Re-imagining Starter Depth Charts & Projections

The idea for this article grew out of an intriguing conversation I had with Cardinals fan Chuck Brownson on Twitter. I am glad to give him full credit for this idea, and I’m sure he would appreciate a follow.

What Chuck is talking about here is Fangraph’s Depth Charts page for starting pitchers. Depth Charts takes ZiPS and Steamer projections and then re-organizes them by position per team to reflect likely playing time. This is a useful tool as player projections like ZiPS prefer to focus on the player itself and less on the actual context of their playing environment. Full ZiPS projections for the Cardinals are not out yet, but Symborski has released information on two key parts of the rotation: Jack Flaherty and Kwang-Hyun Kim (who prefers to be known as KK). Those projections, combined with the Steamer model provides enough data to make some determinations on how the Cardinals should align their starting rotation and draw some conclusions about how the club’s rotation compares with the rest of baseball.

Let’s start with the current Depth Charts projections for Cardinals starters:

Depth Charts essentially has the Cardinals with four regular starters — Flaherty, Mikolas, Hudson, and Wainwright — and assumes that some combination of Ponce de Leon, Reyes, and Gomber will fill the fifth spot with just above replacement level production. Knowing that the Cardinals’ preference is for Carlos Martinez to move back to the rotation, Depth Chats adds him in as an after thought, with 19 innings over just 3 starts and .3 fWAR.

The total is 12.2 projected fWAR — over half of which comes from Flaherty and Mikolas — in 940 innings, or roughly 5 ²⁄₃ innings per start for 162 games played. That currently ranks as the 17th best projected rotation in the game for 2020 by fWAR. Not good.

In 2019, Cardinals starters collectively ranked 15th in the baseball and achieved that ranking with less fWAR produced than their current projections: 10.9 actual versus 12.2 projected. So, Depth Charts thinks the Cardinals rotation will be better in 2020, but only enough to drop a few spots and land in the bottom half of rotations. There are obvious issues with the way Depth Charts has allotted playing time and questions that should be asked of their individual Steamer projections. Let’s explore these projections, make some adjustments to the depth chart, and see how the Cardinals rotation shakes.

Jack Flaherty - 3.69 FIP, 4.2 WAR
I really want to be annoyed by Flaherty’s projections, but I also realize that my angst is entirely irrational. For half a season, Flaherty was one of the game’s best pitchers. His .91 second half ERA is the third best post-break performance since the start of the century (which is about as far back as Fangraphs’ splits go). Why not project Flaherty’s stellar second half for a full season and give that man the 8 fWAR that he deserves! That’s just not how any of this works. A 4.2 WAR projection is actually really strong. Despite his disappointing first half, Steamer is still willing to name him a top-15 starter in the game, placing him within spitting distance of established aces like Strasburg and Sale.

If you’re still not happy with that, ZiPS founder Dan Szymborski throws you a bone. ZiPS has already been published for Flaherty and it’s a bit more optimistic. They have the 23-year-old at 4.4 fWAR with an outstanding 3.13 ERA. Also, Flaherty did produce 4.7 fWAR. Even if he can’t duplicate his 2019 second half, there are reasons to think that Flaherty could outperform his Steamer projections. Still, we’ll play along and leave it as it is.

Dakota Hudson - 4.55 ERA, 1.5 WAR
Hudson is such an odd case that he really deserves his own article. Look for that down the road. For now, the problem with Dakota is that his peripherals — FIP, BB/9, K/9 — did not match up with his actual performance — ERA. He had a low K/BB ratio, a high LOB%, and a unexpectedly low BABIP and HR rate. Fangraphs penalizes Hudson heavily for these peripherals, giving him an atrocious 4.93 FIP in 2019 and just 1 WAR in 174 innings. His actual ERA, though, was a really good 3.35. So, which is it? What Hudson pretty great? Or pretty awful?

The problem with Hudson’s projections is that he always — and I mean ALWAYS — outperforms his FIP. At no point in the minors did Hudson’s ERA match his FIP and he routinely beat his FIP in earned run average by over a run. What are we supposed to make of this? Personally, I’m good with his current 1.5 WAR projection. That’s a nice midpoint between his ERA-centered brWAR from Baseball Reference (2.2) and his FIP-centered WAR from Fangraphs (1.0).

Kwang-Hyun Kim - 3.89 era, 2.5 fWAR
VEB writers have already written brilliantly about KK. If you want more information about the Cardinals recent addition from Korea, you can look here, here, and here. I’m here for the projections, which were recently released by Szymbroski and Fangraphs. Take a look:

That’s such a fantastic projection for KK. For the Cardinals to acquire 2.5 fWAR for about $4M is an incredible steal. Of course, to have a chance at reaching 2.5 WAR, KK will have to start. Will he? At the introductory press conference, Mozeliak predictably danced around questions about his role. KK himself prefers to start, but it’s obvious he will have to earn his role in spring training.

What does that mean for our Depth Chart projections? At this point, I’m inclined to give preference to production and KK’s 2.5 fWAR in 157 projected innings places him as our third best starter in production and second best by fWAR produced per IP. For now, he makes my projected rotation with the projection offered above.

Carlos Martinez - .3 fWAR in 19 projected innings
Depth Charts is probably right about one thing with Martinez: it’s either all of nothing. Shoulder issues forced Martinez out of the rotation in 2018. His struggles to regain strength in his throwing arm were well documented last winter. The result was a late start to 2019 and banishment to the bullpen, where he was a really excellent closer. The Cardinals and Martinez both want to see a return to the rotation and reports from the Dominican and Jupiter are encouraging. If Martinez struggled to stick with his strength-training regimen in 2018, such does not appear to be the case now.

So, what do we do with him? Depth Charts essentially writes him out of the rotation and elevates Ponce, Reyes and Gomber into a platoon in his spot. That’s fair. But, its not the Cardinals current plan. What happens to rotation projections if we give Martinez a starter’s workload? Because Steamer projects Martinez as a reliever and 2020 ZiPS is not yet out, the best I can offer at this point in the season is his 2019 ZiPS projections. Enjoy these!

How can you not love that Bob Gibson comp for Martinez? I mean, it’s a crazy comp, but it’s crazy in the most awesome way. For that comp alone Martinez has to be in the rotation, doesn’t he? Of course he does!

Despite working out of the pen all of 2019, Martinez did not show any ill-effects of his shoulder weakness. His peripherals in the bullpen were comparable to his previous stints in the rotation, and better in some critical areas. I feel relatively confident that if Martinez can stay healthy enough to provide 168 innings, his stats would be similar to those projected by ZiPS heading into last season.

Re-Imagining a Depth Charts Rotation

Where does that leave the Cardinals and their rotation? Let’s return to the original chart provided by Depth Charts and make a few VEB updates to it. First, to keep it in ratio with the current chart, innings will needs to be adjusted downward in order to stay in the 935-955 range. To do this, I will retain the projected innings of the top 3 starters by WAR and then cut the rest by percentage to reach the innings cap. This mostly affects Hudson and Wainwright and pretty much eliminates Gomber from the conversation. I will also re-calculate fWAR for the players whose innings I cut, so the original ratio of fWAR/IP is maintained.

The new totals are encouraging. With that playing time allotment, the Cardinals jump to an amazing 15.2 fWAR in 949 innings. 15.2 fWAR would be the 4th highest SP fWAR since 2000, behind ‘09, ‘15, and ‘12. Obviously there are issues with this exercise. Injuries and ineffectiveness are always problems. Perhaps we can view this as a “if everything goes to plan” scenario?

Finally, to accurately rank the Cardinals compared to the rest of the league, I would need to make similar adjustments to other teams. Who has time for that? So, please view the starter projections for the rest of the league — and the Cardinals ranking among them — with healthy skepticism. Compared to the projections I offer above, a 15.2 fWAR from the Cardinals starters would rank 10th in projections by starting pitcher per team, between the Indians and Red Sox. Will this actually happen? Your guess is as good as mine. Perhaps the take away is that the Cardinals rotation has more production potential built into it than what the club fielded last year. It’s absolutely necessary for Martinez, if healthy, to be in the rotation, regardless of any needs in the bullpen. KK should also get every opportunity to prove his ZiPS projection is real. Hudson also needs to continue to outperform his FIP. Oh, and I’ll just go ahead and put Flaherty down for 8 fWAR anyway. Ba humbug, Steamer projections! Merry Christmas, Cardinals fans.