The Most Vulnerable Rotations in MLB
Every team starts the season with five starters, but the chances a team makes it through the season unscathed is minimal. From research performed by Jeff Zimmerman, starting pitchers who made 20 starts and pitched 120 innings the previous season have roughly a 40% chance of hitting the disabled list during the season. Eno Sarris added to Zimmerman’s work before the start of last season and found that teams have been averaging around 10 starters per season since 2011. That average did not change last season as teams averaged 9.6 starters per team. Even if the bar is raised to a minimum of five starts, teams averaged 7.4 starters per team in 2014, consistent with five-year average of 7.5 starters per team per season.
Attempting to measure pitching depth can be difficult. Jeff Sullivan took a look at overall depth last month, showing how many players were projected for at least one win according projections. Focusing on rotation depth the same way is a good way to look strictly at depth, but with the season upon us, we can see the potential impact of that depth by examining the dropoff from starters to their replacements. FanGraphs completed the Positional Power Rankings with the pitching split into two posts, one for the bottom half, and one for the top half. Without rehashing the entirety of both posts, here is the graph from those posts which should provide a base of expectations for a team’s rotation.
The power rankings are based on the FanGraphs Depth Charts, where the number of innings are picked by humans based on our understanding of how the season is expected to play out. Current injuries play a role as does the risk of injury for some pitchers who are not as likely to pitch a whole season. In every team’s projection, the top four pitchers all ended up with at least 130 innings pitched. That is fairly realistic as exactly 120 pitchers had at least 130 innings pitched last season. The fifth starter can vary greatly with the number of innings past number five dependent on how many innings are expected from the top of the rotation.
Given the innings projected for the first four starters in every rotation, determining depth by separating the top four pitchers from the rest of the rotation can bring some clarity to depth and a potential dropoff. The fifth starter is not likely expected to carry a heavy load on most teams, and could easily fall out of the rotation based on performance, health and the potential of the replacements in the farm system. Using WAR, like above, will provide a good combination of innings and performance, but in measuring depth, innings are the question mark and might not aid a useful answer. Using FIP should provide a good measure of the talent at the top of the rotation, and help determine how far a team will fall should they lose someone for an extended period of time.
Here is a graph showing the average FIP of the top four starters for every team.
The graph above is a rough approximation of the positional power rankings. The top four teams are the same, albeit in a slightly different order with Kershaw factoring in heavily and Gio Gonzalez not factoring in at all. The bottom four teams are the same as well with the Colorado Rockies outlook in 2015 not looking great.
To attempt to make a determination on depth, using the next four starters should provide a pretty good idea of what teams are looking at if they have to make a move in their rotation. While some teams have very good fifth starters expected to make an impact this season, they are included as depth given that a good fifth starter is a luxury item that most teams do not possess. In the FanGraphs Depth Charts, the Washington Nationals were the only team to have a fifth starter projected to exceed 140 innings with only a handful of teams even pitchers expected to reach 120 innings.
As opposed to just looking at the fifth or potentially sixth starter, I am going down to eight starters because on average teams use between seven and starters for at least five starts every year. The remaining starts are likely to be divided among the sixth, seventh, and eighth starters anyway as teams try out different options for pitchers who might lack experience in that role.
Here is the average FIP for every team’s fifth through eighth starters.
With Gio Gonzalez listed as the fifth starter and Tanner Roark available if someone should falter, the Nationals depth appears prepared to aid the team if anyone should falter. The St. Louis Cardinals have a host of young options available to them and the New York Mets have Noah Syndegaard grounded for the moment. The San Francisco Giants have experience with Ryan Vogelsong and the potential of Yusmeiro Petit should the need arise, but the Cincinnati Reds and the Chicago White Sox do not appear to have a lot of help at the end of the rotation with Jason Marquis and Hector Noesi at the end of the rotation, although Carlos Rodon could perform better than his current projections indicate.
The last graph above is a good representation of who has the best depth, but does not necessarily say which team is the most vulnerable to a loss. Subtracting the FIP of the first four starters from the FIP of the last four starters helps tell which team’s projection could suffer the most if a top starter is removed from the equation.
It would be easy to say that the best teams are the most vulnerable, but that does not look to be the case. The Reds and the White Sox have middle of the road rotations, but they are heavy at the top. The negative numbers from the Kansas City Royals and Miami Marlins show that those rotations could improve as the year goes on if they get meaningful contributions from Kris Medlen and Jose Fernandez, respectively. Of the top rotations, Washington, St. Louis, the San Diego Padres, and the Cleveland Indians look pretty good to weather any storms, but the same is not true of the Dodgers or the Detroit Tigers.
The Dodgers already have concerns about Hyun-Jin Ryu, and should anything happen to Clayton Kershaw, Zach Greinke, or Brandon McCarthy, the team could be in some trouble. They have a little bit of a cushion in the National League West, but they are already hoping Brett Anderson can pitch well. Relying on Joe Wieland, Brandon Beachy or Zach Lee for an extended amount of time could be problematic for a team expected to repeat as NL West champs and contend for a World Series title.
Craig Edwards can be found on twitter @craigjedwards.
This is interesting, but it’s hard to say how valuable this exercise this is. Could you repeat this for the 2014 depth charts? Then you could compare those numbers to how the 2014 results varied from the projections.
Did depth (defined as average FIP for projected starters 5-8) actually have any relationship to the errors (actual-projected WAR) from last season’s projected starters? Or is this just an interesting exercise that doesn’t really tell us anything (useful) that we didn’t already know?
I agree. This is something that might make you say “hmmm” but it seems a bit rudimentary. First, I don’t know why you’d use projected FIP instead of projected ERA. Second, it would make sense to weigh the 5-8 pitchers depending on their projected innings. A crappy 8th pitcher just isn’t that big of a deal. You can’t compare that to your 5th starter or primary backup.
2014 would be a data point of 1 and wouldn’t be of any actual use. Injuries involve a lot of variance.
“Or is this just an interesting exercise that doesn’t really tell us anything (useful) that we didn’t already know?”
Oh I’m sorry, did you already know the average drop off each team would experience should one of their starting pitchers get hurt? Did you make a community post on this already?
This is a pretty silly comment. Nothing we read here is actually useful to us, as we don’t work for a team, and if we did we’d probably have better metrics than even FIP and fWAR.
I for one find it interesting to know which teams will be in the best/worst spot should they experience multiple injuries to their rotation, and there wasn’t a place where that information was organized before. The depth charts hold all that information of course, but it needed someone to take that information and organize it in a way that shows us which rotations would experience the steepest drop offs.
Are you Craig’s mom or his dad?
He wrote for my favorite Cardinals blog before he came here. But I would have wrote this response regardless of which of the fine fangraphs writers wrote the article, as it had much more to do with the stupidity of the comment than the author he was critiquing.
Maybe I just get too annoyed at stupid comments, which I admit is my own shortcoming since I voluntarily read them.
I am. Nice job, Son.
I guess you found use in the post in fulfilling your interest. I doubt it has much predictive value though. I looked through the Fangraphs depth chart and despite having a hundred or so people on the payroll, it doesn’t seem to reflect that.
Really interesting analysis. It seems to me that this addresses a large part of how team projections can vary – you can’t assume that every team has the same backups stepping in.
Now, if you could project which rotations were most likely to lose part of their staff, that’d really be something ….
The thing is, the projections do try to cover that somewhat — that’s why some teams have only five or six starters projected to get >80 innings, and other teams have ten or twelve. Obviously, projecting injury is still really rough, but the projections do a better job than they would if they just assumed every starter in the first five pitched the whole season, or had the same injury risk.
Awesomely interesting article Craig. I am very happy to see the Cardinals as 2nd in FIP among the 5th-8th starters. Really shows that they were right to trade from depth in the Heyward/Miller trade, and they were right not to be seriously involved in the Lester/Scherzer/Shields/Hamels markets.
Perhaps you could combine these with Jeff’s injury likelihood projections to see which teams are most likely to see a steep drop off in SP WAR from their preseason projections.
Sullivan and Cameron are on the boat that the Cardinal’s rotation has way to many question marks and could fall apart, in deciding the disagree with nearly every projection system that sees the Cardinals as a great to elite team. They seem to be ignoring the significant depth they have.
Although I know Dave hates Carlos Martinez and thinks they need to stick him back in relief where he belongs. Note sure why he has this opinion, as multiple articles this site have shown Martinez has some of the best stuff and pitchability in baseball.
Martinez does indeed have the “stuff” to succeed as either starter or bullpen piece. But command & pitchability, I’m not so sure. I believe he used nearly 19 pitches per inning last year when starting (per a recent writeup at VEB), and that’s just not efficient enough to be an MLB starter for long….
If this post accounted for emotional vulnerability, I suspect Seattle’s rotation would fall considerably in the rankings.
…what?
If it accounted for vulnerability to boat propellers, their farm system would fall also.
You appear to have hit “j” when you meant to hit “l”. Cheers to jokes about a life ended tragically soon, jackass.
No clue how the White Sox rotation is middle of the road.
Their #3 has a better WAR than Lester since he entered the league. Their #1 is the best pitcher in baseball not name Kershaw and their #2 was one of the 10 best pitchers in baseball last year.
Seriously. And then if they’re middle of the road, how can they be that vulnerable? On top of the fact that the depth charts projections have the combination of Sale, Samardzija, and Quintana for 10.6 WAR, 3.7 fewer WAR than they produced last year, apparently losing either of Noesi or Danks would be a net positive. How can they be vulnerable if 40% of their starting rotation is worse than or equal to the remaining four starters in the system (Rodon, Penny, Johnson, Montas)? Rodon doesn’t even need to outperform his FIP projection to be an improvement over Noesi or Danks, yet the author still qualifies his statement about the White Sox with, “although Carlos Rodon could perform better than his current projections indicate.”
Which is it, guys? If the projections are the null hypothesis, as Dave Cameron has argued, then Rodon is better than 40% of the rotation right now. If they’re not, and we’re allowed to use common sense in evaluating a player’s potential contribution, then the rotation is about 4 wins better than the projections suggest.
It’s obvious that the White Sox rotation is extremely top heavy, but the projections that some writers on this site have inexplicably refused to deviate from in discussing the rotation don’t bear that out. They can’t be both middle of the road and vulnerable.
“They can’t be both middle of the road and vulnerable.”
Just look at the depth charts, it’s pretty self explanatory. They have 4 above replacement options, and one of those is very marginally above replacement and probably won’t even be in the majors until mid-season. They have two guys starting the year in the rotations that are sub-replacement level. Three pitchers can’t carry a rotation for 162 games. Think about how bad it will be if one of the 3 actually good pitchers goes down. How can you not see that as vulnerable?
Now compare that to say the Cardinals, who are 4th on the depth charts and are ranked as the 2nd least vulnerable. They have 9 above-replacement options.
“If they’re not, and we’re allowed to use common sense in evaluating a player’s potential contribution,”
I’ll always take the projections over common sense. Time and time again what most uneducated people think is true is shown to be false.
So you would take Jose Quintana for 2.7 WAR per 200 innings, then? And Samardzija for 2.9 WAR per 200 innings?
I’ll take the over on both.
I’ll also take the over on Noesi and Danks being below replacement level. If they actually pitch that poorly, they’ll be replaced. I think both will have ERAs at their xFIPs for last year of 4.29 and 4.62, respectively. If you buy these projections, you’re betting on 5.23 and 5.1.
Obviously, that’s a big difference. Both numbers are based on metrics, so which one would you characterize as “what most uneducated people think.” It’s not a question of education and proving people wrong. It’s a question of which numbers you choose to believe and why. You’ve defended the numbers used in this article by reasserting that both Noesi and Danks are below replacement level. Again, I’ll take the over.
Sure, they could both get injured and finish the season below replacement level after a bad start, but that’s not the most likely outcome. I think the most likely outcome is that they both continue to produce at somewhere between replacement level and league average, though I admittedly have a lot more hope for Noesi than Danks.
I “believe” both numbers, it’s just that one set of numbers is a description of what happened in the past, and is not a projection of future events.
The projections take a lot more data into consideration and has a much more systematic approach than either of us could ever hope to accomplish on our own. It certainly considers a lot more than just last year’s WAR/200, as that is an incredibly lazy approach.
Both Shark and Quintana’s projections are in-between their 2014 and 2013 numbers, though closer to their lower 2013 performances. Considering the boom and bust of pitching and how pitchers suddenly become worse or better all the time, the projections regress their performances to a degree. They do that because on the whole it works. They also do it for everybody, so you have no business complaining about it just for White Sox pitchers.
Noesi and Danks have been very, very close to replacement level over the last couple years. Their performances were no where near average. Sure you can just replace them, but as this analysis shows, there’s no one worth a damn to replace them with, which is an obvious problem and thus makes them “vulnerable”.
Well, Quintana does seem to break the projection systems, I’ll go along with that premise for sure. By either fWAR or bWAR, he’s been worth 9 wins over the past two years, and there’s *nothing* about his age or workload or velocity or anything else (I can see) to indicate that he’s ready for a dramatic dropoff in performance.
Wish Vegas (in the form of Bovada) had odds on Jose Q winning the A.L. Cy Young this year. Over 20 pitchers available for wagering this year, at odds all the way up to 100-1, and no Quintana!?! I would love to bet him at 100-1 or better. This guy gets no respect I tell ya, no respect at all.
jfc you’re the White Sox everdiso
I notice you didn’t mention the 4/5 starters. Enjoy those 3 wins seasons, not WAR, they should both go about 3-15.
Just looking at it from the view point of the Yankees, 1-4 FIP seems way too high and 5-8 way too low. Of course, I have CC has the 5th starter
Is toronto at ‘0’ because they don’t have any 6, 7, 8 starters? =p