MASTER NOTES: The May-April season (pitchers)

In last week’s Master Notes, I looked at the baseball “season” that started last May 1st and ended this year on April 30th, making some observations about hitters during that six-month span of major-league baseball.

This week, I’m repeating the procedure, but looking at the pitchers. And I’m splitting them up into starters (minimum 10 starts) and relievers. We'll start with the starters, which makes sense.

Finding #1: If You Think Starters Aren’t Getting Deep Into Games, You’re Right

Among the 116 pitchers who started every game they pitched (no relief appearances), only three managed to average 7 innings or more per start: Clayton Kershaw, Chris Sale and Jake Arrieta. Corey Kluber, Dallas Keuchel, Zack Greinke and Marcus Stroman were at 6.9 and the numbers tailed down from there.

At the bottom of the range, starters like Charlie Morton, Mike Pelfrey, Taylor Jungmann, Jeremy Hellickson, Kyle Kendrick, Mike Bolsinger, Wily Peralta and Joe Kelly were under 5.5 IP/GS. That should make owners wonder about their roster value, since their ratios are high, their strikeouts relatively low and their wins potential restricted.

Unlock all of our insights on your way to a fantasy title! See what our season-long coverage offers and then subscribe to all of our insights on your way to a fantasy title! See what our season-long coverage offers and then subscribe to

Finding #2: Getting Deep Into Games Doesn’t Mean Getting Decisions

You’d think Kershaw and Sale, league leaders in getting deep into games, would therefore get the most decisions.


Kershaw, the depth leader, is 66th on the list of decisions per starts (D%): He got a win or a loss in 70% of his starts, keeping company with Anthony DeSclafani and Nathan Eovaldi, both under 6.0 IP/GS. The D% leaders were Josh Tomlin (92%) and Dallas Keuchel (91%), with 18 more starters at or over 80%, including Sale (19th, 81%) and Arrieta (11th, 85%). Fourteen pitchers were under 60%, with CC Sabathia (52%) and Kyle Kendricks (53%) bringing up the rear.

The correlation between IP/GS and D% was 0.39—positive, but weakly so. It was pretty common for a pitcher to be well up the IP/GS list and well down the D% list, and vice-versa.

The situation is even more pronounced with Wins/GS. Just four pitchers—Stroman, Tomlin, Arrieta and Steven Matz—won 70% or more of their starts. Kershaw and Zack Greinke, despite sterling skills and performance measures, won just 52% of their starts. Sale and Jose Fernandez were at 50%. And 11 pitchers with excellent peripherals—ERAs below 3.00 and/or WHIPs below 1.0—had W/GS% below 50%.

My takeaway from all this is that wins is a dumb stat.

Finding #2: Wins May Not Be A Dumb Stat

The common beef about the wins category is that unskilled lucky pitchers get too many of them and skilled pitchers too few. In fact, I basically just finished making that case. But in this period, at least, it ain’t necessarily so. Over the period, the wins leader was Arrieta, with 24 wins. Other top winners include most of the Cy Young ballot—Kershaw, Price, Bumgarner, Sale, Scherzer, Greinke. Their wins percentages might not be great, but they had a lot of starts. The list of top winners also includes Colby Lewis, Collin McHugh and Rubby de la Rosa, so there is clearly an element of luck still at play.

Of the top 20 starters in the Wins category, those with 15 or more “W”s, the aggregate ERA was just over 3, and the aggregate WHIP was 1.10. Ten of the 20 had ERAs under 3, and only three (Lewis, McHugh and Rubby) had ERAs over 4. Similarly, six of the wins leaders had WHIPS under 1, and just five (Lewis, McHugh and Rubby again, plus Michael Wacha and Carlos Martinez) were over 1.25 WHIP.

Appropriately, the bottom of the wins list, those with four wins or fewer, was populated by pretty bad pitchers. Their aggregate ERA was 4.97 and WHIP was 1.46. Fourteen of the 29 four-win starters had ERAs over 5 and WHIPs over 1.5. Shane Greene led the ERA parade at 7.83 while Keyvious Sampson had the booby prize for WHIP at a sparkling 1.83. The diamond in the coal pile was Jerad Eickhoff, who managed a 3.21/1.07 WHIP, despite only four wins in 13 starts. Wins are dumb!

Finding #3: Complete Games Are Vanishing

Arrieta led the way with five CG. five other pitchers—Kershaw, Madison Bumgarner, Max Scherzer, Kluber and the departed Mark Buehrle—had four apiece. Fifty-seven other pitchers threw at least one CG.

This is no surprise. The CG is going the way of the way of the dodo, the passenger pigeon, and pop stars who can sing in tune. In 1980, 20% of all games were CGs, and 160 different pitchers had at least one. The A’s alone had the top three—Rick Langford had 28 (!), Mike Norris had 24 and Matt Keough had 20.

By 1990, the CG number had fallen to 10%, though 133 starters still managed at least one. In 2000, though, CGs had fallen to 5% of starts, with just over 100 different pitchers having at least one and no pitcher in double digits. David Wells led with nine.

Ten years later, the transformation was complete. Barely 3% of starts were completed, just 84 pitchers managed a CG, and Roy Halladay led with nine. Last year, the number percentage of CGs was just over 2%. And this year, so far, just over 1%.

Finding #4: Some ERA Surprises

ERAs keep dropping faster than oil prices, although ERAs have less effect on Middle Eastern economies. In 1996, the aggregate ERA of starters, using the same criteria as here, was 4.46, with just seven starters under 3.00 ERA and 39 from 3.01-4.00. In this current period, aggregate ERA was 3.93—half a run less—with 22 pitchers under 3.00, and 96 between 3.01-4.00. In both instances, the number of pitchers has tripled in 20 years

The ERA leaders for the period start with Arrieta’s 1.62, with Kershaw in third at 1.96. In second? Rich Hill, whose astonishing April lowered his period ERA to a few thousandths less than Kershaw’s. Other surprises include Jaime Garcia’s 2.68; Jose Quintana’s 2.76; Tyler Duffey’s 2.90, just ahead of Bumgarner and Sale; and Joe Ross’ 2.99.

The big surprise at the other end of the list is Michael Pineda’s 4.84, nearly two runs worse than his xERA and a run-and-a-half over his xFIP and SIERA, two other skills-based ERA predictors.

Finding #5: Some WHIP Surprises

Like ERA, WHIP is tracking downward, with 26 starters under 1.10, compared with just five in 2006. And 10 starters in the period showcased WHIPs under 1.00, while just one such starter, Johan Santana, managed to get under the 1.00 threshold 10 years ago.

Again, most of the names are whom we’d expect. Surprises in the Top 20 included third-place Josh Tomlin, at 0.86, narrowly lagging only Kershaw and Arrieta; Hill’s 0.95; Drew Smyly’s 1.03; Garcia, Eickhoff and Ross.

No surprises among the Bottom-20, but up just a little higher we see Carlos Rodon, Doug Fister and Martin Perez.

Finding #6: Few Surprises Among K Leaders

Kershaw, Sale, Scherzer, Archer, Bumgarner, Kluber, Arrieta, Price…

There might be a few surprises in who is not in the top tiers on the list. Matt Harvey, for instance, was under 200, and Jeff Samardzija was further down than Harvey. The top Dom guys pretty much followed the K list, with exceptions due to PT: Rich Hill was #1 at 12.0 K/9, while Jose Fernandez was second at 11.5. But Hill had just nine starts and Fernandez 16.

Toward the bottom of the Dom list, one name jumps out: Toronto rotation leader Marcus Stroman’s Dom is a paltry 5.4 K/9, one of just 28 qualifying starters with Doms under 6.0 K/9. Stroman’s rate has been brought down by a 4.9 in April of this season after an already short low-Dom run last year coming back from injury. While there is that explanation, however, owners with an interest in Stroman should monitor his Dom before making any decisions, because 5.4 ain’t going to cut it in the long run.

Finding #8: Your Saves Leader: Mark Melancon

Melancon has 53 saves in the “season” just past, with a 1.88 ERA and a 0.87 WHIP. Four relievers have 40+ saves: Kenley Jansen, Trevor Rosenthal, Shawn Tolleson, Jeurys Familia and Francisco Rodriguez. Interestingly, most of these guys were widely discussed as closers highly likely to lose their jobs.

Eleven relievers have 30-39 saves and six more have 20-29.

Finding #9: You Know How All Pitchers Regress to a 70% Strand Rate? Nope.

The median S% for the 129 qualifying relievers was 78%, with 46 different relievers above 80%, led by Wade Davis at 93% and Chasen Shreve (!) at 90%. Only 19 were under 70%, including Fernando Salas, whose name gets brought up every time the Angels bullpen is discussed.

Finding #10: Like Wade Davis Needed Even More Value

Ever since Ron Shandler formulated the LIMA Plan, we've known that late-game relievers can add a nice bundle of vulture wins to stats along with superb decimals and, sometimes, a useful number of Ks. During our season, Kevin Siegrist of the Cardinals had nine wins (and five saves) in 73 IP with a 2.32 ERA, 1.15 WHIP and 90 whiffs. The now-injured Keone Kela of TEX had eight wins, a couple of saves, and a 2.86/1.11 with 72 Ks. Orioles Darren O’Day and Brad Brach had seven wins each, with WHIPs right around 1.00 and ERAs of 1.76 (O’Day) and 2.19 (Brach).

Wade Davis had 20 Saves in the season and seven wins, because he was setting up for Greg Holland during the early part of our season. His decimals are both under 1.00 and he had 78 Ks.

And by the way: I multiplied ERA and WHIP to make a metric called ERAWP (pronounced “EE-rop”) and took the top 50 relievers, leaving out all the guys with 10+ saves. Aggregate ERA was 2.67, aggregate WHIP 1.11, with medians right in those neighborhoods. Six non-saves relievers had ERAWPs under 2.00: Shawn Kelley, Alex Wilson, O'Day, Dellin Betances, Joaquin Benoit and Will Harris. Four of these six had ERAs under 2.00 and all had WHIPS 1.01 or better. And of course Bettances’ 136 Ks were more than 107 of the qualifying starters.

As with the batters, looking at pitchers through the lens of an alternate season confirmed some ideas and revealed others. I like the idea of the ERAWP metric as a quick first-throw shortcut to rank relievers, and maybe starters as well.

Finally, remember again the warning I gave last week in the hitters discussion: All of these stats and metrics stabilize over different periods. A “season,” no matter how structured, seems like a long time but for many stats, it is not a large enough sample, while for others, it is more than enough.

Click here to subscribe

  For more information about the terms used in this article, see our Glossary Primer.