starting pitchers. Interestingly, this particular group of starters includes some batted ball extremes — Phil Hughes and Dan Haren are in the 99th percentile with regard to flyball contact, and Tim Hudson is in the 1st percentile, with fellow groundball guy Roberto Hernandez‘ profile almost as extreme as Hudson’s. Individual popup, flyball and groundball percentages tend to correlate quite well from year to year, while line-drive rates are more random.
Now, let’s combine the frequency data above with the production allowed by these pitchers within the three major batted-ball categories to get a better feel for what makes each of these pitchers tick. (Virtually all popups are outs — no need to devote more space to the fourth major batted-ball type at this time.)
|FLY AVG||FLY SLG||R FLY PRD||ADJ FLY||LD AVG||LD SLG||R LD PRD||ADJ LD||GB AVG||GB SLG||R GB PRD||ADJ GB||ALL AVG||ALL SLG||R ALL PRD||ADJ ALL||TRU ERA|
The third, seventh, 11th and 15th columns above represent the relative run value for batted-ball type compared to the MLB average, scaled to 100. The fourth, eighth, 12th and 16th columns include estimated adjustments for team defense, ballpark, luck, etc..
Some really interesting stuff here — let’s look at a highlight or two for each pitcher. Bartolo Colon relies on popups and relatively weak flyball contact, but was helped a bit in 2014 by his home park and outfield defense. Scott Feldman induced lots of grounders, and held opponents to a ridiculously low level of production on them, though some regression should be expected in that area in 2014. Doug Fister also allowed tons of grounders, but yielded higher than MLB average production on those grounders, due at least in part to the Tigers’ limited infield defense. Dan Haren allowed a ton of flyballs and relatively authoritative contact across the board, but was kept afloat by his strong K and BB rates. Roberto Hernandez allowed a ton of grounders, with a relatively low production level, but the fly balls he did allow were hit with significant authority.
Tim Hudson is quite interesting — he obviously is an extreme groundball guy, but the league got average production on those groundballs in 2013. His popup rate, however, has become quite high for a groundballer, offering him another efficient way to retire hitters. Phil Hughes possesses a fairly unique portfolio, in a not-so-positive way. He allowed tons of flyballs, but managed the production on them quite well last season. However, he allowed well more than league-average production on the relatively few groundballs he allowed. That combination is not a likely pathway to long-term success.
Scott Kazmir‘s comeback season had a couple of blemishes — he yielded a very high line-drive rate, and well more than league-average damage on the flyballs he allowed. The LDs should regress; not so sure about the flyball damage. On the plus side, Jordan Lyles has a strong groundball tendency, and once adjusted for team defense, allowed below-league-average production on them. On the negative, his K rate was dangerously low, and he allowed very authoritative flyball contact, a no-no in Colorado, his new home. Ricky Nolasco is what he is — a pitcher who lacks material flaws. He controlled flyball contact quite well in 2013, and his solid K and BB rates offer a baseline for success. Jason Vargas has become quite a success story, relatively speaking. A fringe big leaguer not that long ago, he dialed down his significant tendency to allow authoritative flyball contact a bit in 2013, and now moves to an even more flyball-forgiving environment in Kansas City in 2014.
And let’s finish with Edinson Volquez, potentially the best bargain free-agent signing thus far this offseason. A) His extremely high 2013 line-drive rate should regress going forward, and B) the outlandishly high level of production he allowed on flyballs in 2013 was simply not real — and he’s headed to a forgiving home park with a stellar outfield defense in 2014. His ERA should drop by two full points in 2014, even without major improvements in control.
Breaking down the whole into its components can yield some very interesting insights into starting pitcher performance. Plus, this is just the entry point for analysis. Incorporating traditional scouting concepts to help determine why the above pitchers are missing more or less bats, or yielding different batted-ball mixes or levels of production within batted-ball groups can take such analyses to the next level. Next time, we’ll take some of the concepts discussed today and apply them to a few hurlers who remain on the free-agent market, at least for a couple more days, until the Tanaka Embargo ends.