Hello, and welcome to the second (of four planned) parts of the 2019 Atlanta Braves player projection series. For a bunch of detail on what exactly we’re doing here, I suggest you look back at the position player post. Some pitcher-projection-specific FAQs are below, and then we’ll get into the the projections themselves.
Can you refresh me on the stats being analyzed?
For pitchers, this post focuses on only two stats, which to some extent are really only one stat: FIP- and fWAR. First, what is FIP-? FIP- is FIP, but similar to how wRC+ is wOBA — it’s scaled for park and league, and put on a basis where 100 is league average. So an 80 FIP- means a pitcher who has prevented runs (by virtue of his FIP) 20 runs better than league average; a 105 FIP- is a pitcher who has prevented runs five percent worse than league average. While not an exact relationship due to fWAR taking infield pops into account as a credit to the pitcher, fWAR is, for the most part, just FIP- (which is a rate) converted to a counting stat and made comparable to fWAR for non-pitchers.
One key mental calibration note: FIP- = 100 is the average run prevention rate, which includes both starters and relievers. As a result, an average starter is going to have an FIP- above 100, because relievers as a group tend to be better at preventing runs in their respective stints. However, 2018 featured the smallest gap between average starter and average reliever performance since 2011, and it remains to be seen whether that’s an anomaly or the sign of convergence in terms of effectiveness across roles.
Okay, but assuming I care about IWAG, what’s different in it this year?
Mike Soroka is a problem child. He’s not stealing mailboxes or setting neighborhood lawns on fire, though — he’s more of a problem child in the sense of being too precocious for his own good, but he’s a problem child nonetheless. The existing IWAG calculus for pitchers had such trouble finding comparisons for a very young pitcher just trashing the minor league competition that it sort of gave up and pledged its undying loyalty to him as a vassal state instead. As a result, I rebuilt the minor league pitching component from scratch to contextualize Soroka’s performance more, taking year-league run environments into account. Soroka still remains a problem child because no one as minor-league-successful as he is has gotten injured and held out for such a long time without injury plainly affecting his stats either immediately before or immediately after, but we’ll get to that later.
As noted in the position player post, the “Matt Wisler adjustment” is also in effect, and it matters for a handful of the players here.
Any other thoughts about pitching projections?
A lot, actually. This year, given the Braves’ fairly unique situation, there’s been a lot of pondering on my end regarding the projections, in addition to the usual calculating.
The first thought is about run environment. This isn’t really a big deal, but estimating FIP- and fWAR really depends on what the rest of the league is doing. IWAG makes some assumptions, they’re likely different from the Steamer and ZiPS assumptions. To that end, you may see the FIP/FIP-/fWAR figures not align or seem weird in relationship to either across multiple systems. Where this is the case, it’s likely run environment shenanigans (or I made a mistake somewhere in calculating things). For IWAG’s purposes, I present FIP just as a gauge for those who inherently find more meaning in an ERA-esque number than a number on a 100-scale; for the other projection systems, they only display FIP on the Fangraphs player pages, and I’ve converted it to FIP- myself for comparability’s sake. So, to be very clear: that’s not the FIP- projected by Steamer or ZiPS, that’s me putting their projected FIP on the FIP- scale (and any mistakes/issues are mine alone).
The second thought is about projection tendency. While I can’t be sure, I think Steamer uses a really different methodology for projecting pitchers relative to other systems, which takes into account many more (and different) things than, say, IWAG does. There have definitely been years, potentially as a result of this, where Steamer has done much better at projecting pitchers. The reason why this is relevant to the present exercise is because the Braves have a number of pitchers these days who are young enough, inexperienced enough, and idiosyncratic enough to cause specific problems for IWAG. Mostly, this is really just a way of me saying, “Hey, here are things to watch out for below that IWAG potentially struggles with,” priming you for some of the projections below.
- Walk rates. IWAG hates walk rates. I can’t directly say whether IWAG really hates high walk rates more than other systems, but prospects with high walk rates aren’t highly thought of. While there is indeed a general expectation of slight command improvements (again, in general), these are generally offset from an effectiveness standpoint by declines in “stuff.” As a result, IWAG red-flags hurlers with high walk rates... which happens to describe many of the current crop of pitching prospects. The reverse is true too — if you can minimize walks in this day and age of three true outcomes, IWAG will like you more, perhaps more than seems warranted.
- Stuff gaps (Matt Wisler syndrome part two?). IWAG does not presently use information about a pitcher’s stuff. (Why not? The need for that data for every pitcher would take it from a quick hobby to something more like a half-time job.) As a result, it assumes that pitcher results are earned by pitchers who have the stuff/ability to get batters out and prevent runs commensurate with those results. That sentence is another way of saying it doesn’t know Kolby Allard is going out there lobbing beach balls. So when you see Allard’s projection, this is why — Allard’s minor league track record numbers and his age-to-league comparables resemble, on average, a guy with much better stuff than he actually has.
- Roles. The concept of a “starting pitcher” is hopefully on its way out of baseball. With that said, though, usage as a starter versus a reliever has differential impacts on both pitcher inputs (how much they can let loose and burn through stamina) and pitcher outputs (whether they have to face guys a third time through the order, etc.). I can’t speak to Steamer and ZiPS, and whether they would project different run prevention numbers across different playing time/role/usage estimates. IWAG just tries to skip past all this and say, “based on the ways in which this pitcher has been used before and the success he’s had, here’s his estimate of future success in those same roles/usage patterns.” If you want to make mental adjustments to the numbers presented here, then, think about the general idea that guys used more as relievers have better run prevention but lower overall usage and value, while guys used more as starters have things go the opposite way.
Before I get to the charts, I’m going to paraphrase last year’s discussion on probability distributions below. That way, no one needs to go elsewhere to read it, and I have the peace of mind that I put the relevant disclaimers into the post.
Player projection is, in some ways, really an odds-making exercise, yet many projections that get published are just the estimates. I get why that is: saying “Player X has a 30% chance of being replacement level, a 10% chance of being a 1 WAR player, a 30% chance of being a 2 WAR player, and then a bunch of outcomes of 3 WAR or above with low probabilities” is a mouthful, and can be really hard to wrap one’s head around. The expected value of any distribution can be thought of as its point estimate, and point estimates are easy to digest: “this system says Player X is a 2.7 WAR player!” However, I think a lot of nuance is lost while clarity (perhaps spurious clarity?) is gained. So, I’m presenting all of the IWAG distribution curves for WAR below. I don’t present anything about FIP or FIP- here because in the end, we really only care about WAR and FIP/FIP- are just direct feeds into that number.
In addition to modeling player performance, IWAG also models health (likelihood of injury and missing time), as well as playing time related to performance. Therefore, on the WAR distribution, you will also see a golden line, which traces a player’s likely production based on playing time. Note that the relationship between these, as modeled, can be complex. IWAG figures that if a player is being awful, he may lose his rotation spot, and thus accumulate fewer IP. And, if a player is rocking it and gets even more playing time, he may regress further with that extra playing time. As a result, the tails are shortened on the WAR distributions as opposed to the WAR/200 distributions.
For each player, IWAG also provides a “confidence” measure, which I’ve summarized into four buckets (very low, low, moderate, and solid). The lower the confidence measure, the more splayed out the distribution tends to be, as IWAG has less certainty that the player will have an outcome in any particular range. However, the charts are drawn to avoid showing the distribution across very fringe probabilities. Therefore, the lower the confidence, the more you should keep in the back of your mind that there’s a greater relative chance that the outcome could occur somewhere to the left or the right of the pictured distribution. Where the confidence is solid, the chance of this happening is consequently fairly low.
Also, as a final note, which I cannot emphasize enough: the dots for the Steamer/ZiPS projections on all charts below are placed there simply for illustrative convenience. I am not making any representation whatsoever that I have any knowledge of the Steamer/ZiPS outcome distributions, or that they look anything like the IWAG distributions being presented. Their placement is solely so you can compare all projections for a player on each chart. To save space, I’m placing the Steamer-related dots in blue and the ZiPS-related dots in orange; the IWAG dots will stay green, except for the playing-time-adjusted WAR projections, which will stay gold.
Because projections are provided below for each player as a set of two images (point estimate table, IWAG WAR and WAR/600 distribution with point estimates), I’m only adding a tiny bit of commentary for each of the 13 players examined. You can mostly use the distributions presented to see where IWAG is coming from and how that compares to Steamer/ZiPS, but of course I would be overjoyed to discuss any of this in the comments or via another medium.
A small note here: while I can’t speak for the Steamer and ZiPS projections, and they’re inarguably the most important here, I find the probability distributions from IWAG particularly interesting in the context of the Braves’ array of high-uncertainty young arms. The point estimates are the point estimates and in some ways, they’re useful gauges of what you probably shouldn’t expect more than. But more than a number, it’s the spread, and the varying likelihoods in the distribution, that are perhaps more interesting. Your mileage may vary.
Mike Foltynewicz has a pretty generic distribution and projection, I think, for a young player who had a breakout season. There’s a chance he repeats, and a chance he regresses to some extent. As usual, IWAG is more optimistic than Steamer and ZiPS in these cases, while these systems seem to mostly focus on the central estimate of the “will regress” hump rather than something bimodal.
Another note here is that while the IWAG distribution does not factor in Foltynewicz’ recent health struggles, it still factors in some generic pitcher health risk, which is why the mode outcome for the big right-hander is a very mediocre mid-1.0s fWAR figure. But hey, on the plus side, all three systems feel at least reasonably confident that Foltynewicz will be average or better at run prevention, which represents a decent point of performance stability (if not health-based stability) amidst a sea of pitchers where that is definitely not the case. And, well, the more 3-win performances the Braves get, the much nicer 2019 season they’ll have. (Duh.)
Note: No, I don’t know why Foltynewicz’ Steamer200 WAR figure is lower than his Steamer figure for 192 IP. No, that doesn’t make sense. The Steamer estimates keep minorly changing as I attempt to write this, and while the conclusions don’t really change, it’s been a small nightmare for me to do this with every pitcher’s data at least reflecting the same Steamer update as I’ve had to put this post together over multiple days and things keep moving slightly.
Sean Newcomb is potentially weird in that he’s not that weird, if that makes sense. His projections are quotidian. You may not think that someone with his particular skillset and downsides is average, but that’s really all he’s been: a 101 career FIP- and a 106 career xFIP- are basically that, and so long as he suppresses homers even a little bit, he’ll be closer to an average starter than anything below that. While the walk rate is brutal (since his debut on June 10, 2017, Newcomb has 10 more walks than the next-highest total and a walk rate “lead” of over one percent on the same, who happens to be Lance Lynn in both cases), Newcomb has shown an ability to do the run prevention thing even as he issues free pass after free pass, and the projection systems have likely caught onto that to some extent.
So, his not-that-weird probability distribution is a fairly normal set of curves, centered around average run prevention. Perhaps one of the big downsides: his relative inefficiency limits his innings totals, such that he’s potentially more like to end up under 2 WAR than above it just because he won’t manage to pitch that deeply into games that often, and that lack of innings does eventually add up.
See same note as above re: Steamer200 vs. Steamer
The first thing that jumps out here: geez, these are some seriously similar point estimates. I can’t remember this ever happening across three systems for any other player. So of course, the outcome is going to be something like 1 WAR or 5 WAR that’s not even on this chart, right?
It’s kind of odd how similar Kevin Gausman has been to Sean Newcomb, despite really different skillsets, approaches, and everything else. The projection systems see Gausman as somewhat better, though not much, and with a bit more latitude value-wise due to a greater likely innings total. This gives the Braves two guys reasonably expected to be mid-rotation starters, which is never a bad thing. (It’s just not exactly a reason to celebrate, either.)
The thing about Gausman is that there always seem to be reasons for why he should improve, without said improvement. The split-change is great, but for whatever reason, it’s either underused or simply only good in relatively lower-use contexts. He has velocity on his fastball, but it doesn’t benefit him as much as other pitchers. He’s been pretty healthy over the last three years, but some early scuffles in Spring Training are perhaps putting that into doubt. None of these things really factor into the projections, nor do they really need to be: it doesn’t make sense to ding him for potentially-unfulfilled potential. So, Gausman is likely to be what he has been at this point: an average run preventer. And that’s okay.
If Julio Teheran had an epithet, like personages in Greek mythology (or certain potentially-overwrought videogames) do, it might at this point be “Teeth-Gnasher.” Not because he gnashes his own teeth, no, but because his presence evokes that particular physiological reaction, either metaphorically or physically. You can see why from the distribution above; sure, he has some non-zero chance of being above-average again. But he has a much greater chance of being something like a fourth/fifth starter. Hence, teeth-gnashing. The real gnashing also comes if you discount the possibility of good performance altogether, as ZiPS and Steamer might be doing, in which his case his expected performance drops into fifth starter territory.
Whenever Teheran, Teeth-Gnasher comes up, there’s always the discussion about him beating his FIP. Here’s a funny thing — Steamer literally projects the same ERA and FIP for him. That’s not because Steamer does this for everyone (they’re pretty close, usually, though): Sean Newcomb, for example, is projected by Steamer to have a half-win per 200 IP gap between his fWAR and his RA9-WAR. For Teheran, though, Steamer’s yielding a big ol’ nope. ZiPS, while the most down on Teheran’s run prevention skillset, does expect a 4.54 ERA against a 4.98 FIP. Hey, perhaps some credit to FIP-beating there. But, the reality is that a 4.54 ERA would have placed Teheran as having the 46th-highest ERA among starters with 70 or more IP last year, i.e., fourth starter territory, and that’s before you consider that that number comes from the National League and not a particularly offensively-geared park.
One way to describe Teheran’s status for 2019 could be, “How long will the Braves let him play out his string this year?” But another, perhaps more apt, is, “How long will the Braves’ defense continue to make Teheran look like an acceptable run preventer?” These questions are pretty important ones, because Teheran is pretty durable — if he’s going to lose a chance to rack up innings, performance seems like a much more likely impetus for it than injury. The Braves will likely need to be proactive to upgrade his spot, as it doesn’t seem like he’s going to find himself being Wally Pipped out of a starting role.
And with that, we wrap up the four kinda-sorta known quantities. Ahead lies probability roulette.
I’m glad Touki Toussaint is first in the probability roulette section. His chart above is a great example of why. You might be looking over the above and going, “wtf? IWAG projects Toussaint as worse than Teheran? Burn that mother down!” To, which I would respond — yes, if you only look at the central estimate. Instead, I implore you to look at the distribution. Toussaint’s mode outcome above? Over 3 wins per 200 IP. Why is his point estimate so much lower? Because that left tail is massive. Gargantuan. The reasons why aren’t hard to grasp. The leash might be short. The more opportunities Toussaint gets, the greater chance of him regressing to a mean. Not necessarily the mean, just a mean, like his personal mean. But the fewer he gets? That’s where things can get weird. If Toussaint pitches well, he will get to pitch more, and the more he pitches, the less likely it is he puts up crazy stats. If he pitches poorly, he may have the rug yanked out from under him. If he’s awful, he may only get a start or three, and those awful rate stats won’t go anywhere.
So, across a multitude of universes, the average result for Toussaint on a rate basis, from IWAG, is not that high. That doesn’t mean he isn’t capable of a great run, just that each such run is counterbalanced by some stuff not going awesomely. And why is that tail so enormous? Well, as mentioned, IWAG is really worried about his walks. More walks = worse outings = fewer opportunities to get that walk rate down. On the other hand, ZiPS isn’t that worried, though mostly due to assuming he suppresses homers as opposed to yielding fewer free passes. Steamer is maybe a little worried, though models a lower walk rate (albeit a Newcomb-ian one anyway).
Basically, there’s a lot of room for all sorts of things to happen. The challenge with Toussaint, as well as with every other pitcher below, is for the Braves to identify when throwing potentially-good innings after bad ones makes sense, and when it doesn’t. If they can solve that riddle again and again, this puzzle piece rotation might become a strength. If they can’t, they might be bleeding wins on the field and in the WAR column over and over.
Ah, Mike Soroka. I could talk about his projections for what feels like a decade. But I won’t, in part because his injury status is so ridiculously uncertain that all of this might be moot. The reality is, though, that Soroka’s 2019 projections broke the prior version of IWAG for pitchers, and it was rebuilt. Thanks, Mike. Even being rebuilt, the probability distribution above is super-dumb for multiple reasons:
- Soroka is generally expected to settle in as an average starter, but there’s a small likelihood he becomes a facemelting Cy Young contender and that drags his projection way up. This essentially never happens for anyone — it’s part of Soroka breaking the system because his data are so idiosyncratic.
- On top of that, Soroka did a really weird thing in 2018 both in terms of his overall major league profile, and the fact that he pitched very few innings, but the major league innings he did pitch (at age 20... his first exposure to the majors, after just five outings at Triple-A...) put him on a 4.6 WAR/200 basis. In history, I’m not sure there’s ever been a player around Soroka’s age with so little track record in Triple-A and the majors who did what he did in the majors but then got very little playing time. As a result, the error bounds are ginormous as the model searches for something to draw on... and there really isn’t much. So his non-rate-basis WAR estimate is at the end of the distribution because the distribution charts lop off things with particularly low likelihoods, and for Soroka, everything is a low likelihood and that line has to be drawn somewhere, so there we are. Anyway, it’s a mess.
But Soroka, Soroka isn’t a mess. He’s projected to be really good. His projections are on par with, if not better than, the ones for the 2018 NL Rookie of the Year last year (and in some cases, this year too). (Well, at least the point estimates.) The question is the extent to which he’ll be healthy enough to do anything, with a secondary query regarding the extent to which, if at all, his injuries and layoffs have diminished his ability to wreck opposing batters.
If Soroka is healthy and as effective as expected, that’s a huge boon to the Braves in 2019. It gives them another average-to-above hurler to go with Foltynewicz and reduce the risk of below-replacement or below-average performance for however many innings he’s able to contribute. If he isn’t (and there’s a relatively sizable part of the yellow line that suggests he might provide 1 WAR or less), probability roulette will continue with greater earnestness. What a time to be alive.
Luiz Gohara is kind of like the Ghost of Christmas Future for Mike Soroka’s projections. Last year, believe it or not, the big left-hander had the best projections among Atlanta’s prospective starters. They ranged from 3.0 to 4.1 fWAR/200, and no, I’m not joking. It was a real thing, and IWAG was the low man, not the high man. Fast forward 12 months and a whole buncha stuff has happened to Gohara, essentially none of it even remotely positive. Will this be Mike Soroka’s fate going into 2020? We’ll see.
ZiPS and Steamer aren’t really giving up the ghost to any extent, still projecting Gohara to prevent runs at an average rate if given the opportunity. It’s that latter part that will be tricky, as between injury, diminished playing time last year, and personal factors, he may simply not get the same chances afforded to others at this point. As far as the probability distribution above goes, the low-ish IWAG estimate is again driven by a substantial chance that Gohara performs akin to his replacement-level-ish standard in 2018, as opposed to his heady 2017 performance that got him those gaudy projections last year. IWAG still sees a path for him to make good on his potential; the difference is that now it’s heavily offset by the chance of being generally unproductive. Put those together and the central estimate is a slightly below average performance, a disappointing far cry from March 2018, when Gohara seemed like the best bet to finally deliver unto the rebuild an above-average major league season from a pitching prospect.
I don’t know that I have much to say about Max Fried. He’s pretty similar to other pitchers in this post. Maybe he’ll be good. Maybe he won’t be. That walk rate is pretty scary. People don’t seem to talk about Fried’s walk rate for whatever reason, but it’s really quite bad: he’s never had a BB/9 below 4.00 or a BB% below something that rounds up to 10 percent or higher at any stop where he’s had more than 20 innings. His best walk rate was probably his most recent 10.3% at Gwinnett in 2018; that would be a bottom-20 walk rate among MLB starters with 70 or more IP last year (two of whom are Sean Newcomb and Julio Teheran, so that’s “awesome”). All that and he also needs to stay healthy.
So, it’s the same story: he could be pretty good. Or he could be terrible. IWAG sees a lot of variability for the high-walk rate guys who haven’t shown they can succeed despite it so far. ZiPS is also skeptical, though perhaps less so; it forecasts over 5.0 BB/9 for him in 2019 as a point estimate, which makes my brain hurt. Steamer is hard to compare in this case, as it treats Fried as a reliever. A 91 FIP- from a starter would be great and better than the 2.4 fWAR/200 listed there; the same from a reliever maxes out at under a win, so it’s alright, but not dominant. In the future, people might look back at stuff like this and go, “Man, remember when people had to discuss projections for players in the context of someone being a ‘starter’ or a ‘reliever?’ How archaic!” Unfortunately, 2019 probably won’t be the start of that era.
If Mike Soroka broke IWAG in one way that necessitated a fix that way eminently doable, Kolby Allard highlights a potential break in the system that I’m unlikely to fix any time soon. Stat-wise, Allard looks... fine? He was a teenager doing pretty well at Double-A, and while his xFIP was elevated at Gwinnett, the overall line looked fine if you buy into a bit of homer suppression on Allard’s part. But, then you see him in the majors and get some detailed pitch data, and oh boy... it’s like he’s throwing grapefruits up there.
So, which do you believe? Steamer integrates some pitch data — do you think even its 0.8/200 mark is too high? ZiPS thinks Allard can prevent runs at a rate less than 10 percent above average — how confident are you in that? IWAG’s probability distribution seems pretty reasonable for a backend starter rising up from Triple-A; Allard’s minor league stats are consistent with those posted by that pitchers of that archetype. But, is that the right group from which to draw inferences about Allard’s major league future? I’m skeptical. If I had to pick one projection this year that I felt was more wrong than the others, I’d say it was this one. And, unlike the addition of the Matt Wisler adjustment, taking care of this particular edge case would require loading tons of new data into the model. Or, maybe, Allard will outmaneuver his ostensible pitch quality and make good on this range of projections. We’ll see... but again, I’m skeptical.
You might be thinking, “What’s Ian Anderson doing here?” He’s got all of four starts at Double-A so far, and isn’t really being talked about as a candidate for starts in Atlanta, what with all of the other arms ahead of him developmentally. Yet, the Fangraphs Depth Charts have him getting a start or two for the team, and it’s not like the Braves have mandated full stops at each minor league level before handing out promotions. Moreover, though, I think this projection is instructive and kind of interesting.
The rate-based probability distribution is in large part a display of how guys with similar track records in the minors have fared (with some other stuff layered on, of course). That suggests Anderson’s track record is pretty good so far, with a mode outcome of a mid-rotation starter and a spread that’s limited in ceiling but with lower-percentile outcomes that are still useful to a major league team. It’s pretty cool that as a central estimate, all three projection systems think that Anderson could at least resemble a backend starter right now. Just remember that the 16th time the Braves send a 115 FIP- Julio Teheran out to the hill this season.
Bryse Wilson is a bit of a counterpoint to some of the other, more volatile projections. This makes some sense, as he’s much more of a moderate walk rate, less extreme strikeout rate guy than the Touki Toussaints of the world. He’s suppressed homers in the minors, but it’s unclear how much that’ll translate to the majors — beyond that, he’s basically got the same pressures as many other hurlers: the more guys he can strike out without spiking his walk rate, the closer to solidly average major league performance he’ll get. An FIP- central estimate around 103 to 109 is really impressive for a guy with only around 100 professional innings above A-ball.
As with Fried, I struggle with translating Steamer’s projection into starter terms — Steamer thinks he can be a decent reliever/swingman, apparently, though not necessarily a stellar one.
Every year, there are some guys who are surprisingly-highly rated by some projection systems. This is really just an artifact of a pretty impersonal data processing task — there are always edge cases. Is Wes Parsons such an edge case? Probably. It might be Matt Wisler Syndrome, or it might be something akin to what’s happening with Kolby Allard’s projections this year, or a mix of both.
Steamer doesn’t really care about Parsons — a league-average FIP as a reliever isn’t really all that interesting, though it does get somewhat more intriguing if you think Parsons serves as more of a true swingman while preventing runs at a league-average rate. ZiPS has him as a generic backend starter, which may not be much to write home about, but is still really exciting value for an undrafted free agent who gets like, 35 FV grades. But, back to the surprisingly highly-rated thing: IWAG loves Wes Parsons. In many ways, it always has. For essentially his entire minor league career, Parsons has done the things necessary to prevent runs without really backsliding in effectiveness while moving up the ladder. He doesn’t strike out many but the walk rate isn’t too high either, and even with no homer suppression, that’s a pretty okay pitcher. The uncertainty of him just being flat-out worse does pull him back a bit from being average, but IWAG doesn’t see huge downside in him. I’m not sure if Parsons will get any kind of extended chance to strut his stuff in the majors this year, but if he does, I’m going to find it really intriguing.
Patrick Weigel is more than a little bit of a darkhorse for 2019 Atlanta Braves starts — he may be worked back into the mix as a reliever given that he’s been recovering from June 2017 Tommy John Surgery, though he did manage some live in-game rehab work late last season. With that said, though, the projection systems are not too excited by him. The reality is that Weigel hasn’t really had relatively unequivocal success since A-ball in 2016, as his Double-A starting experience was mixed (and consisted of just ten starts) and he was terrible in Triple-A. That makes him more replacement level than not (or history’s greatest monster, according to ZiPS) with IWAG not seeing much of a useful ceiling distribution-wise, either. Even Steamer treating him as a reliever isn’t that kind, since a reliever with a league-average FIP is actually behind the curve as far as relievers go.
But wait, what about Kyle Wright?
Yeah, yeah, seems like a big omission, right? Well, it’ll be interesting to see exactly what happens with Wright. The Fangraphs Depth Charts currently have Wright only getting nine innings of work for the major league team, with guys like Fried, Allard, and Gohara ahead of him. That probably seems less likely than the converse at this point, but the reality is that Wright is a nightmare to project because his professional track record is minuscule. Last year, to qualify for the ERA title, a pitcher had to have thrown as many innings as Wright has thrown in professional baseball. That’s not really a great basis upon which to project anything. So, attempts to get any kind of useful projection for Wright are, on my end, met with my computer sneering at me — and hence, no distribution or the like.
Steamer sees Wright with a 4.32 FIP, which by my calculations is a 107 FIP-. That’s basically the same average-y territory described as where Steamer and ZiPS have Sean Newcomb pegged for this coming year. ZiPS is pretty similar at 4.38, which is incidentally basically the same mark it has for Allard. And while IWAG won’t really generate a distribution, the best guess it has for a center of a theoretical distribution is a 105 FIP-. Basically, in the absence of a lot of information, you get: Kyle Wright, Average (?) Starter. That seems like a good, safe bet given the sparse track record to go off of.
Putting it together - starting pitchers
Below are a couple of tables that sum up the above to get a total starting rotation WAR number.
Well, those are certainly numbers, though the point estimates aren’t that great. The median rotation earned right around 10 WAR last year, which makes perfect sense given that there are usually five rotation spots and an average player is a 2 WAR player. The Braves come in below that by Steamer and ZiPS, and above it by IWAG, mostly thanks to more aggressive projections on Foltynewicz and Teheran for the latter (though offset by a worse one for Toussaint). No matter which central estimate you look at, though, these are all middling rotation totals. Last year, it would have ranked 15th to 18th. In 2017, 15th to 17th. In 2016, 17th or 18th. You get the idea. The Braves’ actual rotation ranks for the last three years, working backwards, have been 12th, 17th, and 29th. In some ways, it’ll be kind of weird if the pitching prospect smorgasbord yields a decrease in rotation value from last year, but that’s where we are right now, on a central estimate projections basis.
For the position players, I also presented a table of playing time as IWAG saw it, rather than the Fangraphs Depth Charts. This is much harder to do for starters. Instead, I’ve lamed out of the exercise by instead just using the four established veterans, plus Mike Soroka as a proxy for “good outcomes” and Touki Toussaint as a proxy for “highly variable but not that great on a central estimate basis” outcomes. Here’s what that table looks like.
Nothing much really changes. While the fate of the rotation may really be based on health and high-level production from any young arm that steps up, that’s hard to see when only talking about central estimates where playing time is locked in, and three of the four “set” rotation spots go to average-and-Teheran players. So, to wrap up, let’s look at this rotation probabilistically, like we did for the position players.
This chart is really different from the position player one, and not really in a good way. The ceiling shown here of around 17 wins is good but not great — four teams managed more than that from their rotation last year alone. Scarier though, is that unlike the really high floor for the position player group, there’s a ton of downside at the lower end, if everybody decides to break or be ineffective in unison. Probably the real kicker for me, though, is that it’s relatively unlikely that the Braves transform their current group of starting pitchers (at least this year) into something like a top 10 unit (14-15-plus WAR) — there’s just too strong of a tendency towards the center with something like 60 percent of the total available innings going to guys fairly unlikely to be above average enough to allow the full rotation to hit this mark, regardless of what their teammates accomplish.
On to the bullpen, next time!