Thank god someone actually write a article on Melky Cabrera that doesn’t blab on and on about him being El Slappy with no power and no tools.

In all the Johan Santana trading wind thing it got a little nutty how badly people bashed Cabrera. it’s getting pretty stupid when people are oblivious to the plus but only see the negative on a 22 year old player.

In all the Johan Santana trading wind thing it got a little nutty how badly people bashed Cabrera.

I think it was a way for people to deal with the possibility of losing Melky.  I saw similar stuff with Hughes, although not as much as with Melky.

Yes, but doing so you lose track of what you really have. I think there is a very very significant chance that Melky Cabrera out do Carlos Gomez for his career. it’s not a bash on Carlos Gomez’s crazy tooooools but the actual most important toooooools is you know… actually hitting the ball and not running really fast back to the dug out after you pop out or strike out.

As for Hughes, there were indeed a slew of articles questioning the young man’s work ethic - although I find it a bit strange how a lazy bones could make the majors by the age of 20.

With regards to Melky, my main quibble with him last season was the dip in his OBP.  On numerous occasions, Melky seemed to swing at balls above his shoulders in 2007.  In 2006, he seemed to lay off those pitches with more frequency.

The plate discipline (rather lack thereof) worries me too, especially b/c IIRC it’s more in line with what he did in the minors than what he did in 2006… his patience in ‘06 may be the outlier.  Hard to say with a 22 year old, though.

I’d take that 65% projection.  The 80%... that’s basically Derek Jeter.  I don’t see it for Melky.  Prove me wrong, lil guy!

-.8 bases - I read that as -8 at first.  The physics instructor in me pipes up to ask, Please to use -0.n consistently.

Thanks for the continuing fascinating work.

“Carlos Gomez’s crazy tooooools”

Here’s a question about talent evaluation question—you hear about “toolsy” guys, but what’s the difference between having a tool and using it?  If one of the tools is power, the kid should hit for power in HS, or College, or Legion ball.  Now I guess you could as a scout say well, the kid’s got the body to hit for power if he gets more selective. 

But what about average?  How can you look at a guy and say, “Well, I think he’s got the ‘hit-for-average’ tool”?  Unless he’s actually, er, HIT FOR AVERAGE—but that’s statistical analysis.

You hear also about guys in the minors with tools, but it’s even crazier to say someone in pro ball has a “tool” if they’re not using it.

Sorry for the rant.

There’s a warrant out for Scott Spiezio’s arrest for drunk driving.  The real surprise is that the Cardinals actually cut him.  Man, is that a franchise with a sickness.

If one of the tools is power, the kid should hit for power in HS, or College, or Legion ball.  Now I guess you could as a scout say well, the kid’s got the body to hit for power if he gets more selective.

I think bat speed is the first thing scouts look for when they say someone has the tools to hit for power.

How can you look at a guy and say, “Well, I think he’s got the ‘hit-for-average’ tool”? 

I think the scouts look for a somewhat level swing and they also look for guys who make consistent contact.  I’d think line drives are also a positive indicator for the hit for average skill.

You hear also about guys in the minors with tools, but it’s even crazier to say someone in pro ball has a “tool” if they’re not using it.

It depends on the tool I think.  If someone’s fast or has a good arm that’s pretty clear.  If they are projected to hit for average or power but are doing neither in the minors then yeah, I think it’s pretty risky to think they’ll all of a sudden turn those tools to skills.

I’d be shocked if Carlos Gomez even approaches average with the bat, although he could end up average or better overall if he’s a good baserunner and plus defender.

Okay, okay, Melky.

I’m in the “optimistic but realistic” camp.  I grok all the reasons why he could turn out to be better than average, but until I see it, he’s an average player.  Which is valuable in itself, of course.

I was in favor of the Johan trade with Melky/Hughes as the big chips.  Yet at the same time, I’m glad to see the homegrown guys stay.

Man, I loved it when Melky robbed Manny over the LF fence.

Okay, thanks SG for being the voice of reason.  I was talking about the hitting tools, it’s pretty easy (even for me) to see how speed/glove/arm can be evaluated even if they’re raw.

Interesting article on hardballtimes today.  This is the type of stuff that *could* help scouts (and “stats-geeks”) try to quantify if a toolsy-player could be expected to suddenly improve.  E.g. if a player has a lot of balls that have high velocity off the bat, but are turning into outs, you could quantify that player is unlucky.  Or a number of line-drives that are caught on the warning track.  Type of thing where if the #‘s say the player probably won’t realize what the scouts say his potential are, we could probably have a new tool to help decide who is “right”.  Hopefully there is enough data to make this kind of thing possible in the next decade or so.

Whether you think Melky will get better or not, what’s undeniable is that he’s a tremendous bargain RIGHT NOW if he just meets his baseline projections.  That’s a great property to have:  a guy who doesn’t hurt you in center field at minimal cost, who, at 23, still has time to develop and improve.

And that doesn’t even factor in the youth and enthusiasm he injects into the team.

I’m glad he’s still here, and count me among the optimists who think he can eventually develop into the player his 80% projection points to…

The plate discipline (rather lack thereof) worries me too, especially b/c IIRC it’s more in line with what he did in the minors than what he did in 2006… his patience in ‘06 may be the outlier.

If you take out that horrific April when he basically hit like me and walked only three times, then the difference between Cabrera’s discipline in 2006 and 2007 is about one walk every eight games.  We have to expect some year-to-year variation, don’t we?  I don’t think this is a big enough difference to qualify as a red flag.  For the most part, you aren’t going to walk as much when you aren’t hitting as well.

Thanks for the link, Mike K.  Interesting side note in there about how positioning can affect a stat like ZR.

“And that doesn’t even factor in the youth and enthusiasm he injects into the team.”

Some day I’m going to going to go back through old articles and create a list of all the players in the last 10 years who’ve been touted as bringing “youth and enthusiasm” to the Yankees.  This year we’ve already seen the Joba article.

I don’t doubt Joba’s Y&E, or Melky’s, or Cano’s…but maybe there are young, enthusiastic players on the team every year.  I bet I can find articles about Shane Spencer, Ricky Ledee, and Andy Fox saying the same thing.

Also, this is Melky’s last cheap year.  If he doesn’t take the next step forward, then next year he’ll be an average player making an average salary.  Not exactly a tremendous bargain any more.

MC in VA,

1 walk every 8 games… isn’t that similar to the Crash Davis line about the difference between hitting .250 and .300?  And that’s taking out April, which, well, you can’t really do unless there was an injury.

I’m just trying to be realistic here.

If Melky hits his 80% baseline, which is rather unlikely, he will perhaps earn $4 million dollars.  If he remains a league average CF more or less, I can’t see him costing more than $2 million.  I don’t see how those numbers will hurt the Yanks salary-wise.

“I don’t doubt Joba’s Y&E, or Melky’s, or Cano’s…but maybe there are young, enthusiastic players on the team every year.  I bet I can find articles about Shane Spencer, Ricky Ledee, and Andy Fox saying the same thing.

Not sure what the point is, but all three of the guys you mentioned (defintely not stars)were more than just Y & E at various times. Spencer in particular did pretty well and had some big moments in the post seasons. Didn’t he start the Jeter flip play and I remember him taking Shilling to RF for a HR in 01 series. Perhaps your comment wasn’t to disparage these particular guys but more to say that the ” clubhouse intangible guys” are present every year in one form or another.

““ clubhouse intangible guys” are present every year in one form or another”

Yes.

Also that those “intangible guys” who “loosen up” the Yankees are in large part a media creation.

1 walk every 8 games… isn’t that similar to the Crash Davis line about the difference between hitting .250 and .300?

Not quite.  One more hit every eight games would be the difference between hitting .250 and hitting .288, which isn’t nearly as sexy as hitting .300.

I’m just trying to be realistic here.

So am I.  Realistically, I think the most likely explanation of the difference between his walk rate in 2006 and 2007 is random variation.

Melky was worse in just about every category.

Melky managed to figure out a way to get hit with the ball 63% more than last year.. so thats something. He’s also getting better at preventing those inside the park homeruns. I like Melky. 65% is doable, 80% will have people cursing anyone who ever thought Melky should be put in a deal for that Santana guy.

I don’t see how those numbers will hurt the Yanks salary-wise.

I’m not saying they will.  I’m just saying that he ain’t playing for the minimum any more.  It’s the difference between a reasonably good, cheap player and a “tremendous bargain.”

Didn’t he start the Jeter flip play…

Yes he did.  But I’m not sure how much credit he deserves for heaving a nine hopper in the general vicinity of home plate from the RF corner.

Are walks a matter of random variation?  After all, they seem to predicated far less on luck than, say, batted balls.  If Melky’s BB rate went down, it was almost certainly because he was being less selective.  Also, pitchers probably realized he wasn’t much of a threat to take them out of the park so they may have been more willing to groove a 2-0 pitch and let the defense take care of things.

I also think it’s worthwhile to look at Melky’s arm in contrast to the other option the Yanks have available right now: Damon.  And the man can’t dent a loaf of bread with a thrown ball.  1st to 3rd just rarely happens on line drives to center anymore, and they were just about guaranteed with both Damon and Williams.  This swing is just huge.  Do any of the defensive metrics take into consideration the reduction in baserunning advance attempts that comes with certain players?  I know that Ichiro’s arm (whether deserving or not) supposedly cut down siginficantly on extra bases taken on balls to right for a while.  I’d be interested to see if Melky has reduced even the attempts to take the extra base.

The article I linked to at the Hardball Times in the defense section considers holds as well as assists when rating arms.  Melky vs. Damon is probably at best 10 run swing on arms alone (-5 for Damon, +5 for Melky).  What’s interesting is that same article rates Matsui’s arm as a +3 or so.  If we look at the different OF configurations with the assumption that Abreu stays in right:

1) Damon in LF, Melky in CF.  Let’s say Damon is +5 on fly balls and -5 on arm, and Melky is +5 on fly balls and +5 on arm.  That means LF and CF is +10. 

2) Matsui in LF, Damon in CF.  Matsui is probably around a -10 on fly balls, but let’s give him the +3 on his arm.  Damon is around -1 on fly balls and -5 on arm, so that means LF and CF is -13? 

That’s huge, a two win difference.

What’s interesting is that same article rates Matsui’s arm as a +3

I think that’s because most left-fielders don’t throw well, which is something we don’t appreciate when we see a slow runner go from second-to-home on a single to left.  I read somewhere once (Rob Neyer maybe?) that if you can both run and throw, you play center.  If you can throw but can’t run you play right.  Run but can’t throw you play center.  Obviously not a hard-and-fast rule, but I imagine you’d generally see outfields arranged that way.

you play center

play LEFT.  I need a nap.

Are walks a matter of random variation?

Who said that?  But come on, everything is going to vary some from year to year.  My point, which I thought was pretty clear, was about the difference between walking 10.7% of the time and walking 7.6% of the time.  If someone’s “true talent” plate discipline is walking 9% of the time, then would it really be so surprising to see 11% one year and 7% the next?

Melky’s 43 walks estimate his true walk rate over last season to about +/- 7 walks or 15%.  Seeing the observed walk rate falling or rising by 1.5 sigma (for 43->9%, and going to 11% or 7%) wouldn’t be that surprising.

Check out Giambi’s BB/PA over his career.  He’s been consistently good or better throughout his career, but still a fair amount of year-to-year variation.

“This is one of the problems with getting defensive metrics more widely accepted. When two systems purported to measure the same thing show a 40 run difference, that’s pretty tough to justify.”

I have to say that I’m still skeptical.  I remember one of the systems giving Hunter approximately +20, -20 and +20 again in consecutive seasons without any known injury.

What makes me particularly skeptical is the source of the information.  Caught balls & the like: not a problem.  But chances?  If a ball’s hit over second base, is that a shortstop chance or a second baseman’s chance?  If the call is made by observers - well, does anyone recall the analysis of official scorers and errors from “The Diamond Appraised?” (I found it shocking when I first read it.)

In other words, the “garbage in, garbage out” argument seems pretty compelling to me in this instance.  If you can’t trust the raw statistics, doesn’t it invalidate the more heavily processed statistics that appear to depend upon that raw data?

Things like a 40-run disparity only confirm the skepticism.  Please let me know what I’m missing!

Also, the 80% CAIRO spreads seem strangely low on home runs, although I’m sure that this is due to my not understanding them.  Surely a really, really good season by Melky could do better than this!  I also recall the 80% on Rodriguez being lower than his (admittedly stellar) last year’s total.  Seems odd to me, but - once again - I’m sure I’m just unaware of the reasoning.

If he doesn’t take the next step forward, then next year he’ll be an average player making an average salary.  Not exactly a tremendous bargain any more.

What?  SG has Melky worth 8 million dollars in 2008.  Melky isn’t going to be making 8 million dollars anytime soon.

...those “intangible guys” who “loosen up” the Yankees are in large part a media creation.

Good grief.  I get it. We’re smarter than everyone else because we don’t believe in “intangibles.”  I happened to think Melky’s enthusiasm was worth mentioning.  I never said the Yankees should pay any kind of premium for it.

“Good grief.  I get it. We’re smarter than everyone else because we don’t believe in “intangibles.””

That wasn’t my point either.  I’m sure it’s nice to have those guys around, maybe it helps the team win, at least it’s more pleasant in the locker room. 

My beef is that every year there are stories about how Player X has really loosened up the Yankees, when they used to be so serious.

Can that really be true every year?  If it were, by now the Yankees would be smoking herb in the dugout and shagging fungoes wearing only a cap and shower clogs, they’d be so loose.

It’s just something the media says because they can make a narrative out of it.

Sorry if it came off as a personal attack, SAS.

If someone’s “true talent” plate discipline is walking 9% of the time, then would it really be so surprising to see 11% one year and 7% the next?

Can’t you just look at a BB/PA+ stat (if such a stat exists)? You can use that to normalize against good years for pitchers, good years for hitters, umps, etc.

I don’t know if anybody got the “bargain for me” reference, but I am pretty sure it is Eddie Murphy.

SG has Melky worth 8 million dollars in 2008.

As a free agent.  It’s certainly an interesting thing to look at, but again, if one is going to talk about a “tremendous bargain,” then the relevant comparison is to all players, not just to the subset of players who as a group make substantially more money than the ones still under club control.  This isn’t hard to understand, and it shouldn’t be controversial—good players making the minimum are a better value than good players in their arb years, who in turn are a better value than good players who have reached free agency.

Things like a 40-run disparity only confirm the skepticism.  Please let me know what I’m missing!

Well, here’s one thing you’re missing: guys have good and bad years hitting, so why can’t they have good and bad years fielding?  Andruw Jones went from +23 to -8 to +23 batting runs without any known injury.  Does that make us suspicious of linear weights, or do we just say that Andruw had an off year at the plate?

I’m not saying that defensive metrics don’t have a long way to go, but I don’t know why we should necessarily expect less year-to-year variation in defensive performance than in offensive performance.  The disagreements among the different systems in evaluating the same player in the same season are a much bigger issue to me.

but I don’t know why we should necessarily expect less year-to-year variation in defensive performance than in offensive performance.

I think we should expect LESS, simply because defense has less of an impact than does offense.  E.g. you’re never going to find a player that “creates” 100 runs on defense, but on offense you’ll see a number of those players every year.  However, your general point holds; players are going to have good years and bad years both in offense and defense.  Doesn’t mean the metric isn’t good.

E.g. you’re never going to find a player that “creates” 100 runs on defense…

Well, you don’t find very many players who create 100 runs more than average for their position on offense, either.  So while you might expect less variation in absolute numbers of runs when the totals are lower across the board, I’m not sure why there wouldn’t be just as much variation relative to league average (or to an individual player’s established performance level).  And you can always convert to RS+ or some such if there’s a scaling problem.

MC, your point is well taken, but I think there is a reply - in fact, two kinds of reply.
First: when Andrew Jones went through that, people noticed.  IIRC, however, Hunter was blithely championed as God’s gift to defence throughout the period in question (granted, this is all vitiated by the fact that I don’t recall exactly where I saw this, nor, even worse, can I tell you which metric was being used) - in other words, no one noticed.  And while I admit that variation in offensive success is more noticeable precisely because it is more easily quantified & therefore more easily tracked, still, in almost any metric I can think of, a swing from the +20s to the -20s and back would just have to be a glaringly obvious thing, no?

And second: you’re right, of course, that variation wouldn’t necessarily make us suspect the metrics.  But where the metrics are already controversial, where the raw data depends to a certain extent upon subjective judgment (it does, doesn’t it?), where we have studies that show that the subjectivity in such things can be far, far greater than we’d like to hope (e.g., the referenced chapter from “The Diamond Appraised”), where we encounter extremely non-trivial divergence in the results of systems based upon this raw data (e.g., the divergence that evoked the comment by SG that spurred this issue), THEN the meaningfulness of the metric really might be called into question.  And in THAT case, cases where a given metric shows very large variations in performance and those variations go unnoticed in the practice of the game… well, those cases then become supporting circumstantial evidence in an argument for subjecting the system and the data it uses to scrutiny and doubt.

Again, if there are aspects of the current approaches to defensive analysis that I’m not taking into account, I’d be happy to learn about them!

Also, the 80% CAIRO spreads seem strangely low on home runs, although I’m sure that this is due to my not understanding them.

Basically, I use standard deviations of the component stats to build the percentile projections.  The idea is that there is about a 60% chance that a player’s performance will fall within around one standard deviation of his baseline projection, so 30% better and 30% worse.  Since Melky’s power output has been so low in the minors and majors as far as HRs, he’d have to perform outside that range to get a higher HR total.  I don’t think given the limitations of projecting players that there’s any use in projecting outside that range, but we can say there’s a 40% chance that Melky will perform outside of his 20% to 80% percentiles.

And second: you’re right, of course, that variation wouldn’t necessarily make us suspect the metrics.  But where the metrics are already controversial, where the raw data depends to a certain extent upon subjective judgment (it does, doesn’t it?)

Year to year defensive fluctuations are expected, as MC says.  Players face easier or harder chances in a given season, their positioning may change, they may have injuries.  As far as the subjective judgement, at least for zone rating there are three separate scorers who sit apart from each other, which cuts that down a lot.  Also, what about balls/strikes and singles/errors.  Aren’t a lot of those subjective judgement calls too?

Defensive metrics can be flawed, and in certain cases where there are big divergences there is certainly reason to question what they are saying, but those cases are the minority.  For the most part I think they can tell us more than our eyes can.

where the raw data depends to a certain extent upon subjective judgment (it does, doesn’t it?)

It really isn’t as subjective as lots of people seem to think it is.  There are well-defined zones, there are areas of the field that are defined as not being any individual fielder’s responsibility, and since multiple observers are used for each game, the generation of the raw data doesn’t depend on just one person making a judgment about where and how hard a ball was hit.  Adding video analysis should make it even less subjective.

Hunter was blithely championed as God’s gift to defence throughout the period in question (granted, this is all vitiated by the fact that I don’t recall exactly where I saw this, nor, even worse, can I tell you which metric was being used) - in other words, no one noticed.

It’s pretty hard to evaluate this without knowing what metric and what years you’re talking about.  As for Jones, the truth is that most people didn’t notice his off year at the plate in 2001, because he still had 34 HR and 104 RBI.  In fact, most people would probably call his bounce-back year in 2002 the off year in that three year period, since his runs scored and batted in were lower.  And the fact that the MSM might continue raving about someone’s defense in a year when the stats say he’s slumping is pretty meaningless.  That’s sort of the whole point of trying to develop these advances defensive metrics in the first place.

...by now the Yankees would be smoking herb in the dugout and shagging fungoes wearing only a cap and shower clogs, they’d be so loose.

Now my brain hurts because picturing Godzilla (even though I think he’s awesome) do that induces sanity loss.  This actually applies for many others on the Yankees as well, it’s just that Matsui comes to mind first, bizarrely.

“Also, what about balls/strikes and singles/errors.  Aren’t a lot of those subjective judgement calls too?”

The problem with singles/errors is exactly the model I’m using to cast doubt on this stuff.  The study I keep referring to in “The Diamond Appraised” focused on that issue and demonstrated convincingly (to most people at the time, at least) that the subjectivity in determining errors really is more than great enough to make the ‘error’ a genuinely meaningless statistic.  Thus, the singles/errors isn’t a counterexample, it’s the example I’m using as a potential (if unhappy) model.

On the other hand…

“As far as the subjective judgement, at least for zone rating there are three separate scorers who sit apart from each other, which cuts that down a lot.”

I didn’t know this.  There’s no doubt that this is likely to make a huge difference.  Are the observers tied to a given ballpark, as are official scorers?  If not, if they rotate around the country, that would get significantly alleviate the central problem documented vis-a-vis the error in “The Diamond Appraised.”

“It’s pretty hard to evaluate this without knowing what metric and what years you’re talking about.”

I know, you’re absolutely right.  I’m hoping someone else might have encountered the same study, or at least the statistic in question, and inject the facts I’d like to have been able to add myself.

“...by now the Yankees would be smoking herb in the dugout and shagging fungoes wearing only a cap and shower clogs, they’d be so loose. “

That was really funny.

Well, you don’t find very many players who create 100 runs more than average for their position on offense, either.  So while you might expect less variation in absolute numbers of runs when the totals are lower across the board, I’m not sure why there wouldn’t be just as much variation relative to league average (or to an individual player’s established performance level).

The 100 runs I was talking more absolute numbers.  E.g. ARod typically (according to runs-created) creates 120 or more runs a year.  There are several players that create 100 or more runs a year with the bat (though still a subset).  There are NO players that save 100 runs with the glove.  Adjusting against an average for position (+/-) lowers the range, but a good offensive player could still vary by 60 or more runs-created between a good/bad season.  I doubt you’ll see that much of a difference in RUNS on defense. 

But sure using a % change, or number of Std deviations, etc, you’d probably see a similar variance between offense and defense.  I think we’re talking the same thing, I was just trying to clarify a point, and instead muddied it.

As a free agent.  It’s certainly an interesting thing to look at, but again, if one is going to talk about a “tremendous bargain,” then the relevant comparison is to all players, not just to the subset of players who as a group make substantially more money than the ones still under club control.

The wording could use some work, but I think that free agents are the right universe to compare to when considering value—not because free agents have representative salaries, but because they are the players who the Yankees have the option to acquire on short notice.  From the perspective of the Yankees decision making, how Melky’s cost effectiveness compares to Jacoby Ellsbury isn’t interesting, because the Yankees don’t have the option to hire Ellsbury at his listed salary (without giving up players to trade for him).  In contrast, if there’s a free agent with a higher expected net 2008 value (value minus salary) than Melky, that’s an indicator (if not always a decisive one) that the Yankees should sign that free agent and give him the center field job.

What on earth does any of that have to do with the conversation we were having above?  Melky was described as a “tremendous bargain.”  I pointed out that this is only true for one more year.  He will be more expensive next year no matter what happens.  He will be a lot more expensive next year if he has a breakout season.  A young player who gives you league average performance for the minimum salary is correctly described as a tremendous bargain; a three to six year veteran who gives you league average performance for a salary in the range typical of arb-eligible guys who are regulars may still be a very good value, but he’s no longer a tremendous bargain.  Why does everybody want to make this so complicated.