The Problem With Only Focusing on Improvements
A few days ago, Shannon Drayer wrote a post about the Mariners potentially pursuing Robinson Cano, based on comments made by Jim Bowden. The Cano rumors don’t interest me much, because I don’t think there’s any reason to believe the Mariners should or will go after Cano, nor do I believe that Cano would have any interest in relocating to Seattle, and I think the idea of a big free agent signing turning around a franchise’s reputation is pretty much 100% BS. But in that piece, Shannon wrote another thing that was a little more interesting, and something I think is worth mentioning.
If you read my blog on a regular basis, you know I hate to make predictions. I will predict this, however: The Mariners’ No. 3, 4 and 5 starters will be significantly better next year. I know I am going out on a limb, but James Paxton and Taijuan Walker will be an upgrade from 3, 4 and 5 and most likely 6 on that list above. General manager Jack Zduriencik is planning on adding a starter from the outside as well. Great. Add a pitcher, do not trade Paxton or Walker and you can pencil in (I am done with my predictions so we are going with “pencil in” here) a 100-run swing.
Zduriencik has said that upgrading the defense is a priority as well and there is a lot of room for improvement. That translates to runs saved, which you can tack on to that 100-run swing. Go ahead and add a few more for an improved bullpen as well. That 754 runs allowed in 2013 should come down significantly in 2014.
She’s right that the back-end of the Mariners rotation last year was dreadful. Whether you’re looking at Joe Saunders, Aaron Harang, Brandon Maurer, Jeremy Bonderman, Blake Beavan, or even Erasmo Ramirez, the results were lousy. A lack of starting depth was one of the main reasons the 2013 Mariners were terrible. It should not be hard to improve upon what the team got from those three spots in the rotation. If Walker and Paxton are what some people think they are, and the team acquires a “legitimate #2” — or someone they’ll stick that label on, at least — then the 2014 rotation should project to be significantly better.
But Shannon makes a pretty common mistake that a lot of people make when projecting the future; she focuses only on the positive improvements from replacing lousy performances from the year before. When you do projections like this, and note that Awesome New Guy X is some number of runs better than Old Crappy Guy Y, you’re inherently treating everyone else on the roster like their performance is fixed from year to year. And that is simply not the case.
As great as Felix Hernandez is, and as good as Hisashi Iwakuma was last year, those two simply cannot be expected to repeat their 2013 performances again in 2014. It’s not that they couldn’t possibly throw another 423 innings while allowing just 143 runs (3.04 RA9, combined), but that their performance from last year represents something very close to the upper limit of their potential, and there’s a significant probability that the Mariners will get less from their top two next year. And you absolutely have to factor the expected regression from those two into any kind of forecast for runs allowed by the team in 2014.
For instance, right now, the Steamer projection system forecasts the Mariners rotation to post a combined 4.07 ERA in 969 innings, barely any improvement over the 4.18 ERA the Mariners got from their starters over 960 innings last year. Part of that lack of improvement is because there is no “#2 starter” included in the depth chart yet, so Shannon’s projecting some improvement from a pitcher the team doesn’t yet have, so you could go ahead and make some adjustments for adding that guy to the mix. And I’d imagine she’s probably more optimistic about the short-term performances from Walker and Paxton than Steamer is, since that system is forecasting something close to league average pitching from both. No one is saying the Steamer projections are the gospel truth and can’t be underselling the expected performances of the team’s 2014 rotation, and perhaps Shannon’s right to be more optimistic about the young kids than the numbers suggest.
However, I think we can say with near certainty that the Mariners rotation will not improve by 100 runs next year. In fact, we can basically prove that they won’t do it just through looking at the recent history of major league rotations.
The Mariners starters allowed 481 runs in 2013. Over the last 30 years, no American League team has allowed 381 runs over a full-season — a bunch of teams did it in 1994, when the season ended in mid-August — and in fact, no team has even come close. The fewest runs allowed by an AL starting rotation over the last 30 years? 412, by the 1990 Red Sox. That was a team that had Roger Clemens throw 230 innings with a 1.93 ERA, and a bunch of good hurlers behind him.
A few other teams have gotten close to that mark recently, including the 2012 Tampa Bay Rays (David Price, James Shields, and a bunch of good young arms) and the 2013 Tigers (Justin Verlander, Max Scherzer, Anibal Sanchez, and Doug Fister), but have maxed out around 415 runs allowed. Those are two of the best run prevention rotations we’ve seen any team run out in recent history, and they topped out at about 65 runs better than the 2013 Mariners. Realistically, it would be impossible to expect the Mariners rotation (and defense) to be better than any of the recent Rays teams, last year’s Tigers, or even the ’85 Blue Jays, ’89 A’s, or that 2002 Red Sox team that featured Pedro Martinez and Derek Lowe.
Even if we narrow the timeline to the more recent years, as offense in baseball has declined and so comparing the current game to the one seeing 15 years ago is a little bit of apples and oranges, we still only find a handful of teams even getting under the 450 run barrier. Over the last three seasons, only seven different pitching rotations have allowed fewer than 450 runs, and five of those seven were between 432 and 442. If you were to look at the runs allowed totals by the best rotations (and defenses) we’ve seen in the last few years, we’d peg the expected upper limit around 435 or so. It’s possible to push it to 415, but it takes such a remarkable performance from so many elite talents that it’s basically impossible to expect anyone to match those levels.
435 runs allowed isn’t even a 50 run improvement over what the Mariners rotation put up in 2013, less than half of the 100 run improvement that Shannon wrote about. And that’s the level reached by the best starting staffs in the AL over the last few years, which it isn’t entirely clear that we should expect the 2014 Mariners to be. Yeah, Walker is a great prospect, and Paxton had a nice final start to his season, but the performance range of young pitchers is all over the map. Pitching prospects are as flakey as anything in baseball, and it’s not like either Walker or Paxton destroyed Triple-A in a way that suggests that the inconsistencies of young pitchers shouldn’t be expected to apply to them. There’s as good a chance that one or both of them just fall on their face — as Brandon Maurer did a year ago after winning everyone over in spring training — and get shipped back to Triple-A as there is that they pull a Michael Pineda and dominate from day one. The reason we were all impressed with Pineda is because that kind of performance from the start was unusual. You can’t expect that from every young kid who throws 95.
And even if you could, you’d still have to expect less from Felix and Iwakuma. They might stay perfectly healthy and make 64 starts between them again, but there’s basically no room for upside beyond that, and even short DL stints from either one could really cut into their overall production levels. And, realistically, a 2.66 ERA for Iwakuma is almost certainly not happening again. He needed absurdly low rates of hits on balls in play and stranding runners in order to post that mark last year, and those numbers fluctuate greatly from year to year. Even if he pitched the same in terms of walks, strikeouts, and home runs, you’d expect his ERA to go up significantly just due to different timing of events.
It’s fine to expect the Mariners 2014 rotation to be better. It might even be a lot better. But, in reality, a lot better is a 30 run improvement, not a 100 run improvement. If the Mariners really do commit to upgrading the defense, and the bullpen gets some positive regression as well, maybe the overall staff can be 50 or 60 runs better than they were last year. But anything beyond that is really pushing it. 100 runs just isn’t realistic.
Comments
28 Responses to “The Problem With Only Focusing on Improvements”
Leave a Reply
You must be logged in to post a comment.
While I agree with you that 100 is too many, I think there’s something you’ve overlooked–an improved staff should throw more innings as well, which would reduce RA by the bullpen. Last year, the sucky sixsome of Sanders-Harang-Maurer-Bonderman-Beavan-Ramirez averaged about 5.47 innings per start last year. If the improved rotation was able to cut runs allowed by 30 while getting the back of the rotation’s average up to 6 innings a start, that would take about 47 innings away from the bullpen (and presumably largely from the worst pitchers in the bullpen), which could reduce RA by around 25. Still not near 100, but maybe a good staff that allows 450 runs and gets 6 IP/GS from the back of the rotation could reduce total RA by 55. Thoughts?
So Shannon was a little over enthusiastic with her projection, it was just commentary not sworn testimony before the Supreme Court. This overly long rebuttal in defense of The Holy Number is much like a nit harvest.
So why say 100 if you don’t mean 100?
Sigh.
Projections suck the fun out of being a fan. I understand their use and accuracy, but to go into a season expecting all your players to suck a little more than last year just sucks.
So if the staff improved by 60 runs, they would be really good.
Short a record setting rotation or an overstatement by Drayer (maybe she should say ’47 runs’?) perhaps the realistic take away from this is: there is hope. This could be a tipping point.
A slight improvement in the rotation.
A significant improvement in the defense.
A slight or significant improvement in the bullpen.
A slight improvement in the offense.
This could mean a > 100+ swing in run differential…..
I didn’t read the piece as a specific criticism of Drayer, she was just the example cited of a common mistake.
I’ve seen that particular analytic error committed by literally dozens of commentators and have fallen victim to it myself.
I like Shannon’s thinking a lot more than last year’s strategy of let’s hit homers. I think an approach where you aim to give up less runs through improved pitching and defense could go a ways to creating a foundation of a .500 ball club.
I’m not sure Dave is criticizing her thinking – more just a discussion of the details about how that might or might not work.
Well, Dave has to do something now that he doesn’t have Geoff Baker to kick around anymore.
I like reading Drayer’s stuff, but analysis has never been her strong suit – she’s just too close to the team and is sometimes just a little too ready to accept what they say as gospel. However I’ll cut her some slack here by ignoring the specific numbers (which Dave has convincingly argued are way too optimistic) and by noting, like the other beat writers, she’s still tasked with generating stories regularly at a time when there’s just not much going on in Mariners-land. That’s hard work!
With regard to “significant defensive improvements”, I’ll believe THAT when I see the opening day roster and Raul isn’t penciled in as our starting left fielder. Seriously… Jack and the others can talk till they’re blue in the face; but in that regard I’m an honorary Missourian.
I can only hope the M’s will improve their outfield defense. We’ll see how Brad Miller does in a full season at SS and cleans up his mental errors. So with better 3, 4 & 5 starters and hopefully improved defense, we could see a huge improvement. Here’s to more competitive baseball!
Getting a #2 type certainly would help. Since Kuma was third in the league in the Cy Young that only leaves two guys that would be an upgrade over him. STOP the labels let’s get the best guy we can and add him to Felix and Kuma. Personally I think Kuma is our #1 unless otherwise proven so we “only” have to get someone better than Felix to upgrade our #2 (sarcasm). I would not mind seeing Paxton and Walker as our 4th and 5th guys to start the season. If Ramirez can regain (health) form and we add another that would be six arms for five spots which would find a way to whittle it’s way down to five. I will be disgusted if we sign another aging “he was great three tears ago” veteran as a starter (or anywhere). Improving the St. Louis Cards would be a tough task. Improving the M’s is easy, just don’t suck. Quit dissing Kuma we already have the Leagues best #1 and#2
This is a very good article in that it is instructive in a number of ways. First, Shannon’s error is an error that is made by a lot of mainstream media — not taking into account all factors that bear on a subject. ERA is the same idea, and BA, and pitcher wins. Second, that taking all factors into account that bear on a subject takes time to explain. The importance is including all factors, so if corners are cut to shorten the explanation, the explanation suffers. Third is that sabermetrics cuts against the grain of how Americans like their information delivered: short, sweet, and fast. Then move on to the next tidbit. Some things are worth taking time with. And last is that there is a wide range in understanding of sabermetrics even within the sabermetrics community. Whenever there is a range of understanding there will always be too much and too little written. No way to win in both directions.
If Shannon reads the article she will likely not see it as personal, and I’d bet she’ll take the ideas and incorporate them into her thinking. That’s a good thing.
The team’s wording of a #2 was odd, almost implying they undervalue Kuma or plan to trade him. But if they’re just talking about getting a good starter, then great. It’s too early to rely on Paxton or Walker.
Kuma’s great, but I’d bet on him regressing this year. It’d be hard to duplicate his 2013. I’d quibble a little on Felix. He had a very good season, but didn’t reach his ceiling. For the money we’re paying him, I sort of expect that performance and I think assuming a comparable performance from Felix in ’14 is not unreasonable.
to go into a season expecting all your players to suck a little more than last year just sucks.
What are you talking about? There is no reason to expect that from “all your players”. Felix overachieved a little, and Iwakuma a little more, so your 50% percentile projection for them is slightly less. But a) they’re not “all players”, just those who overachieved. (Just as you can expect more from last year’s underachievers) and b) it’s not like Dave or any projection system is saying it’s impossible for them to repeat their season. It’s just mildly (Felix) to moderately (Iwakuma) unlikely. But as a fan, we root for unlikely things all the time.
It should not be hard to improve upon what the team got from those three spots in the rotation.
Of course, we are talking about the Mariners, who managed to punt a bunch of defense in order to add 5 runs of offense last year.
Dave sounds absolute when talking about Felix and Iwakuma’s 2014 season campaign failing to equal or improve from what they did in 2013. He makes it sound like Felix is 35 and Kuma found success when never attaining it… ever. There is nothing presently, Dave, that suggests Felix and Kuma CAN’T have an equal or better year. They didn’t throw 20 combined complete games, they didn’t each throw a few no-hitters, and Felix fell off quite a bit at the end of last year. Always so doom and gloom with your posts Dave… Cheer up!
Dave sounds absolute when talking about Felix and Iwakuma’s 2014 season campaign failing to equal or improve from what they did in 2013. He makes it sound like Felix is 35 and Kuma found success when never attaining it… ever. There is nothing presently, Dave, that suggests Felix and Kuma CAN’T have an equal or better year. They didn’t throw 20 combined complete games, they didn’t each throw a few no-hitters, and Felix fell off quite a bit at the end of last year. Always so doom and gloom with your posts Dave… Cheer up!
djw… would love to hear your explanation on how Felix overachieved. His stats were better when he won the CY Young, and he’s still so young and healthy. Just doesn’t make sense to me that a guy who is looked at as such a great pitcher all over the entire league goes 12-10 with 31 games started, NO complete games, and a ERA of 3.04…… and he overachieved. In my opinion he underachieved and fell off late last year.
Another way of looking at what Dave is saying is simply this:
If the M’s replaced three starters in last year’s rotation with three new ones who projected to X fewer runs in the same number of innings as last year’s starters, that doesn’t mean you automatically project the M’s rotation, as a whole, to give up X fewer runs, because you can’t simply pencil in Felix and Kuma to do the same thing they did last year. The likeliest result for them is that they will do worse.
That doesn’t mean they will. But in designing the team for next year, a projection that they will be marginally less wonderful than they were in 2013 is the more realistic and appropriate way to plan.
Dave sounds absolute when talking about Felix and Iwakuma’s 2014 season campaign failing to equal or improve from what they did in 2013.
You’re not thinking this through. He isn’t saying it’s impossible they’ll repeat, he’s saying you need to peg projections at the 50% percentile outcomes, and they exceeded those last year.
There is nothing presently, Dave, that suggests Felix and Kuma CAN’T have an equal or better year.
Of course they can. It would be absurd to suggest they “can’t” be better–that it is impossible. No one is suggesting that. Your don’t project a best case scenario, or even an 80th percentile outcome scenario. You project a 50th percentile scenario.
would love to hear your explanation on how Felix overachieved. His stats were better when he won the CY Young, and he’s still so young and healthy.
This is an incredibly easy question to answer. Your factual claim is wrong; by all useful metrics I’m aware of his performance was better (more Ks, fewer walks, lower FIP, lower xFIP, higher %GB) in 2013 than 2012. Is there an advanced pitching metric that shows 2012 as a better year than 2013? I’m not aware of it. What aspect of his performance, using stats that actually evaluate pitcher and not team performance (that is, not ERA and W-L) show him as better in 2012 than 2013?
Felix had an xFIP of 2.66 last year. The previous three years his xFIP was a full half run higher (3.14, 3.15, 3.20 in 2010, 2011, 2012 respectively). He had an established level of performance, and he exceeded it.
Now it’s possible–no one’s categorically denying it–that that improvement was him building on his skill set and becoming an even better pitcher. But it’s much more likely (that is, it happens much more often) that he simply had bit of luck and overachieved slightly, and his well-established level of performance from 2010-2012 is his most likely output going forward.
Looking over at the Steamer projections, I must say it’s a lot less grim looking than I expected. 32 WAR as currently constructed, when the average team in the league has 35-36 WAR. My lack of faith in the front office leads to believe the ~3.5 projected WAR gap between the Mariners and the league average team is at least as likely to grow as it is to shrink over the offseason, but that’s a lot closer to average than I would have guessed. A few further thoughts:
Curious where the projection for negative defense is for Seager is coming from.
That slashline for Zunino seems awfully optimistic. He didn’t seem to be improving at all on his horrible, not close to show-ready contact rate at all over the course of the year.
Holy Cow does Steamer love Brad Miller. Here’s hoping.
I get the need to regress Iwakuma’s LOB%, but why does steamer take it all the way below 70%? Shouldn’t it just regress to slightly better than league average, since he’s a better than average pitcher?
If Montero plays regularly and has a .333 wOBA, I’ll be both shocked and thrilled. Has my frustration and disappointment caused me to vastly underrate him as a prospect?
shannon = fish in a bucket.
why bother?
MLBTR says Chuck is retiring!
Wow
Half of HowChuck is retiring. Now there is an improvement!
eponymous, I was just going to post the same thing.
Btw, my source on this was the Mariners themselves, with Chuck’s statement, so I think we can assume it’s true.
Now this is an off-season improvement plan I can get excited by! All this talk of on-field moves is pointless because the people behind the moves are faulty towers. Its hard to imagine ever caring for the team until the Lincoln-Armstrong debacle is completely gone.
The biggest “downturn” by Felix this year was innings pitched. The funny thing about advanced metrics is that if you do nothing you get a zero not a negative. When Felix won the Cy Young, his cause was advanced by modern metrics, not W-L or any other pre-historic way of evaluation. Felix was the 2nd best starter on the M’s this year, that is a step or two down from first in the league. If Felix had pitched in September things could have been different. Felix’s health issues prevented a full season and lessened his value in reality if not metrically.