PITTSBURGH PIRATES

PREVIEWING ROLES IN THE STARTING ROTATION

Pitching roles are thrown around, with the question seemingly revolving around if a pitcher is an ace or not.  What are the Pirates pitchers roles?

Last season, Pittsburgh Pirates starter Jameson Taillon pitched to a 3.20 ERA, 3.46 FIP, and 3.7 fWAR in 191 innings.  His teammate Trevor Williams tossed a 3.11 ERA, 3.86 FIP, and 2.5 fWAR in 170.7 innings.  At the deadline, the Pirates acquired Chris Archer in a deal that makes you stop and turn your head just out of the name they both acquired and traded away.  This can lead to the question of what role is each pitcher?  Taillon and Williams seemingly can be pointed to as a low one/high two based on their ERA’s and Archer has that name value, but what role do they actually serve?

To answer the question, I downloaded seasonal data from Fangraphs from 2016-18 with the criteria narrowing down to innings that were thrown just as a starter with a minimum of 100 innings.  This leaves 390 player seasons over three years, which on average is 130 player seasons per season and 4.3 starters per team.  Their different run prevention metrics are as followed:

Run Prevention By Year
Year	Seasons	RA9	ERA	FIP
2016	137	4.38	4.06	4.10
2017	125	4.52	4.18	4.25
2018	128	4.24	3.94	4.01
Total	390	4.38	4.06	4.11
*Numbers representing time as a starter, minimum 100 innings

Then, breaking down roles in percentiles, we can get the average RA9 for each role in the rotation.  The 85th percentile is the number one starters (here we’re classifying the number one starter differently than an ace, who is the 95th percentile), the two starters are the 65th, three starters are the 45th (ie around the average major league starter), four starters are the 25th, and the rest are your five starters (percentiles/classifications can be moved as these definitions are arbitrary).  The count of each pitcher in a season over the last three years are:

Count of Pitcher by Percentile (Role)
Season	85-100°	65-85°	45-65°	25-45°	0-25°
2016	21	27	27	28	34
2017	19	25	25	25	31
2018	20	25	25	26	32
Total	60	77	77	79	96

Average RA9 by Role
Season	1 Starter	2 Starter	3 Starter	4 Starter	5 Starter
2016	3.04	3.71	4.30	4.93	5.73
2017	3.08	3.92	4.44	4.91	5.84
2018	2.83	3.68	4.18	4.69	5.57
Total	2.98	3.77	4.31	4.85	5.71

The number one starters in baseball have on average around a 3.00 RA9, where the biggest surprises in terms of name recognition were 2016 Junior Guerra (3.71 FIP), 2017 Chase Anderson (3.58 FIP), and 2018 Dereck Rodriguez (3.81 FIP).  On the other end of the spectrum, in just 2018, we see pitchers Felix Hernandez, Alex Cobb, Marcus Stroman, and former top prospects Jon Gray and Lucas Giolito.

Of the 33 players who changed teams over the course of a season, only Rich Hill (2.37 RA9 with the Athletics and Dodgers) and Justin Verlander (3.50 RA9 with the Tigers and Astros, though was the 85.5th percentile, so a low one/high number two might be more accurate).  With respect to the Pirates transactions, 2016 Ivan Nova (4.44 RA9) was a three, 2018 Chris Archer (4.67 RA9) was a four, and both 2016 seasons of Jon Niese (5.74 RA9) and Francisco Liriano (5.31 RA9) were fives.

In terms of players who pitched entire seasons with a team, the Dodgers had the most one and two starter seasons (10 of 14 total), and they acquired 2016 Hill and 2017 Yu Darvish (two starter, 4.00 RA9).  The most amount of four and five starters are the Tigers and Orioles, both having nine seasons out of 12.

The small amount of pitchers who can be labeled as having an ace level season (an ace is always a one starter, but a one starter isn’t always an ace), the year-by-year RA9s are:

Aces 2016-18
Season	RA9	Count
2016	2.62	7
2017	2.81	7
2018	2.35	7
Total	2.59	21

The only pitchers who have had multiple ace level seasons are 2016 and 2017 Clayton Kershaw and 2017 and 2018 Chris Sale.  For the 2018 season, the seven pitchers who were in the 95th percentile are Jacob deGrom, Blake Snell, Chris Sale, Aaron Nola, Walker Buehler, Justin Verlander, and Trevor Bauer.  Both Guerra and Anderson’s seasons qualified as “ace” level, as 2016 Jose Fernandez, 2017 Dallas Keuchel and Gio Gonzalez, and 2018 Kyle Freeland just missed being in the 94th percentile.

Just looking at the Pirates from 2016-18, the club has had one starter, a trio of two starters, and a five starter:

Pirates Rotation Roles
Season	Name	Team	Starter
2018	Taillon, Jameson	Pirates	1
2016	Taillon, Jameson	Pirates	2
2017	Williams, Trevor	Pirates	2
2018	Williams, Trevor	Pirates	2
2016	Cole, Gerrit	Pirates	3
2017	Cole, Gerrit	Pirates	3
2017	Nova, Ivan	Pirates	3
2018	Musgrove, Joe	Pirates	3
2017	Kuhl, Chad	Pirates	4
2017	Taillon, Jameson	Pirates	4
2018	Nova, Ivan	Pirates	4
2016	Locke, Jeff	Pirates	5

The biggest surprise here is that in 2017 Williams qualified as a two starter based on this method.  For the season, his RA9 was 4.37, which would slide him in as roughly a three starter (average RA9 for that bucket was 4.44 in 2017 and 4.31 over the three-year period).  But this only takes in his numbers as a starter, where he posted a 4.22 RA9 over 138.7 innings.  This season was in the 67th percentile, so right at the edge of the number two bucket, so calling that season a three starter level season certainly is more than fair, and it’s what the Rule of Six (6.0 – pitcher WAR = slot in rotation) generally suggests of that season, seeing him as a three/four starter more than a two starter.

Using the STEAMER projections for 2019 (projections lean conservative), the Pirates pitchers project at the following RA9s:

2019 STEAMER Projections
Pitcher	Innings	R	RA9
Archer, Chris	191	85	4.01
Taillon, Jameson	190	90	4.26
Musgrove, Joe	162	79	4.39
Lyles, Jordan*	53	28	4.75
Kingham, Nick	113	61	4.86
Williams, Trevor	145	82	5.09
*Doesn’t fit 100 innings requirement

STEAMER projects a big step back for Williams, mainly because of what was outlined above, but I think he’ll be better than a 5.09 RA9, more around the 4.80 RA9 the four starters have produced from 2016-18.  Similarly, Taillon seems likely to better than a 4.26 RA9, especially with the addition of his slider/cutter that led to more success in 2018.

Archer continues to project well, but he has underperformed his peripherals (27.5 percent strikeout rate, 7.5 percent walk rate, 44.8 percent ground ball rate, and 12.9 percent swinging strike rate) over the last three years while posting greater than 4.00 ERAs.  The underlying numbers are there for Archer, he just has to go out and pitch better and be the pitcher those numbers indicate he has the potential to be.

Kingham and Lyles (who doesn’t have the requirement and projects as a reliever) are in a battle for the last spot in the rotation, with other likely serving as the long man in the bullpen.  Lyles projection would get worse if projected as a starter, and Kingham currently projects as a four.  Maybe the curveball is real and Lyles can be a projected four in the rotation, but based on the projections Kingham seems like a better fit.

For the most part, for the Pirates to exceed their projected 78 wins by Fangraphs and 80 wins by PECOTA, the pitching will have to over perform their projections, which projects as a two (Archer), pair of threes (Taillon and Musgrove), a four (Kingham), and a five (Williams).

Taillon and Williams seem like safer bets to perform a role above their projection, but that still might not be enough to will the Pirates into the playoffs.  Pitching is the club’s best attribute, but they have lacked the multiple top of the rotation arms in a given season they will need to compete in the National League Central.