Though the deal is not yet official, ESPN's Jerry Crasnick (author of License to Deal) is reporting that the Rockies are trading outfielder Matt Holliday to the Oakland A's For starting pitcher Greg Smith and two other players, perhaps including closer Huston Street and OF Carlos Gonzalez.
Unfortunately, at this point, nobody's exactly sure who will go from Oakland to Colorado, though one AP report speculated that closer Huston Street and outfielder Carlos Gonzalez could be Denver-bound as well. So let's go with that.
For the Rockies:
The Rockies, who finished 15th in the 16-team National League in ERA, are obviously in need of pitching help, as they always are. Smith, they hope, can eat up innings for them effectively and somewhat improve the dearth of pitching.
Though he finished the year just 7-16, his 4.16 ERA was the best among qualified Oaklanders (all both of them), and 25th best in the American League, and his 190 innings were 22nd best. At least that's how Scott Boras would spin it.
Smith came to Oakland in the Dan Haren trade and he significantly outperformed what anybody expected of him, but he walked 87 batters and struck out only 111, and his minor league numbers do not suggest that he'll ever be much more than a LAIM.
Smith came to Oakland in the Dan Haren trade and he significantly outperformed what anybody expected of him, but he walked 87 batters and struck out only 111, and his minor league numbers do not suggest that he'll ever be much more than a LAIM.
In fact, before this year, Baseball Prospectus had his closest comparables as a bunch of lefties with very, very short careers. That he's done as well as he has already defies all logic.
In any case, he's going to Coors, where the thin air and his lack of a reliable out pitch will likely cause his ERA to balloon over 5.00. Baseball-reference.com suggests a modest ERA increase of just half a run, but their algorithm doesn't compensate for how outs get made and runs get scored. A finesse lefty, without blinding speed or a sharp sinker or a big curve, naturally has to rely on his defense, and Smith, a severe fly ball pitcher, is no exception.
How severe? Out of 89 qualified MLB pitchers in 2008, Smith's Ground/Fly ratio was 82nd. And he's leaving sea-level Oakland, where a lot of those pop flies were either caught in the outfield or in the expansive foul territory.
In any case, he's going to Coors, where the thin air and his lack of a reliable out pitch will likely cause his ERA to balloon over 5.00. Baseball-reference.com suggests a modest ERA increase of just half a run, but their algorithm doesn't compensate for how outs get made and runs get scored. A finesse lefty, without blinding speed or a sharp sinker or a big curve, naturally has to rely on his defense, and Smith, a severe fly ball pitcher, is no exception.
How severe? Out of 89 qualified MLB pitchers in 2008, Smith's Ground/Fly ratio was 82nd. And he's leaving sea-level Oakland, where a lot of those pop flies were either caught in the outfield or in the expansive foul territory.
Without a true strikeout pitch, he'll be forced to throw pitch after pitch of his breaking stuff, which won't break as much as he's accustomed to. Those offerings can be repeatedly fouled off, until eventually he'll have to throw his middling fastball over the plate, at which point the National League's hitters will tee off.
Smith pitched 190 innings this year with an adjusted ERA almost average (an ERA+ of 97), but I'll be surprised if he can even stay in the Colo-rotation next year.
comments (11) write a comment »
write a new comment
about 1 month ago
It started strong, until you got to Holliday...
"Still, there's no question that he's benefited, and no doubt that he'll see a decline in his numbers, at least on the face of them, in Oakland."
You did utilize some ballpark factor stats, but not accurately.
How many games did Holliday play in Coors Field last year? By your analysis, all of them. Which is obviously incorrect.
Further, you utilize Holliday's road splits as rationale that he will take a substantial step backward. While everyone knows that split stats are terribly inaccurate, often-times they can tell a story. But you chose to ignore the entire story, instead, flipping to the last chapter to figure out the results.
Did you consider that 70% of Holliday's career road games have been at pitcher havens in Los Angels, San Diego, and San Francisco? It's then safe to gather that Holliday will have a vastly improved road split as an A.
But even if you ignored where Holliday was playing his road games, you ignored an upward trend-a significant one at that-within his road splits. Your split stats utilize figures that date back to Holliday's rookie season, five years ago. For an understanding of how poor that analysis is, how about checking in with the top 3 pitchers in the NL during the 2004 season (I'll save you the legwork, Johnson, Sheets, and Schmidt led the league in strikeouts, want to bet on any one of them doing so in 2009 based on that? I'll give you 100 to 1 odds and thats still a $10 bet you don't take).
But let's check out that upward trend. In 2007 Holliday posted a road OPS of .860 while 2008s landed at .892. Neither number is going to win him an MVP award, but in the context of predominantly in pitchers parks, AND on the road, that's fairly impressive, wouldn't you agree?
The fact is, Holliday has grown as a hitter and developed a superior grasp of hitting on the road. His road OPS has increased every season he has been in the league.
Because of those two obvious factors, why is it that you would believe Holliday to be a WORSE hitter in Oakland, then he was on the road in 2008? You certainly did not assert anything regarding league differences. But I would say playing for a contract would easily negate that. If not, being at a stage in his career where improvements can be assumed would.
In other words. Your analysis ignored a lot of key factors. Too many, in fact.
about 1 month ago
It does not seem that I mentioned how going to the tougher league would affect Holliday, though I intended to do so. I think I figured that any deficit derived would be offset by the fact that more runs get scored in the DH league anyway, so I didn't mention it specifically, though I should have. If anything, that's a knock against him, too, though a small one.
Otherwise, though, Brandon, you've got me wrong on several counts.
1) I *did* acknowledge the trend that his road and home stats are not as divergent in 2008 as they once were. The sentences just before the one you quoted read:
Holliday's been great in Colorado over the course of his career, hitting .357/.423/.645 in 359 games, with a respectable .280/.348/.455 on the road. That split was not as severe this year as it had been in the past, only about a 100-point OPS gap, instead of about 250 or more.
2) The adjustment I did *was* based only on his home stats. I left the road stats alone. As I said:
ESPN's park adjustments show the effects on individual types of hits and runs overall, and if you apply those ratios to his 2008 home stats, and then re-tally...
I skipped some of the messy details because, frankly, this article was messy enough without them, but I only adjusted his home 2008 stats based on the ratios of park factors reported on ESPN.com. I realize that one-year park factors can vary a lot, but it seemed the closest thing to fair.
For example, Holliday hit 15 homers in Coors this year, with a PF of 1.299 for HR. Oakland's HR PF is 0.988, a ratio of 1.31 between them. 15/1.31 = 11.45 ~ 11 HR in Oakland, and then adding back in the 10 HR he hit on the road in 2008 he gets 21.
I did the same for Runs, Hits, 2B, 3B, Walks, RBIs (using the Runs PF ratio) and SO (had to figure out my own PF for that one). It's not perfect, but it's a little better than the adjustments that baseball-reference.com does. If I were going *forward*, say, taking his projection for 2009 at Coors and adjusting for playing in Oakland, I probably would have used 3-year or 5-year PF's, but since we were just asking the question of what would have been in 2008, it didn't seem so outlandish to use 1-year factors.
3) Not sure why you think I used 5-year old park factors, or for that matter, what exactly Ben Sheets has to do with any of this, but I didn't, so you and Ben can relax.
4) As for the road games, well, I couldn't exactly make every little adjustment, could I? I’d be doing calculations until May, and then my taxes would be late, and we can’t have that.
I left them alone, as they were, in fact, respectable, and there was little reason to think that there would be a huge adjustment needed for them. Dodger Stadium and Petco are certainly severe pitchers' parks, but AT&T is simply not. It had an overall factor of 103 this year (see baseball-reference.com) and is usually about neutral. It's hard to *homer* there, sure, but it’s basically neutral with regards to run scoring overall.
5) And even if it *were* a pitcher's park, that's still less than a third of his career road games, not 70%. Not sure where you got that number. In truth it's only about 21% (35 at LAD and 40 @ SD) / (359 total on the road). 75/359=0.21
Plus Phoenix is as good a hitter's park as any this side of the Rockies, and he won't get to play there much anymore either, not to mention Cincy, Philly, and Wrigley. That stuff should more or less come out in the wash.
6) And why shouldn't his overall numbers look worse playing half his games in the best pitcher's park in the AL? There are some hitters' parks and some pitchers' parks throughout the rest of the NL, but Oakland's pretty harsh for hitters, so that brings him down quite a bit. I said he'd have hit .299/.393/.486 overall playing for Oakland, and though I didn't mention that his 2008 road stats were .308/.405/.486, the difference there is all but negligible. An improvement in health (or playing for a contract, as you suggest) could easily compensate.
7) So you can see that I did not "ignore the story" but actually I read every relevant detail I could and tried to summarize for you, and I think I actually said so, more or less.
It seems that your factors ignored a lot of my analysis.
from about 1 month ago
1. That's pretty brief. When the rest of your article harps on how much his numbers will go down by not playing at Coors Field, it tends to be assumed that you were suggesting that this season was an aberration.
2. "ESPN's park adjustments show the effects on individual types of hits and runs overall, and if you apply those ratios to his 2008 home stats, and then re-tally..."
That is what you wrote. You then took Holliday's 2008 season (half on the road, half at home) and configured it to McAfee. You more or less stated that "Holliday ONLY played at Coors Field and that if he ONLY played at McAfee, this would be the results."
What you should have done is taken a prorated version of Holliday's home stats in 2008 for an entire season (ie 81gms=162, 15hr=30, etc). Then you could have gotten an idea of how Holliday's numbers would transition to McAfee (don't forget to divide by two after though).
I think we can both agree that is incorrect. Major error.
3. You used 5 year old split stats to help your argument that Holliday can't hit away from Coors. Ben Sheets is probably your best bet in my aforementioned drill, I'd take him ;)
4. One year park factor? Really? How about the year before? Or the year before that? AT&T is traditionally a more pitching friendly environment. It isn't as bad as PETCO, but it is comparable to McAfee.
Oh, and reason enough not to use one season of data, in 2005, McAfee rated as the #7 park to hit in.
5. Error on my behalf. Quickly did the math and just looked at v. specific team rather then at the actual park. Even still, 35% of his road games are played at parks which deflate scoring. This is the highest figure of any team in the Majors.
That is, no other team has to spend so many games on the road in pitchers parks as the Rockies do. Thus, their split stats must be taken with a grain of salt.
Chase is a nice park, but in case you didn't notice, the AL West has the Ballpark in Arlington, an equally as strong place to hit. Safeco is essentially the same as AT&T. And we can both agree that Angel Stadium is VASTLY superior to PETCO.
In other words, Holliday's road schedule improved significantly.
6. "Best"? Well its ONE of the best, but not THE best. It's probably on par with Safeco and the MetroDome.
Also, no one is arguing that his numbers won't look worse in Colorado. What I am suggesting is that he will be a vastly superior hitter then what you are suggesting. Posting an incredible park and league adjusted EQA (tied for 12th in the Majors).
7. I'm sorry, it seems as though you did miss a fair amount of facts.
about 1 month ago
1) Look, man, I'm sorry you *assumed* that I *suggested* something with which you disagreed. If it seemedas though I was "harping" on the subject, it's because there are a lot of factors to consider. If you simply read what I actually wrote, you'd see that.
2) I don't know how I can state this any more clearly. I did NOT assume that he only played at Coors. I took the stats compiled at Coors (half his games, 73 of them this year) and adjusted them to see what he would have done playing those games at McAfee. I then added those back to his actual road stats, another 63 games. We can't assume that he wouldn't have pulled a hammy just because he was playing in Oakland, can we? I'm not saying I only expect him to play 139 games next year, I'm just adjusting what he actually did this year for what he might have done in a different uniform. That to me seems as correct and fair as possible, in so much as pretending that he did something he did not do is neither fair nor correct, technically speaking.
3) I did use 5-year old splits, as well as 4-year old and 3-year and 2 and 1. That's called a *career*, and it's all fair game. And I didn't use them to say that he can't hit away from Coors, only to demonstrate that he, like most players, does not hit *as well* away from Coors. I mentioned his whole career splits because it's just irresponsible to throw away four seasons worth of stats and pretend they never happened when evaluating a player's quality. But when I adjusted his 2008 stats, I only used the splits from 2008, looking at what might have happened, not what will happen.
His annual home/road OPS differences were 355, 273, 313, 297 and 105 points, from 2004-08, respectively. If anything, I went easy on him, and his home road splits in 2008 look more like the abberation, and his 2004-07 numbers are much more consistently and severely skewed.
I still don't know what Ben Sheets has to do with this discussion.
4) Since 2004, Annual Batter's park factors for AT&T (from baseball-reference.com):
104, 100, 99, 100, 103
It's NOT a pitcher's park. It's just not. Not during Holliday's career, anyway.
I stay away from one year PFs when they seem like an abberation, like McAfee in 2005, or Dodger Stadium in 2005-06, or Wrigley in Y2K. But AT&T has been roughly neutral overall for a while now, so I don't see the problem.
5) Yes, Arlington's about as good a hitter's park as Chase, but Safeco is about as tough a place to hit as Dodger Stadium, with an average hitter's PF of 95.6 (range from 91-100) since it opened 9 years ago, and SF is roughly neutral, as I showed already. That means they've got one notable change, from Petco to LAnahfornia Angels of Anaheim Stadium, which has fluctuated a lot from year to year, but admittedly has got to be a nicer place to hit than SD. I would say his road schedule got marginally better, at best, and that this improvement is offset by having to face better pitchers who throw more breaking stuff in a tougher league. Which is why I didn't bother to make any such adjustments. As I said.
6) Fair enough. *One* of the best. No problem.
As for the other part of point #6, I think you've mischaracterized my argument. I really wasn't trying to say that he would be lousy or anything. I fully expect him to still be a very, very good player, and I do hope he's healthy all year, as he could easily hit 30 homers and drive in 100+. But hitting .299/.393/.486 in Oakland in 2008 would have made him one of the best hitters in the AL, you see? He can perform exactly as I suggested and still be one of th 10 or 15 best players in baseball, even though the stats look less impressive.
from about 1 month ago
RE: #4 And I will leave it at that...
You cannot simply take a 3, 4, or 5 year ballpark factor that states an average above '100' and claim that is 'nuetral'. You must look at the rank. Being a 'pitcher friendly' ballpark is relative to the league.
There may only be 5 teams that negatively affect a hitter (say a score of 80 or less), does that mean that ballpark 6 thru 10 are on par with ballpark 20 thru 30, simply because their factor is over 100?
Obviously not.
That being said, if a ballpark (like AT&T) regularly ranks as the 19th worst ballpark for hitters (11th best for pitchers), it is obviously a pitchers ballpark.
This is what I meant about improper analysis. You cannot take a number and draw a conclusion without understanding it. Which is what you did with ballpark factors.
Oh, and if you did what you are NOW saying for #2, you did a very poor job at suggesting it (ie you placed Holliday's 2008 FULL season stats, not his home splits, prorated). Again, improper analysis.
about 1 month ago
Those ratios (100 being average) are only relative to the league, not all of MLB, leaving just 16 teams for AT&T to compare with, not all 30. That's as it should be, since the leagues differ with the DH, and the extra runs scored due to that in the AL will naturally lead the ratios there to be muted a bit. But if you like rank, we'll use rank:
From ESPN.com, AT&T's Runs Scored PF rank (for hitters), since 2004:
8, 20, 16, 17, 11 (average: 14.4)
That is to say, it's usually in or close to the middle third of all MLB stadia, and therefore neither a severe hitter's nor a severe pitcher's park. The average 5-year PF for Runs is 1.012, which is 12th among the 30 current stadia (7th out of 16 in the NL), a little better than average.
And still not a pitcher's park.
As for the adjusted 2008 stats, I still don't get you. I posted his actual full season stats for comparison to what his full season stats would have been if he'd played half his games in McAfee instead of Coors. And I said so. What else would I do? What else has any relevance? Why would I prorate his stats for 81 games, if he only played 73 of them in Colorado? Why would I only post his adjusted home stats? He still has to play the other half of the season.
I adjusted the relevant portion of his 2008 numbers, the home portion, and then added those back to his actual road numbers, unadjusted. I said that in the article: "you apply those [park factor] ratios to his 2008 home stats, and then re-tally..." meaning that you add back in whatever the road numbers were. I'm sorry if you missed it, and I'm sorry I didn't post the damn Excel spreadsheet I used. I can email it to you, if you like.
With all due respect, Brandon, your profile says you're in grad school in Buffalo, at a school that does not offer any grad courses for engineering, science or math, though I suppose you could already have a degree in math and are working toward a secondary Ed certificate or an MBA or something.
I've got an engineering degree from a pretty good university, and have been working in the engineering field for over 11 years, not to mention doing this kind of stuff on the side for more than 6 of those. I've read (among dozens of baseball books) Win Shares, The Fielding Bible, The Physics of Baseball, and the newest Bill James Abstract (all 1008 pages of it) cover to cover. I have read the annual Baseball Prospectus book in its entirety every year since 2002. I have a pretty good grasp of how numbers and equations work, and I know a good deal about baseball statistics.
None of this is to run you or your school down or anything, and is certainly not to say that I know everything or am infallible or anything of the like, only to hopefully reassure you that I have some idea of how to use a calculator and which numbers get applied in what ways.
I appreciate feedback and don't mind having my tie straightened if I'm off base (sorry for the mixed metaphor...I don't have a degree in English), but for the most part, it seems, your corrections all stem from not having read my article carefully enough. It's easy to get lost in a 2,000+ word article, I know, and perhaps some of that is my own fault, as word economy has never been my forte, but math and baseball are among my strengths, so I'm reasonably confident that my analysis is as good and proper as anybody's.
from about 1 month ago
Travis,
I'll get to the rest of your post later. But do you agree with the following logic?
2 students received a grade of 100%.
2 students a grade of 90%.
2 student a grade of 80%.
2 student a grade of 70%.
2 students a grade of 10%.
The aforementioned grades result in an average of 70%. But notice that the middle point is 80%.
By your logic, you are simply looking at the average and suggesting that it is a typical bell curve (where 96.4% of the population falls within 3 standard deviations). HOWEVER, that clearly is not the case in this example NOR is it the case in the ballpark factors.
That is, while San Francisco may play as a 'nuetral' ballpark, relative to the ENTIRE LEAGUE the club favors pitching.
To revisit my previous example. Do you see the 2 students sitting at 70%? They are technically 'average', however, compared to the class, they are below average. There are 2 students at the bottom of the class (let's call them PETCO and the Metrodome) that are pulling them down.
Can you imagine what would occur if we removed the 2 worst and 2 best grades?
The problem is, a '100' suggests that the ballpark plays nuetral although it doesn't consider 'outliers'. In baseball's case, the bottom third are boosting the friendliness of the middle and upper third.
I'm not sure I can explain this in a more simple form.
from about 1 month ago
Travis,
You've got an engineering degree, correct? Have you ever heard of outliers (ie the season where AT&T finished 8th)?
Also, I find it interesting that you began your investigation at the PEAK season. Why not another 3 clicks to the beginning of ESPN's data? Interesting, those three clicks had you miss out on three of the four worst hitting years in AT&T (SBC) history. Those additional three years would promote some havor with your runs numbers (definitely dropping it below 100).
While I won't claim that you did so intentionally, it does have a major affect on the results (dropping it from 14th to 18th-upper half, to bottom half).
Going back to the outliers, it is common for data analysis to eliminate both the high and low number of a series. In the case of AT&T park, we will eliminate the year 2001 and 2004.
In addition to that, a proper analysis would check where the other ballparks 'averaged' and then ranking them accordingly. Simply because AT&T 'averaged' an 18th place finish, does not simply mean it is the 18th rated ballpark for offensive output. You should know that!
about 1 month ago
Hey Travis,
I liked the article. Thought you were right on the money regarding what the Rox got (except I don't remember you mentioning Smith's elbow surgery). I don't totally agree re the A's but can't say you're factually wrong about anything significant.
But about Pac Bell. I have watched pretty much every game played there either in person or on television. Stats are great but they simply cannot tell the whole story. I absolutely promise you, Pac Bell is one of the tougher parks in which to hit.
You're clearly a stat guy and I've got no problem with that, but I won't be able to explain adequately why this particularly one isn't capturing reality. I simply don't have your knowledge of its formulation, but my guesses:
1. Barry Lamar Bonds - he's gotta skew numbers, even if he is one chemically-enhanced man, considering how many ABs he had there from the lefty box.
2. Giants have been really, really bad since 2004. Sure, their offensive numbers should depress the offensive production, but a pitcher working without a net is much more prone to explosion. Especially young pitchers. That combo has created a lot of ugly and crooked numbers on the scoreboard that had nothing to do with the park. Some 10+ run games and the numbers are out of whack.
3. Giants defense. Until last year, they were really, really old. Which meant not a lot of errors, but a lot of hits since they couldn't get in position to make an error. Again, a lot of offense unrelated to the park.
4. The mental aspect. Left-handers come into Pac Bell knowing they cannot pull a homerun. That messes with their approach and, depending on how lefty leaning the lineup is, the entire offensive approach.
Maybe the formula accounts for all this, but it's not accounting for something. Just ask any NL West hitter. Pac Bell is death valley, regardless of what the stats say.
Obviously, I'm not going to convince you. But, if you are a math guy, I'm sure you've heard that statistics can be used to prove anything.
about 1 month ago
Andrew:
Thanks for your insight, and for the respectful tone. I don't always disagree with someone so nicely.
There's no question that AT&T park is a tough place to hit home runs, but overall, either relative to the league or to the whole of MLB, it does not depress run scoring, not appreciably or consistently, anyway. The annual park ratios seem to show that while HR are depressed there, doubles and triples seem usually to be increased, perhaps because of the spacious power alleys and quirky foul lines or something. I've only been to one game there, so I can't say from experiece, but that makes me a little less biased and more willing to believe the numbers instead of my perceptions.
Getting back to Brandon:
I know what outliers are, and I understand their uses. Throwing out a few of them is fine if you've got a sufficiently large sample size, but you can't call 40% of your sample an "outlier" and throw out 2 of your 5 measurements. And you can't use park effects from 2001-03 to evaluate a player who debuted in the majors in 2004. Your agrument was that AT&T is a pitcher's park and has therefore hurt Holliday in his road games, right? Well, he didn't play any games in AT&T before 2004, so I don't see how any of that is relevant.
The changes in ballparks since then (openings and closings, renovations and etc., perhaps changes in the game with fewer homers being hit overall since they started pretending to crack down on steroids) mean that even if AT&T was a pitcher's park then, which it undoubtedly was, it doesn't seem that it is any longer. It's about neutral, both in 2008 and for the last half a decade, roughly. It's reputation as a pitcher's park has long outlived the accuracy of that description.
As for the averages I took, I did both. I averaged the ranks (14.4) AND I ranked the averages (12 out of 30) as I mentioned previously, to avoid just such a pratfall as you describe, and AT&T is not a pitcher's park by either measure, not during Holliday's MLB career. Using the averages of 5-years' worth of data (for most parks, some haven't been around long) AT&T comes out as the 12th best place for run scoring in MLB, albeit by a slim margin above average, 1 or 2%. Or, as I wrote a few hours ago:
"The average 5-year PF for Runs is 1.012, which is 12th among the 30 current stadia (7th out of 16 in the NL), a little better than average."
This is the proper analysis you just requested, which I had already done. As before, it seems that you simply did not read carefully enough.
about 1 month ago
Travis, stop going back-and-forth with this clown. This is what he does. He goes to other people's baseball articles and looks for reasons to talk down to them and make himself feel good. I don't get it, but that's what he likes doing. You'd be best off just ignoring this ho.
write a new comment