The Rockies Interesting Rotation Experiment
The Rockies rotation has been a disaster this year. As a group, they have a 6.31 ERA, and while their FIP (5.15) and xFIP (4.44) are somewhat better, both marks are still among the worst in the league. The main culprits — the rockies are 29th in walk rate (3.71 BB/9), 29th in home run rate (1.58 HR/9), and 30th in BABIP allowed (.349). Putting guys on base before you give up hits and homers is a sure way to allow the other team to score, and while the park and defense are significant non-pitching factors, their starters xFIP- is 114, worst in baseball, and that’s all on the pitchers.
So, Jim Tracy is going to try something different. Very different. In lieu of just changing the names on the back of the jerseys, Tracy is essentially throwing away the standard construction of a starting rotation, and is going to use his pitching staff in an extremely unconventional manner — four “starting pitchers”, each one limited to 75 pitches per outing.
To be honest, I love the idea of trying something different, but I’m just not sure this roster and this location is the right place to try it.
Let’s start with the logic behind the idea, though. We know that relief pitchers perform better than starting pitchers, and in general, a pitcher who knows he doesn’t have to conserve his bullets for a longer outing can throw more pitches at max effort and see an uptick in the quality of his stuff. That jump in velocity and movement, along with more frequently getting the platoon advantage and not having to face the same hitters multiple times in one game, helps relievers allow about one run fewer per nine innings than starting pitchers. It’s a significant effect, so transferring innings from starters to relievers could allow a team to reduce the amount of runs they allow.
Indeed, the ideal pitching staff probably doesn’t look like the current five starters/seven relievers format, and I wouldn’t be surprised if teams eventually move away from that system and into something that takes more advantage of the effectiveness of pitchers in shorter stints. However, there are some issues specific to the Rockies experiment that makes me think that this probably isn’t the right long term solution either.
For one, the fact that the shorter outings are being combined with one fewer day of rest means that the Rockies starters probably won’t see a significant uptick in stuff. The Rockies are essentially combining two offsetting effects here – the reduction in quality of stuff from pitching on three days rest with the improvement in stuff from not having to save your arm for pitches 76-100.
Overall, the starting pitchers will still be throwing about the same number of pitches as they were under the standard five man rotation, only spaced slightly differently. Whether its seven or eight outings of 75 pitches each of five or six starts of 100 pitches each, the Rockies are still looking at getting between 500-600 pitches per month from their four remaining starters. While we don’t know for certain that relief pitchers can throw harder because they also throw fewer pitches, it seems that the quantity versus quality trade-off is more likely to come to fruition if there actually is a reduction in quantity.
Perhaps the Rockies will find that pitchers can throw harder for the same number of pitches if just spaced out differently — it is certainly worth watching to find out — but my guess is that Christian Friedrich, Alex White, Jeff Francis, and Josh Outman aren’t going to see a significant impact on the quality of their stuff under this new system. And if they’re not throwing better pitchers, then the entirety of the rotation’s improvement would have to be based on each pitcher getting to face opposing hitters fewer teams within the same game.
That could have a real (but probably small) impact on the starters’ performances. Here are the splits for each time through the order in MLB this season:
1st PA vs SP: .247/.310/.393
2nd PA vs SP: .260/.321/.417
3rd PA vs SP: .271/.332/.444
(We’re ignoring 4th time through the order because of huge selection bias issues – a guy only gets to face batters a fourth time if he’s pitching really well.)
Just for reference, the first PA against a reliever in a game is .241/.316/.375 – not that much different than the first PA against a starting pitcher, which does suggest that a starting pitcher going through a batting order once might not perform that differently from a reliever. But, the Rockies aren’t putting their starters on a once-through-the-order limit, as 75 pitches per outing means they’re probably looking at about 20 batters faced per game, or two full trips through the order and then a couple of hitters a third time. So, if we just reallocate number of PAs based on these averages, what size impact could we see?
This year, the average starting pitcher is facing 25 batters per start, so they’re basically facing the order all the way through twice and then facing seven batters for a third time. So, if we weight the average OPS by those situations as 9/9/7, we get the .737 OPS that is basically a dead on match for what starters are allowing this year. If we re-weight that to 9/9/2, to account for 20 batters faced per game, the average OPS allowed would drop to .727. In other words, having to face those five extra batters a third time through the order pushes a pitcher’s OPS allowed up by about 10 points. If nothing else changed, that would knock about .05 runs per game off the Rockies runs allowed total. A real change, but a small one.
Of course, that assumes all things staying equal, and this structure doesn’t make that likely. Fewer batters faced per game by the starters means more batters faced by middle relievers (or in this case, Jeremy Guthrie), so the trade-off will be switching from having batters face a starter for the third team to batters getting to face a bad pitcher. Guthrie might see some boost from pitching out of the bullpen, but again, he’s going to function more like a long reliever than a setup guy, and those pitchers don’t see the same uptick in stuff as arms who can come in and throw max effort for 15-20 pitches at a time.
So, the overall net benefit to the Rockies under this structure seems likely to be negligible. They don’t have a strong deep bullpen of guys who can come in and dominate for a couple of innings each before handing it over to the next guy, which is really what you’d want in order to make something like this work. They also play in a stadium that is notorious for inflating offense, meaning that they have to face more batters and throw more pitches than an average team when they’re at home. By taking out your starter after just 75 pitches, you’re essentially guaranteeing that you’ll burn through your bullpen in a high scoring affair, and either not be able to play match-ups at the end of the game or be completely screwed if the game happens to go to extra innings.
Ideally, a team trying this kind of radical departure would play in a very low scoring environment, reducing the number of batters faced per pitcher and letting them go deeper into games with their lower pitch counts. San Diego, for instance, would probably be a much better testing ground for this kind of strategy. But even then, the organization that tried it would have to deal with the fact that they’re essentially stripping away starters abilities to rack up wins, which affects their actual paychecks in arbitration, and there are financial realities that are tied to the five man rotation that can’t be completely ignored. If a team is going to abandon the structure of the current rotation, they’ll also have to figure out a way to properly compensate their pitchers under the system that they’re employing.
I love that the Rockies are willing to abandon tradition and try something new. I’m not sure they have the pitchers or the environment to make it work, and I’m guessing this will be a short term experiment that is abandoned by the end of the year. But, if it does fail, I hope that doesn’t preclude some other team from attempting to overthrow the notion of the strict five man rotation in the future. The five man construct is probably not the best way to manage a pitching staff. I’m just not sure the Rockies way is a real improvement either.
Dave is the Managing Editor of FanGraphs.
This is something I suggested to Rob Neyer years ago, and he ignored me.
I suggested / thought if there were 12 pitchers on a staff, you could use four to six per game at two innings each, you’d have more fresh pitchers available and more ability to manage a game and keep your opponents confused. Your team could benefit from having your ace pitcher start 2/3rds of the games for 2 innings each – same number pitched, but it gives you the chance to get ahead in a game more often. Many lineups improve the second time through the order, so wouldn’t it make sense to prevent this a bit?
Anyway, I can understand if it sounds stupid, because there is probably some logical reason I don’t understand as to why a pitcher has to throw this many innings, this many pitches, etc.. I mean, beyond “that’s how it’s always done”. Right?
I suspect I’m completely wrong but I never understood why.
Doing something that rigid makes it hard to take advantage of match-ups and leverage. You’d either need to have a bullpen full of guys who can get out both LHBs and RHBs (harder to find than you might think) or you’ll be giving up the platoon advantage in critical situations because you need the pitcher on the mound to work a couple of full innings, regardless of who is due up.
There are also going to be scenarios where the game gets out of hand early, and you’re better off just giving your good pitchers the night off to save them for stretches where you have multiple close games in a row. By not having anyone on staff who can just soak up innings, you’ll end up reallocating innings that have no impact on wins and losses to your better pitchers in order to keep everyone on their strict usage patterns.
There are positives and negatives to such an approach. Doesn’t make it stupid, but ideally, you want a plan that minimizes the drawbacks.
It doesn’t have to be super rigid, I’m sorry if that’s how I meant it, I just meant to break the mold of 5 starters, 7IP each. Switch it sideways a bit, make it more like the AS Game in a series than not.
Anyway, your reasoning makes sense, I just thought that with a lot of thought, something of the sort could be done to great benefit instead of the current “established” method, in case it wasn’t the best one; which is something I’ll always consider.
How about an early minor league style piggy backing rotation?
8 “starters” who work off each other. 4 pairs of 2 (obviously), where one starts, the other relieves, for 3-4 innings each. You have another 3-4 situational guys, a closer, RH and LH set up man, and a long man if someone gets hurt, or is off. Same 12 man staff, but different set up.
I don’t think it’s true that “the entirety of the rotation’s improvement would have to be based on each pitcher getting to face opposing hitters fewer teams within the same game”.
Another benefit is that whoever the 5th starter is is probably terrible and he is not part of the rotation. He becomes a reliever of either high leverage if he’s good as a reliever and low leverage if he’s not.
I think the big problem with rotations is that there is such a difference in the quality and durability of starting pitchers that it’s tough to shoehorn them all into a “one size fits all” system. The 5 man rotation works because it accommodates those who can pitch into the 7th/8th regularly and also accommodates 5 inning pitchers by having relievers available in any game situation. Anything that reduces the flexibility of managers’s use of relievers is going to cause problems IMO.
@rudegar- would you really wanna do that if you had Justin Verlander or Cliff Lee, Roy Halladay and Cole Hamels. Pitching any of those guys for four innings once every 4 days would be a disservice. That’s162 innings per year per guy. That does not even get into giving subpar pitchers 4 innings at a time.
A “platoon advantage” in most parks derives from breaking balls – ie a slider thrown by a lefty has a different platoon effect depending on the batters handedness. Coors itself, because of altitude, negates much of the value of breaking balls – all breaking balls – and any platoon split there is going to be based simply on the effectiveness of their other pitches.
There are plenty of pitchers who just don’t have very good breaking balls – but who have a good repertoire of changeups, sinkers, cutters, splitters, etc – the pitches that mess with the timing of a batter’s swing (regardless of handedness) rather than relying purely on ball location to mess with the batter. the Rockies have never really figured this out – and have thus not ever gone hard after the pitchers who really can pitch well at Coors without the standard “platoon splits” meme getting in the way.
But regardless – shorter outings will always work better at Coors than elsewhere because it can maximize fastball velocity and give batters fewer chances to “figure things out” with a particular pitcher.
@whatever I fully understand your point. One would not want to limit the truly elite pitcher. But how many of those guys are their? Also, think about how much they cost? A less conventional approach (though perhaps not as extreme as 4-6 guys every night) could give you almost as many high quality innings at a fraction of the cost. Further, it’s generally not the best idea to structure a strategy around needing to acquire a very rare commodity (Verlander, Halladay et al)…which is why I think a shift in pitching staff strategy could do some teams (esp. the poorer/cheaper ones)