RADIO
Very few fantasy baseballers have Clayton Kershaw ranked as the top player on their board. But should they?
This article will examine just that topic using z-score rankings based on Steamer projections.
Editor's note: for even more draft prep, visit our awesome 2017 fantasy baseball rankings dashboard. It has lots of in-depth staff rankings and draft strategy columns. You will find tiered rankings for every position, 2017 impact rookie rankings, AL/NL only league ranks and lots more. Bookmark the page, and win your drafts.
How we Performed the Analysis
First, a quick primer on how we calculated these rankings and values: we used the standard deviation and mean (components of z-score) based on a sample of the top 120 starting pitchers, as ranked by 2017 projected WAR. We then calculated z-scores for the four standard 5x5 starting pitcher categories (W, K, WHIP, ERA) for the top 190 pitchers, and averaged the four categorical z-scores together.
The reason we used a 120-starting pitcher sample for the z-score calculation is that in most fantasy leagues, at most 120 pitchers will be relevant. Using larger samples would have skewed the results inappropriately. The reason we ranked the top 190 pitchers is, well, we like to go deep at RotoBaller!
Additionally, because starting pitchers only contribute in four categories, compared to hitters’ five, we reduced the pitchers’ averaged z-scores by 20% to make them comparable to outfielders. The reason this is necessary is that we want to look at total value returned by a player. If Kershaw only contributed strikeouts, and had a z-score of 5, it wouldn't be fair to use that 5 as a point of comparison to Trout's average z-score of 2.174 across five categories. Thus, the 20% deduction makes sense. Below is a chart of the top ten hitters and top ten pitchers with these adjustments:
| TOP TEN PITCHERS | ADJUSTED Z-SCORE | TOP TEN HITTERS | Z-SCORE |
| Clayton Kershaw | 2.768 | Mike Trout | 2.174 |
| Max Scherzer | 1.952 | Mookie Betts | 1.864 |
| Noah Syndergaard | 1.690 | Miguel Cabrera | 1.615 |
| Madison Bumgarner | 1.488 | Manny Machado | 1.591 |
| Chris Sale | 1.416 | Jose Altuve | 1.586 |
| Corey Kluber | 1.400 | Nolan Arenado | 1.531 |
| Stephen Strasburg | 1.248 | Trea Turner | 1.498 |
| Carlos Carrasco | 1.176 | Bryce Harper | 1.498 |
| Jon Lester | 1.112 | Paul Goldschmidt | 1.394 |
| David Price | 1.112 | Kris Bryant | 1.335 |
The Argument for Kershaw as the Top Pick
Wow. Kershaw’s adjusted score is 27.3% higher than Mike Trout’s and 41.8% higher than the next highest pitcher! How could he not be number one? Moreover, a starting pitcher with over 200 innings takes up a larger percentage of your innings than a hitter does of your total games, so you could argue Kershaw’s exceptional rate stats are undervalued relative to hitters. However, these pitchers’ wins and strikeouts are overvalued by the z-score, as they are not rate stats and get a boost from the number of innings.
While many feel starting pitchers are not as valuable due to the ability to stream pitchers, this line of thinking is flawed with respect to elite pitchers who you will start every week. Rosters are not large enough, nor waiver wires deep enough, to stream all your pitcher slots. In fact, the ability to stream the back end of your rotation makes backend starters less valuable, increasing the relative value of top-end starters. Thus, I rank the middling starting pitchers lower than my peers.
But, there are surely arguments against Kershaw as the number one starter as well…
The Argument Against Kershaw as the Top Pick
Ranking Pitchers by These Scores Requires Moving the Other Pitchers Far Up in the Rankings
By these rankings, the other pitchers’ overall rankings are: Scherzer (3), Syndergaard (5), Bumgarner (12), Sale (13), Strasburg (14), etc. Thus, using this method requires the conclusion that we are undervaluing elite starting pitching generally, rather than undervaluing Kershaw per se.
Is this the case? Maybe not. Starting pitchers are higher injury risks, and their value needs to be decreased to account for that risk. Moreover, one of their four categories (wins) is dependent on a significant amount of luck; while we can argue how much luck is involved in a pitcher’s win loss record, most would agree that it is more than is involved in the hitters’ categories.
For these reasons, we can adjust the scores so that the other pitchers’ rankings are closer to their ADP. Adjusting the scores so that Scherzer falls to his ADP of 9th has Kershaw (2), Syndergaard (14), Bumgarner (17), Sale (20), Kluber (21), Strasburg (24). These rankings still seem to rank starting pitchers higher than our perception. If we adjust Scherzer to half way between Turner/Harper (tied for 8th) and the next hitter (Goldschmidt), we get Kershaw (2), Syndergaard (14), Bumgarner (19), Sale (20), Kluber (21), Strasburg (24)…Lester (39). These rankings are still a little higher than ADP would suggest, but they are fairly close and still have Kershaw second.
Steamer Is Overprojecting Kershaw Relative to Other Starters
Steamer projects Kershaw for 17 wins, a 2.29 era, 0.93 WHIP and 258 strikeouts. Those seem reasonable. However, Steamer’s second lowest ERA projection is 3.05. The second lowest WHIP is 1.02. The second highest win total is 15. That explains why Kershaw’s z-score is nearly 50% better than the second-best pitcher, and more than 100% higher than the seventh best pitcher, despite an apparently reasonable projection. Kershaw certainly could lap the field in all these categories as Steamer projects, but that is the optimistic bound of outcomes. If we adjust his relative projections down, the case for him as the No. 1 overall pick loses strength.
The Sample Size for Our Calculations May Affect the Relative Value of Pitchers and Hitters
The z-scores were based on 120 starting pitchers versus 250 batters. The question is whether the relative number of batters versus starting pitchers was off, such that the sample size unnecessarily favors one versus the other. If it favored pitchers relative to hitters, as ADP suggests, that would also weigh against Kershaw as number one.
Conclusion
Using z-scores, Kershaw is the clear No. 1 pick. However, due mainly to injury risk and partially to the luck involved in win-loss record, I adjust the pitchers downward. While I have Kershaw close to Trout, I have him third due to how highly I value Mookie Betts. Having Kershaw third also allows me to be logically consistent with where I ranked my other pitchers, whereas putting him first would require an upward adjustment of the others.
