18 November 2013

Evaluating What Shelby Miller Would Cost the Orioles

One of the most difficult aspects of running a big league club is figuring out exactly what you have as well as what others have.  The ability to do so opens up a world where you have a good idea about the value of the pieces you have as well as what those pieces add up to in comparison to the other franchises.  Much of this is qualitative and rather difficult to provide definitive explanations of specifically what a prospect is worth or what a "proven" veteran is worth.  The problem is that predicting the future is not possible.  Projection models like ZiPS, MARCEL, and others do a great job in figuring out what will happen, but that comparative accuracy still only means the models tend to be accurate within 8 games for only about three quarters of the teams.  That is a lot of wiggle room.

So, yes, we can write about how things are so hard to know, but the other perspective is that, yes, we do know things.  In the extreme, yes, we do know that Cal Ripken Jr. is not able to go out next season, suit up for the Orioles, and break Barry Bonds' home run record.  We actually know that.  The chance of it happening is so infinitesimal that it is facetious to say otherwise.  Perhaps more apt, within the realm of possibility, we know Brian Matusz will not provide front end of the rotation performance.  The projection models, even with their inaccuracies, help provide some basis for items like this often more so than simple judgement.  However, this post is not exactly about those models.  It is about another model.

One of the more brilliant ways to define prospect value was determined by Victor Wang back in the day.  It suggests that by evaluating where prospects were ranked or rated in the past by Baseball America and John Sickels results in certain probabilities for success.  In other words, as these two entities appreciate the value of a prospect's future, it is actually correlated to an increase in performance.  Yes, two thirds of the prospects on Baseball America's top 100 list will bust out and never be a meaningful contributor to a big league club, but you can put a performance value on where a player stands.  If you can measure performance, you can put that into runs.  If you can put that into runs (or wins), then you can convert that into money as suggested by the free agent market.  This does not exactly mean that a specific player is worth 15 MM, but that historical performance of players like him suggest that his value is roughly 15 MM.  I hope all of that makes sense because this post largely depends on your understanding of my explanation here.

One of the topics getting the most attention for us over the past week has been the allegedly discussions between the Orioles and Cardinals that touched on J.J. Hardy and Shelby Miller as well as our response here to those rumors.  Perhaps four out of five responses regarded the following suggestions as absurd:
This was not exactly a blind stab into the dark.  It was the result of a quick conversation with a talent evaluator.  Anyway, Orioles fandom was not happy and here is an example of that:
This made me think that this might be a good moment to provide context to prospect and player value.

What is Shelby Miller worth?

Coming into the 2013 season, Miller was rated as the third best pitching prospect in baseball as considered by Baseball America behind the Orioles' Dylan Bundy and the Marlins' Jose Fernandez.  Bundy and Fernandez had leap frogged Shelby due to the formers' solid 2012 minor league campaigns and Miller's homer prone venture in the Pacific Coast League.  Even with his issues in AAA, Miller advanced from 8th best prospect overall in 2012 to the 6th best prospect overall.  In other words, it is what Bundy and Fernandez did as opposed to what Miller did not.  Miller still profiled as a top of the rotation arm.

At the age of 22, Miller entered into the Cardinals rotation and proceeded to show why everyone had so much faith in him.  He excelled the National League with the exception of a few starts in August.  It was not the kind of domination that Jose Fernandez enjoyed in that Fernandez right out of the gates has shown himself to be a top of the rotation arm.  Fernandez was so impressive that he finished third in Cy Young voting even though he threw only 172.2 innings.  Miller also performed better than Dylan Bundy who began Spring Training injured and eventually had reconstructive surgery.  Some may wish to waive off Tommy John surgery, but it certainly is something that can result in a weakened arm and at best simply has resulted in the loss of a year in development.  With Miller, he had a full year in the MLB and showed that he could performed as a middle of the rotation arm on a first division team as such a young age.

Year Age W L ERA G GS SHO IP ERA+ H/9 HR/9 BB/9 SO/9 bWAR fWAR
2012 21 1 0 1.32 6 1 0 13.2 296 5.9 0.0 2.6 10.5 0.5 0.5
2013 22 15 9 3.06 31 31 1 173.1 119 7.9 1.0 3.0 8.8 3.4 2.1
Provided by Baseball-Reference.com: View Original Table
Generated 11/17/2013.

Of course, when you trade for someone you do not trade for who they were, but who they will become.  His age, based on historical performance of others like him, would suggest that he will perform his best in his late 20s.  That is, Miller would be expected to improve upon these numbers and perform at a higher level.  Ignore the numbers and go with the qualitative scouting, you will find that there is a popular opinion among baseball folks that Miller will improve.  However, those perspective might not move you and we will take a slightly more conservative perspective in our evaluation of Miller.

In this post, we will assume that Miller will do worse than either fWAR or bWAR had him this past season.  We will call him a 2 WAR pitcher per year.  With that in mind, we will project him as a 2 WAR pitcher each year from 2014 through his last controlled year (2018).  We will consider his value as surplus value compared to what a pitcher of his performance would likely earn on the free agent market.  This past season saw the cost per win at about 5 MM.  This off season has been suggested to increase that mark to 6.5 or 7 MM due to new media money coming into play locally and nationally.  However, we'll be conservative and suggest an increase of 10% per year over the course of Miller's controlled years.  As we will be keeping this value constant for all parties involved in this potential deal the amount of money we assign per win is somewhat arbitrary and simply scaled up only for the numbers to make more sense.

Here is what we get:

WAR $ / Win Market Control Surplus
2014 2 5.5 11 0.5 10.5
2015 2 6.1 12.1 2.4 9.7
2016 2 6.7 13.3 5.3 8.0
2017 2 7.3 14.6 8.9 5.9
2018 2 8.1 16.1 12.9 3.2

67.1 30 37.3
If we expect Miller to be a 2 WAR pitcher, an average starter, then we expect him to provide 37.3 MM in surplus value to the team.  That means that over the course of the next five years, the Orioles would be able to devout that money to help acquire or maintain players.  That also means that the Cardinals should be interested in trying to acquire pieces back from the Orioles that come close to that surplus value as well as positionally making sense for the organization.

How much is J.J. Hardy worth?

Hardy has hardware, winning the gold glove twice.  His offense leans heavy toward home runs and little else.  That is not a bad thing.  There certainly is a cult of on base percentage, but OBP is not all.  It certainly is very important, but that does not negate the importance of things like home runs which Hardy has.  Still, it is not a special bat.  It is an average bat paired with an above average glove, which gives you a player who is a fringe first division shortstop according to fWAR (10th overall at SS in 2012, 9th in 2013).  That is a valuable player.

What makes Hardy valuable in another way is that McPhail signed him to a three year extension that is now in its final season with his salary pegged at 7 MM.  That is a great underpay based on his performance over the past several seasons.  How did the Orioles sign him to such a cheap deal back then?  He had a couple rough seasons at the plate and had gotten significantly injured in the three seasons prior to signing the deal.

We will use an optimistic projection and declare that we expect Hardy to play 150 games and earn 3.5 WAR.  At 5.5 MM per win, his value would be 19.3 MM.  With a 7 MM salary, his surplus value comes to 12.3 MM.  A trade of Miller straight up for Hardy would be 25 MM in the Orioles' favor and would constitute a highly unfair trade.  If I were the Cardinals general manager, I would look to Jhonny Peralta, Brendan Ryan, or someone else freely available.  I may want a shortstop badly, but I do not wish to fritter away valuable pieces for a shortstop in his 30s who is probably going to look for a big payday after the 2014 season and is not an elite shortstop.  Luckily, the Orioles have some pieces in the minors that could be of some use.

What is Eduardo Rodriguez worth?

Rodriguez is a prospect.  He is a good prospect, but he is not exceptional.  He was one of the youngest players in AA when he was promoted mid-season and struggled.  Not greatly, but he performed at a level that was below average using ERA.  Mind you at this age, Miller had already established himself as one of the best pitching prospects in baseball.  Rodriguez is likely to see his age 20 season in Bowie with perhaps an advancement to Norfolk if he performs well.  It is not a given and there are some split opinions on him (some envision him as a setup man).  Rodriguez has never been found on a top 100 list.   Roughly two thirds of all prospects who make a top 100 list never yield a significant contribution to their team.  If you wish, you can call that a bust rate.  It really illustrates how difficult it is to make that climb from the minors to the majors.

Remember Victor Wang, well, using his methodology and updating for changes in the price per win we can attribute a surplus value amount on Rodriguez.  This value considers performance probability of pitchers similar to him in terms of grading.  If we use his 2013 ranking as a B level prospect by John Sickels, but not a top 100 prospect, then his value is 8.9 MM.  However, if you wish to predict that he will be a 75-100 prospect on Baseball America's list then his value would be 12.0 MM.  Again, these numbers may seem arbitrary, but they are based on the historical population of players ranked as such and then taking that performance and converting it to a dollar sign.

Our low end deal of Miller for Hardy and a low projection Rodriguez would result in the Orioles seeing surplus value of 16.1 MM.  The high end deal would have a surplus of 13.0 MM for the Orioles.  In both scenarios, it seems that the Orioles would need to add to the package in order to acquire Miller.

What is Jonathan Schoop worth?

Before the 2012 season, Jonathan Schoop enjoyed a Baseball America ranking of 82nd.  As one of the youngest players in Bowie, he struggled to make contact and did not make adjustments.  That lead to a souring on him that knocked him out of the top 100 list before this past season.  2013 saw him struggling to make contact, struggling with his mechanics, and suffering from injuries.  Although he had a sip of coffee in the Majors, he is not expected to play any meaningful role with the team.  This was further emphasized as he struggled at the plate through much of his time in the Arizona Fall League.  He is still quite young and perhaps could regain that shine.

If we use the same methodology as we used for Rodriguez, we can project him as he was before the 2013 season (a B level prospect) or optimisitically as 75th to 100th Baseball America prospect.  On the low end, his surplus value is worth 6.7 MM.  On the high end, his value is 15.3 MM.

The low estimate for the deal with Schoop included would be a 9.4 MM surplus for the Orioles.  The high estimate would be a 2.3 MM loss for the Orioles.


I hope this article brings home just how long the road is for Schoop and Rodriguez to becoming useful players for the Orioles.  It may seem exciting and convenient to pencil in Schoop as the second baseman next year or Rodriguez as in the rotation in 2015, but this is an incredibly optimistic envisioning of success.  Miller's value should also be a silver lining for Kevin Gausman if he manages to successfully perform as a starter or Dylan Bundy if he can get back to where he was before his injury.  Neither are givens, but they do provide hope for the Orioles being successful and being free to dedicate money elsewhere.

In my opinion, I think a Miller for Hardy/Rodriguez/Schoop deal is incredibly fair.  Though, if the Cardinals have something against Rodriguez or Schoop, I could see the deal going nowhere.  This may be a difficult thing to hear, but it is possible.

Some may scoff at this valuation and note that no pitcher in recent history has been dealt for so much.  This is true, but no pitcher has been dealt with only one year of service time and with such a high degree of favor as Shelby Miller.  Miller was seen at a top of the rotation arm and did nothing to suggest otherwise during his age 22 MLB season.

Some may contend that the Orioles are robbing SS or 3B to pay SP.  That perhaps counting on Manny after a significant knee injury to play SS reasonably well is foolhardy.  I would counter on the latter point first in that as knee injuries go that this is not an incredibly troubling one and that it would be just as troublesome for Manny as a 3B.  Second, the team should not be constructed to win April because by planning for that month it probably means you only win that month.  Finally, Miller means more to the long term health of this organization than anyone else in this deal in terms of performance and likelihood to continue to perform.  Hardy is not a great extension candidate and might well be a fringe qualifying offer candidate given his full career history.

I would also contend that the increase in performance with Miller in the rotation is probably equivalent to the difference between Hardy and whoever the team could easily field at shortstop or third base.  The Orioles should consider players like Brendan Ryan, Jamey Carroll, Juan Uribe, and Jhonny Peralta to help Machado out on that side.

Anyway, to get good quality, you tend to have to give up good quality in baseball.  Cheap deals are not always plentiful.  It can make these things hard to swallow especially when we have invested so much of our time growing up with these prospects developing in the Orioles' system.


Lou Proctor said...

I would make the Hardy / Schoop / Rodriguez for Miller deal in a heartbeat and your analysis of the matter is excellent.

Frankly, I still think even this would be a steal for the Orioles, because starting pitching has been such an issue for them. Any time you can trade non-top 50 prospects for young cost-controlled players that have demonstrated MLB effectiveness, I think you should seriously consider it.

I'd miss Hardy and wish him well in red, but come 2015, a rotation headed by a 24 year old Miller, 24 year old Gausman, and 22 year old Bundy would have me giggling all the way to season tickets.

Matt P said...

There are a few important things to realize when we talk about Shelby Miller.

The first is that while he did put up 2 WAR as a rookie, he did it in 173 innings. Most rookies don't throw 173 innings their first year in the majors and therefore what appears to be an impressive WAR is less so. Jordan Zimmermann for example put up 1.8 WAR as a rookie but did it in only 91 innings.

The second is that he did it in 31starts which means he averaged less than 6 innings a start. He averaged 5 innings a start in the minors in 2012. Either he'll need to throw more innings per start or his value will be limited.

He'll need to have exceptional performance to put up average value.

I'd be cautious of assuming he'd be worth 2 WAR for each of the next five years. I'd probably put him at 1.5 WAR for each of the next five years.

Jon Shepherd said...

Matt...that is a pretty weak point. A 22 putting up a mid 2 or 3 WAR is impressive.

Lou Proctor said...

Regarding innings, Miller's season high pitch count was 113 and the median was 100. Given that he was 22, had only thrown 150 IP in 2012, and was subject to being in a NL lineup, I suspect the Cardinals weren't trying to push him to his full IP/GS, in an effort to ensure he didn't throw too many innings over the course of the season and would be available in September. He's still building up the arm strength to be a 200 IP starter, which at 6'3" 215 lbs, he has the ideal build for.

Also, Baseball-Reference clocks his WAR at 3.4 and his ERA+ at 119. Frankly, I think that's everything you could ever want from a 22 year old pitcher and then some. Imagine if Gausman had that kind of production for the Orioles in 2013.

Anonymous said...

Baltimore cannot afford all the players you can sign at a WAR/$ profit. It is that simple. Hardy is an example of a player who can extract a very high percentage of his value as salary. The Orioles can only have so many of those players and a $90-$100 million budget. If we re-sign Hardy at $15 million a year, and five more players like him, we can't field a ballclub. We would have six fine players and a bunch of holes.

The Orioles simply do not have enough cost-controlled players for their business model. If you can trade a non-cost controlled player for a cost-controlled player, then you get closer to a sustainable cost-structure for Baltimore. What I don't understand is why St. Louis would do it.

Overall the trade would be getting pitching and selling hitting. Why wouldn't Baltimore do that?

Matt P said...

First, my post got swallowed. Now Fangraphs is breaking on me. So, quick version without the link I want to show.

Bottom line, remember that he threw 170 innings last year. Most 22 year old rookies don't do that. When most rookies throw 90 innings and Miller throws 170 innings then of course his WAR will be higher.

Millers performance was impressive because he was good enough to throw 170 innings. That's a positive. Now let's compare him to rookies who did that. Suppose we look at all rookie starters who threw at least 140 innings that are 23 or younger.

His performance is good enough for 17th out of 22. Roughly the same goes for his innings per start.

Jon Shepherd said...

17 of 22. And how many 22 yo who become star pitchers are not in the majors. I think there is a population issue here.

Jon Shepherd said...

Another thing to look at would be how that taking related to future performance.

Bermanator said...

I like how the projections for Miller assume he gives his team five injury-free seasons, coming off a 173-inning campaign as a 22-year-old. Because, you know, young pitchers are safe bets to stay healthy when worked that hard. That's even better than the "Oh, Machado will be fine -- there's no issues with his knee injury so he'll be ready to transition back to shortstop right away," as though Jon has seen the medicals.

Baltimore's never gonna make this deal, so I'm not so worried about it. But the only thing worse than falling in love with your own young talent is falling in love with someone else's.

Stay the course. Let Hardy play out the year. Give him the qualifying offer, he'll turn it down, and you get the first-round pick. Filling a hole in the rotation only to create one at shortstop, when the non-Machado options at short on the market are so underwhelming, doesn't make much sense given this current team's window.

Jon Shepherd said...

So, yeah, you could delay creating a hole a year.

Meds typically are fine in injuries like these. Read our site more often. Stuart wrote on this.

A more realistic assessment would have him much higher than 2 WAR and earlier in hos career path. Likely far more valuable in relation to cost control and minimal cost liability.

Bermanator said...

"A more realistic assessment would have him much higher than 2 WAR"

No, a more realistic assessment would have him getting injured at some point along the way. There's a reason "there's no such thing as a pitching prospect" is such a cliche.

The bottom line is that Hardy-Schoop-Rodriguez for Miller is one of those deals where both teams probably say now. I'm sure the O's want Miller, but not at that price. We know the Cardinals are interested in Hardy, but they're probably looking to give up Lynn or someone else instead. Guessing it's a kick the tires and move on conversation that happens so often between GMs.

Lou Proctor said...

"No, a more realistic assessment would have him getting injured at some point along the way."

Why? He's a large, strong pitcher who has, to my knowledge, never been on the DL or suffered an injury in his years in pro ball.

If the assumption is that every pitcher will get injured and is therefore not worth acquiring, then I suppose the Orioles should trade Tillman, Chen, Gausman, Bundy, etc. and go with an extended bench.

Besides, if Miller is an injury risk because he's a (until now completely healthy) pitcher, the obviously the same must be true about Rodriguez. We already know Schoop is an injury risk, with his back issues. Hardy was, until recently, considered injury-prone.

I don't see any rational reason to assume Miller is more likely to get injured than improve on his 2.1 fWAR / 3.4 bWAR as he ages.

Greg Pappas said...

A very solid analytical post, as always Jon.

I'd like to know what percentage you'd attribute to Eduardo Rodriguez becoming a better than average ML starter? 20% chance? I ask, because it helps me to make sense of such hypotheticals.

I, for one, agree with you assessment, and would pull the trigger. Alas, the O's will likely not.

Jon Shepherd said...

TINSTAAP is sadly a useful term, but one that is grossly overused to the point where people don't really consider what it means. Injury rates are a significant part of why pitching prospects have troubles with matching the scouting assessment, but not all of it. A large part, perhaps more significant, is the difficulty in accurately assessing how well a package of pitches performs from one level to the next amd how those pitches change over time. Injury issues simply are not as big a deal. I once thought they were but Daniel Moroz and I noticed that performance issues were basically equivalent between position players and pitchers once they hit the Majors. So that suggests a scouting issue of how pitches play level by level.

Matt P said...

For sure it's a small population. After all, Kevin Gausman was 22 last year and didn't pitch 170 innings because he wasn't good enough. It's impressive to be good enough to get to throw 170innings.

As for those 22 pitchers (140 innings or more in their rookie year defined by the Fangraphs rookie filter from 2005 to 2013), 3 of their rookie years were in 2013.

For the 19 remaining pitchers, their mean Rookie WAR was 2.53. So far, their average career WAR (not counting that rookie year) has been 1.94. 12 of the 19 regressed.

The six rookie pitchers that put up a WAR between 1.7 and 2.5 were Porcello, Niese, Britton, Scott Olsen, Josh Johnson, and Matt Moore. So far, they've had an average WAR of 1.7.

19 data points is far from conclusive. But it does seem to indicate that 22 year old rookie who pitched a lot of innings and had a good year are merely due for some regression rather than being studs.

Anonymous said...

A solid analysis.

However, speaking as a Cardinals' fan, if the Orioles only value Miller as "a 2 WAR pitcher" there is no reason for the Cardinals to even discuss trading him. One of the principle goals of any trade like this should be to find a trading partner who values your prospects more highly than you do and is willing to give you proportionally more in return. The Cardinals clearly value Miller more highly than "a 2 WAR pitcher" and it makes little sense to trade a player/prospect like Miller to a team that values him less than you do.

Further, with the current depth of the Cardinals minor league system, they are not in need of adding prospects like Schoop and Rodriguez. If the Cardinals were interested in trading for Hardy, they'd simply look to assemble a package of prospects that didn't include Miller for Hardy alone.

Anonymous said...

O's fans are right. Miller to big of a risk. Cardinals should trade L. Lynn for J.J. Lynn is not a rookie, no injury history, pitched more than 200 innings last year, and won 18 games in '12and 15 last year. I would gladly give you him over Miller. Miller would be our best young pitcher, out of a heap of 'em, if it wasn't for Wacha. Or you know what, we'll take Machado since you guys are worried about injuries.

Liam said...

@anonymous Would Lynn be worth much less than Miller? One less year of team control but similar peripherals (Lynn might actually have the edge) and Lynn's got two solid years under his belt. Not totally sure who I'd rather have, to be honest. Also, the point of a trade isn't just finding someone who values your guys more than you do, its finding someone who has surplus in an area you don't, kind of like international economic trade.

Back to the point of the article, great analysis, and I agree the deal is pretty darn fair. We'd have to fill some holes in the middle infield but Cano's a free agent, so there's one down, right?

Anonymous said...

@Liam - certainly when trading you need to find a trading partner who has the kind of resource that you need, in this case a SS for the Cardinals. However, it should be recognized that the Cardinals have other options than trading for Hardy. They could look for a trade with Texas, Arizona, or a few other teams, or they could simply go the FA route with Peralta, Drew, etc.

With the options they have, I don't believe the Cardinals are in a position where they have to make any type of "negative value" trade that they are not comfortable with - and I don't believe they would ever be comfortable packaging Miller in a trade for Hardy. If the Cardinals are going to move Miller it would have to be to a team that projects him as a 3.5+ WAR mid-to-front of the rotation starter very soon, and one that would give the Cardinals suitable value in return.

You just don't see very many 3-for-1 trades in baseball where the "1" isn't a veteran making a near full market value contract. If Miller is the best player in a deal that the Cardinals make, it will be because someone approached them and bowled them over with an offer for him.

Honestly for Hardy (signed for only one more year and not the most consistent in terms of offense over his career) I see the Cardinals offering maxing out at Joe Kelly + a "B" level prospect.

Anonymous said...

@liam - On paper Lynn may look as good or even better than Miller, but you have to look at more than numbers. Things such as composure, especially for a pitcher. Lynn can be lights out. But if someone makes an error or he gets into trouble, usually a crooked number is going up on the board. That's why I would give you Lynn over Miller. I've been waiting two years to see Miller in the bigs and he didn't disappoint. Now everyone expects us to trade him for little return. No way! I would love to have J.J. But I can wait a year and keep Miller/Lynn. Just about anybody could have a better year at the plate than the duo we had last year. Why wouldn't we just sign furcal to a one year deal, then sign J.J. next year. Yes you would get a first round pick, but how long will it take that pick to be MLB ready? So I agree with the article and would like to thank the poster for identifying the value of Miller.

Anonymous said...

@ Matt P. You're way off. A rookie's war doesn't automatically go up if they pitch more innings. Zimmermann's war would have most likely been lower if he pitched 170 innings due to arm fatigue. Miller's war was at a 2.5 at the all-star break, around 100 innings pitched. It went down from there. And according to your innings pitched = higher war theory, Miller would have a higher war next year when he throws 200+ innings

Liam said...

@anonymous, to be clear I wasn't saying Hardy for Lynn or Miller strait up, I was saying Hardy+Rodriguez+Schoop would be a fair deal. Miller's valuable but that's potentially a ton of value in return depending how the Cards value Schoop and Rodriguez.

I don't think anybody's saying Miller can't be a 3.5 WAR pitcher or even better, but when projecting average value based of a rookie season there's so much uncertainty that you have to low ball it. You can't just trade based on upside, and every team knows that. Matt P's point shows that talent and upside are far from a guarantee of future performance, even with a strong rookie year. For every 2 WAR rookie who blossoms into the guy he looks like, there are far more who fall flat, regress or plateau.

Matt P said...

Anonymous - So why is it that when one compares Miller's performance to other rookie pitchers 23 or younger who threw 140 innings or more from 2005 to 2013 that his WAR was good for 17 of 22? It's almost as if the amount of innings thrown matters.

The guys with similar performance to Miller (unless you look at ERA) didn't really turn into stars. Johnson is a star. Moore has promise. Olsen and Britton are busts. Porcello, Niese and Cueto are decent but not great.


Jon Shepherd said...

I am still not sure why this population of 22 is so important. There is a major assumption here that I think has not been validated.

Matt P said...

"Further, with the current depth of the Cardinals minor league system, they are not in need of adding prospects like Schoop and Rodriguez."

With the current depth of your major league system, you're not really in need of adding pretty much anything but a shortstop.

Is there anything else your team would be interested in besides a star outfielder with years of control like Trout or Stanton? I don't think so but maybe you'll disagree.

Sure, your minor league system is loaded also but it's sorta prospects or nothing.

Sorry that the link came out poorly in the previous post.

Matt P said...

"I am still not sure why this population of 22 is so important. There is a major assumption here that I think has not been validated."

It puts into context what Miller actually did. It's not like the 19 guys who did it before 2013 are all stars. Two of them were Matusz and Britton. Most of the ones who are stars were considerably better than Miller.

Matt P said...

Here. Suppose I lower the innings limit to 100 and keep everything else that same.

If I do that, then Miller's performance is good for 26th out of 56.

Suppose I compare him to all qualified rookie starters with everything else the same.

If I do that, his performance is good for 31st out of 165.

His innings are inflating his WAR. That's all.

Jon Shepherd said...

It informs what he did, but does not necessarily mean that it informs us what he will do. That has not been established.

Jon Shepherd said...

His innings are inflating his WAR? I don't think that is a valid conclusion based on the data. In fact, as mentioned, the innings worked against him. What about the others? Just not a silver bullet stat here.

Matt P said...

I find it odd that innings would work against Miller but not against other similar rookie pitchers. And why shouldn't his second half performance count? If it was arm fatigue after a mere 100 innings then that's a problem. If it was regression then that's a problem.

I do agree that if I only saw his first half performance then I'd be more impressed. 2.5 WAR in 100 innings is impressive. 2.1 WAR in 180 innings is less so.

Anonymous said...

@ Matt P - I'm sorry Matt, my post was confusing. My point was innings does matter. And that's why comparing zimmermann to miller is way off. If the cardinals didn't sustain injuries to their starting pitching(carpenter, Garcia), Miller would have been around a 100 or so innings, but he was asked to step it up and forgo numbers for the greater good, which he did quite well (15 wins) So believe what you want but he is the real deal. He's not the next verlander but certainly a top of the rotation guy

Jon Shepherd said...

Matt...the issue is that it is a potentially incomplete data set due to selection bias and that its predictive capacity is unfounded.

Jon Shepherd said...

If you wish to deal with the predictive issue...use a predictive stat like xFIP, not fWAR.

Population issue means accounting for players not in MLB, which is difficult and at least should be noted as a weakness.

Matt P said...

"My point was innings does matter. And that's why comparing zimmermann to miller is way off."

That was my point. Sounds like we agree.

"Population issue means accounting for players not in MLB, which is difficult and at least should be noted as a weakness."

When I say that he's 17th out of 22 that doesn't mean I think he sucks. Even if he was 22nd out of 22nd that still wouldn't mean I think he sucks. Even the bad pitchers are very good relative to the rest of the population (pitchers not in the MLB).

The fact that he's even in the population is impressive and means he has an exremely high chance of pitching multiple years in the majors provided he doesn't suffer a serious injury. There's a reason why it's such a small population.

His performance already is better then most pitching prospects even if he never pitched another inning.

But most rookie pitchers end up being worth far less than either 1.5 or 2 WAR a season. And very few end up being worth more than 2WAR.

The ones that do end up being worth more than 2 WAR are very special.

Jon Shepherd said...

I guess I am not sure where the conversation is at.

Anonymous said...

@Matt P. We agree, but in a different context. You are saying that Shelby Millers WAR is better because he pitched more innings. I'm saying it's unfair to compare him to someone who didn't pitch 170 innings because their performance could be significantly different with more innings. Miller had more strikeouts in his first 70+ innings then the final 90-100 innings. His era ballooned from sub 2 to above 3. He never threw more than 140 innings in the minors then you ask him to shoot for 180 in the majors? This is why we didn't see him in the playoffs. Is he hurt? Will he get hurt next year? Cardinals GM doesn't seem to think so otherwise he would be traded already. I just disagree with how you're evaluating Miller based on rookies over the last eight years. I could be wrong, but I watched him pitch in a tough division and I think he's going to be special regardless of what his rookie season stats looked like.

Matt P said...

"I'm saying it's unfair to compare him to someone who didn't pitch 170 innings because their performance could be significantly different with more innings."

I agree with this.

"I guess I am not sure where the conversation is at."

The reason why I made my comment is because I assumed when I looked at his stats that he was going to be a sure-fire TOR pitcher.

And if we could erase his second half and just looked at his first half stats then I'd think of him as a definite TOR who it would be reasonable to project at 3 WAR a season. Not a #1 pitcher but a definite #2. Guys like Harvey, Fernandez, Strasburg, King Felix were all better their rookie season.

Let me just put this in context for Anonymous. If Miller becomes 3 WAR pitcher for the next five years then he'd be better then every single Orioles starter since Mike Mussina with the possible exception of Bedard (and I'm going with Miller). He'd very arguably be in the top 5 Oriole starters since 1980 especially if he stuck around. I know our pitchers have sucked but...

But when you add his first half and second half together then the picture gets more shady. Instead of looking like a surefire #2, it's closer to a #3. It's because his IP/S was terrible and he pitched a lot of innings. I don't believe you can just ignore his second half.

I thought that was interesting. I was expecting him to look like a guy who would be worth 3.5 WAR a season. When it ended up being between 1.5 and 2 WAR I was surprised.

He could end up doing better. Even if he doesn't very few prospects end up becoming 1.5 WAR a year pitchers. Just didn't seem to be as good as I would have thought.

Jon Shepherd said...

Just an FYI...I used WAR here simply to provide context of value. We really should use something like xFIP or FIP if we want to be predictive. Remember that WAR in either of the main forms has potential to misjudge talent.

Unknown said...

As a Cardinals fan, no way do we give up Shelby Miller for that package. Hardy plus prospects might get you Lynn, but Miller has too much potential for one year of Hardy.