Part of CSNPhilly.com


« Phillies Notes: Dom won't stop; dealing Yuni? | Main | Brown makes Law's list, homers again »

Tuesday, March 19, 2013

Comments

From the end of last thread:

Jack: "why should I care at all about Howard or Utley's batting average? That is certainly meaningless."

Ok, I'll take a stab at answering this -- at the risk of opening myself up to ridicule. I care because, deep down, I have my doubts about the proposition that spring training stats are meaningless. I've seen a lot of lousy spring trainings turn into lousy seasons -- or, at the very least, lousy starts to the season. Mayberry, Wigginton, and Halladay last year; Ibanez in 2010 (lousy start); Moyer in 2009; Kendrick in 2008; Eaton every year he was here. I know that, for every bad spring training-turned bad season, you can point me to a bad spring training-turned good season. But at the same time, I strongly suspect that, within the generality that "spring training stats are meaningless," there are some very specific, but poorly understood, areas in which they are NOT meaningless. The slugging pct. thing that I pointed out is one such example and, until about 2 or 3 years ago, no one knew about it. There are undoubtedly other similar areas out there that we don't know about yet.

Hence, I want to see our guys do well because there is much uncertainty about the meaning of spring training stats. In a word, it's safer to see them do well.

BAP: "When a guy's spring slugging pct. greatly exceeds his career slugging pct... it actually DOES have predictive value."

Ron Stone is proof of that.

Lopey is nopey.

If Yuni gets traded for anything more than Michael Cisco, then ST stats are obviously not meaningless to opposing GMs.

The easiest decision of the spring: Betancourt over Utley.

I mean, the stats speak for themselves.

Utley: .200/.347/.325
Betancourt: .450/.455/.525

bap, I see your point, but the reasons most ST stats have had little predictive value in the past are because of the sample sizes and the level of competition faced.

b-r.com has an ST stat line now for players that shows the level of competition, which is "helpful". One wonders if it's a prelude to try to eventually make something predictive out of it?

Hypothetically:

"Player X had a slash line of .355/.410/.500 in ST, but it was against competition with an average Opposition Quality rating of 7.2 - slightly above AA. That has been shown to be negligible in it's predictive value at the MLB level.

OTOH, Player Y had a .305/.350/.475 slash line against competition with an OppQual rating of 9.8 - slightly below MLB level. that has a much higher predictive value and we believe he should have a godd MLB season."

clout: Ron Stone? Did he have a great spring training or something?

I said predictive, not perfect. If it were perfect, Ben Francisco would be an All Star right now.

BAP--

I am curious about this study that shows a surge in slugging % in the spring having predictive value. What years did this study evaluate? Was it during the PED era?

Because if it was, might I suggest that a surge in spring slugging % may be predictive of PED use.

rolo: Except you still run into the problem that a 3-week sample even in the middle of a major-league season is not particularly predictive.

Again, I get it. It makes it more interesting to watch these ST games if you delude yourself into thinking it means something. But it just seems it's still worth reminding people that there's a 99% chance that 99% of what happens in these games means absolutely nothing for the regular season.

rolo: Clearly, there are sample size issues in spring numbers. But what is a meaningless sample size for some statistics may be less meaningless for others. If a hitter gets 80 ABs, and hits .375, you can pretty safely deduce that you're dealing with a meaningless number. If a hitter gets 80 ABs and hits 9 homeruns, that would be considerably harder to dismiss.

You're also ignoring that some of these guys may not even be trying the way they would in the regular season.

I mean, let's say Ryan Howard has decided he wants to work on hitting breaking balls (I know, I know, just go with it). He may let a perfect fastball go by middle of the plate that he would normally crush in the regular season, just so he can see the next breaking ball (and swing and miss, most likely).

So would we be saying the same thing if Dom was batting .125 right now?

Wes: It was a few years ago, because I remember discussing it on Beerleaguer when Ben Francisco was having a great spring training. You can probably find a few articles about it with a Google search.

And, sure, an uptick in spring slugging numbers could certainly be predictive of PED use, since PED use often brings about a high uptick in slugging. But some players also have sudden upticks in slugging without PEDs. I can't see how spring numbers would be predictive of ONLY those in the former group but not in the latter.

I picture a huge 'curve' on a graph with # OF MLB PA and ST RELEVANCE on the axis. Players with 0 MLB PA have very high releveance in their ST numbers while the more MLB PA they get, the lower and lower the ST relevance becomes, to a point where it just doesn't matter anymore around 2000 or so PA.

Being predictive is probably a harder sell, but relevance to the proposed decision makers on the team is surely there.

BAP, you say this:

"I've seen a lot of lousy spring trainings turn into lousy seasons -- or, at the very least, lousy starts to the season. Mayberry, Wigginton, and Halladay last year; Ibanez in 2010 (lousy start); Moyer in 2009; Kendrick in 2008; Eaton every year he was here. I know that, for every bad spring training-turned bad season, you can point me to a bad spring training-turned good season. But at the same time, I strongly suspect that, within the generality that "spring training stats are meaningless," there are some very specific, but poorly understood, areas in which they are NOT meaningless."

Isn't that just a way of saying that, yes, spring stats are meaningless? Because you can say the same thing about any two variables.

"I've seen a lot of people who drink lattes get cancer: my friends Dave, Jill and Mike, for instance. I know that, for every latte-drinker who got cancer, you can point me to a latte-drinker who didn't get cancer or a cancer patient who didn't drink lattes. But at the same time, I strongly suspect that, within the generality that "drinking lattes isn't connected to cancer," there are some very specific, but poorly understood, areas in which drinking lattes DOES cause cancer."

That makes your argument sound dumber than it actually is, but the point is that, when you say that in general you can't correlate a particular pair of variables, but in some individual cases there is a clear correlation, it really sounds like in fact there's no reliable correlation to be found and you're just applying confirmation bias to the cases in which there appears to be a correlation due to chance. The fact that a few cases exist where, in hindsight, Event A was followed by Event B doesn't mean that there's any reliable predictive value to be found in Event A; You have to be very careful to not ignore the huge number of cases that don't play into a particular theory of what Event A means, predictively. I'm agnostic on the specific case of spring training stats... that's just a general point.

That Opposition Quality rating is not completely useless, but there are still major flaws in using even that to help make ST stats predictive.

Suppose Justin Verlander is pitching, but he is working on his curveball and throws way more curves and far less of his other pitches than he would in a real game. And suppose player A goes 5 for 5 with two homers. Using the Opposition Quality metric, it would predict player a to have an MVP type season. (I know I am way over simplifying this for the sake of brevity).

The reality is that in a real game player A would be lucky to go 1 for 5 with a weak single off Verlander.

Stats can be fun. Stats can be useful. But the use of stats will always have flaws. This is why the manager, coaches and scouts actually watch the spring games. They can see the things that stats fail to quantify.

ST stats are certainly not meaningless. Rarely are they predictive. But they do have their use. Just as watching a players approach at the plate and seeing how he is swinging is valuable.

This is why they play spring games. Why they keep track of the stats. Why the coaches watch the game. Take each part by itself and it may not mean very much. But when you factor it all in and look at the big picture, there is meaning in that. That is what teams use to make their roster decisions. Even using the big picture has its flaws. But it's the best we have short of using a crystal ball to see the future.

Is it any wonder that Clout Day is on March 20th when the forces of light finally catch up to the offseason forces of darkness.

and during ST you have to separate performance and stats. If doc had an ERA of 7 but was throwing in the low 90s with lots of movement on his offspeed pitches, i would chalk it up to ST doesn't matter. But since hes throwing in the 80s and laboring on the mound, its reason for concern.

The choice for the last utility spot isn't limited to Galvis or Yuni. There are several guys likely to be on waivers as spring training closes, including:

Jason Donald (Reds)
Carlos Rivero (Nationals)
Juan Uribe (Dodgers)

Rivero is stretched defensively at short (though he appeared in 17 games there last year), but he's 24 and still has promise as a hitter (.303, 10 HR in AAA at Syracuse). He'd give the organization another option at 3B next year if Asche isn't ready. To bring Rivero back and get him away from the Nats, I'd be willing to stomach the prospect of seeing him, Frandsen, or Young play the occasional game at short.

two phormers, nice.

Rivero, Frandsen and Young are all stretched defensively at 3B let alone SS.

"when you say that in general you can't correlate a particular pair of variables, but in some individual cases there is a clear correlation, it really sounds like in fact there's no reliable correlation to be found and you're just applying confirmation bias to the cases."

Yep. I don't disagree with you at all. But there may well be unknown statistical methods for identifying, with at least with some degree of success, the guys whose spring numbers are probably meaningless & the guys whose spring numbers are probably not meaningless.

Maybe I'm splitting hairs here. And I didn't mean to turn this into a prolonged debate on what spring stats mean or don't mean. I was just addressing Jack's question of why people care about what particular players are doing in spring training. The fact is, nearly all of us DO care at least a little bit. There must be a reason. So I'm telling you what my reason is: because, while I acknowledge what the research tells us about spring stats, I'm not entirely persuaded that it's quite as simple as: spring training stats don't matter.

"rolo: Except you still run into the problem that a 3-week sample even in the middle of a major-league season is not particularly predictive."


Jack, we're not disagreeing. I was merely speculating as to what b-r might be up to. Personally, I believe it's a Fool's Errand. I've posted a link a number of times to an article about meaningful sample sizes, and ST just isn't long enough.

Jermaine Mitchell is getting another start. Maybe the team is considering him after all.

BAP--

I did not mean to suggest that the only reason for a players increased slugging would be PEDs. I am sure plenty of non PED using players have have sustained a ST surge in slugging over a season (although your mention of Ben Francisco does not help your argument). I was just attacking the predictive quality of that stat.

Hypothetically, lets say the study examined the years 1995-2005 (prime PED years). During this period there were 100 players who had a spring slugging % above their career numbers. Of this 100, 90 maintained the surge though the season.

Continuing the hypo...if the study then examined the years 1975-1985 and found 100 players who outperformed their career slugging in ST. Of that 100, only 25 sustained the increase over the course of the season.

In the first group the success rate was 90%. That is very predictive.

In the second set, the success rate is only 25%. Not so predictive.

What is the difference between the 2 groups? The first group likely had many PED users in it.

That is what I meant by suggesting the study may be flawed. Of course I am arguing hypotheticals, as I don't know what sample the study was based on. It's possible the study looked at the last 100 years of baseball. I don't know.

My main point is that statistics can be easily manipulated or fail to account for all variables (PED use). Which brings into question their validity.

"I don't think anybody thinks they should trade Mayberry."

MG said it's the one move he hopes they make before the season. Trade Mayberry and another piece for rotation help.

"I've seen a lot of lousy spring trainings turn into lousy seasons -- or, at the very least, lousy starts to the season. Mayberry, Wigginton, and Halladay last year; Ibanez in 2010 (lousy start); Moyer in 2009; Kendrick in 2008; Eaton every year he was here. I know that, for every bad spring training-turned bad season, you can point me to a bad spring training-turned good season. But at the same time, I strongly suspect that, within the generality that "spring training stats are meaningless," there are some very specific, but poorly understood, areas in which they are NOT meaningless."

This is really something. You can't find any definite correlation, and use a few random examples that don't prove anything, but figure there must be some reason that exists because if there isn't, then why would anyone worry?

I rarely agree with Jack, but he's right when he says fans just want to make these games mean something. Some fans just want to get excited or worry about things because if they don't, then what's the point of watching games that don't count?

fumphis' analysis of the latte example isn't actually fair to what bap is saying though. All stats start out as random, seemingly meaningless relations between things. But if there is a potential mechanism for why some types of lattes may be giving people cancer, while others are not, or for why certain people are susceptible to a cancer-causing agent in lattes that others aren't susceptible to, then more research could eventually uncover that connection, even though right now it just looks like random chance deciding who gets cancer and who doesn't among latte drinkers.

By that token, stats folks may eventually be able to tell us that, in certain scenarios, with certain qualifications, spring stats DO predict regular season performance. We don't know them yet, but bap is recognizing that lack of knowledge on that front now doesn't mean that spring stats are definitely, 100% meaningless.

Timr, that's true and a useful qualifier, but right now we and the stats industry are definitively in the pre-theoretical stage regarding spring stats. I don't think anyone would say the stats are "definitely, 100% meaningless," but someone needs to actually run some regressions using the "certain qualifications" you refer to if we're going to break out of the essentially useless anecdote-level analysis that the latte hypothetical and BAP's earlier post represent.

Iceman, all I need to know about ST is how the team operates when evaluating players.

Don't Charlie and some of his staff sit in chairs outside of the dugout closer to the batter's box? Isn't that to get a better and closer look at what each batter is doing?

It seems to me that the team has a wholly different way of evaluating players in ST than using stats, but that's just the impression I get.


Completely different topic:

Nik brought up the fact that Mitchell is getting another look today.

Can Mitchell possibly make Nix redundant where he could be moved as part of a deal? I was looking at his MiL splits, and while he had a bad year against RHP in 2012, he seemed to do pretty well against it in 2011.

"Can Mitchell possibly make Nix redundant where he could be moved as part of a deal."

I'm probably Beerleaguer's leading champion of older prospects, and I'm no great fan of Laynce Nix, but the answer to your question is a resounding NO.

Mitchell doesn't make any roster-lock "redundant." Mitchell seems to be in competition with Ruf and Inciarte for the 5th OF spot.

At the start of spring training, that would have been a ridiculous suggestion. Today, it's valid. And, frankly, it makes a bit of sense. Inciarte's defense may be ML ready, but it's hard to argue his bat is. Ruf's bat may be ML ready, but it's pretty clear his defense isn't.

Mitchell may do neither great, but he may be the Goldilocks choice.

I feel that ST stats are not predictive of what the regular season will be showing; however, I'd rather see a player be hitting well in ST as he heads north than one who is having a hard time in ST. Hopefully Dom's hotness continues in Philly, while Utley's coldness gets left behind in Florida,

Lake Fred sums it up for me. I've come to understand not to get worked up over ST results. But I'm not "stat hardened" enough to pretend I'm not enthusiastic about how Dom is playing, nor to pretend I don't have some concern over Utley and Halladay.

I put Halladay in a different category. When we know he had arm troubles last year that led him to only throw 88 MPH, and then he comes out this year only throwing 87 MPH, then yes, that is a real concern. If we saw Ryan Howard visibly limping walking back to the dugout after a strikeout, that would be the concern--not the strikeout.

Agree with Bob about how to look at these spring #s and overall appearance of guys. Brown being hot is definitely a positive no matter who the competition is. This needs to be the year he takes off and it's better to see him doing what he's doing now than falling flat on his face this spring. As for Doc, I'm concerned. Anyone who isn't is foolish. As much as I wanted to not believe it the guy is human.

NEPP: "Rivero, Frandsen and Young are all stretched defensively at 3B let alone SS."

So is Yuni. At least those guys are useful in other aspects of the game.

Before sitting the final 3 games, Rollins started 152 of 159 games last year. Is it worth carrying Yuni for approximately one start per month? I'm guessing the defensive difference in 7 starts between Yuni and any of those guys might be, what, a run over the course of the season?

If J-Roll goes on the DL, Galvis is up and starting. If J-Roll is out for a few days, you can live with poor defense at SS for a bit, juggle the roster to get Galvis up, or cut Rivero.

What's the deal with Lohse? How much longer is he going to wait to sign?

Halladay says he's lost 10lbs in 2 days.

That do wonders for arm strength?

Lohse is probably waiting for a team that thinks it can contend have a potentially serious injury problem with one of its top starting pitchers.

Hmm. Wonder if anyone knows of a team like that?

Zolecki:

"Halladay said if he is not strong enough to throw a bullpen Wednesday, the possibility exists he could be pushed back a day or two at the beginning of the season. In one scenario, right-hander Kyle Kendrick could start the second game of the season and Halladay could start the home opener against the Royals at Citizens Bank Park on April 6."

"Halladay says he's lost 10lbs in 2 days.

That do wonders for arm strength?"

If you don't catch a stomach virus early, it can easily spread and become an arm virus.

Halladay should start whenever he is ready. The Phillies can steady the ship while he gets completely up to game shape. It is April after all. A month when typically pitchers are ahead of hitters.

I remember last year when all the Phillies had to do was tread water until Howard and Utley came back.

"Halladay says he's lost 10lbs in 2 days."

Does Roy have the same weight loss clauses in his contract as D Young???

Who knew Roy would be so hurting for some extra cash?

That was before Gio and Jordan Zimmerman turned into better pitchers than Cliff Lee and Roy Halladay.

Anybody able to actually watch today's game on MLB.TV? I caught the first half-inning, and KK looked good, but lost it shortly thereafter.

He still getting weak contact overall?

Michael Young makes a tough play down the line and makes a strong throw to first... but it's high and just over Howard's outstretched arm.

I've actually been impressed recently for the way Young has handled balls hit to him. This is the first throwing error I remember in awhile.

Yes, KK looking good.

I'll go ahead and say I'll be very surprised if Halladay starts in the first week of the season.

This is how it starts. They float out that he "may" miss the first series and just be "pushed back" to the home opener series. Then, though you hear he "looked great in a bullpen session," he gets pushed back another start, because they just "want to be careful to not rush him back."

I don't have a good feeling about this one, I'll be honest.

Defensively as well. Just turned a DP on a smoking line drive right at him.

Philibuster: Mostly weak contact. Just had a line drive up the middle that Kendrick snagged and turned into a double play.

Other than that, no hard contact yet against KK in 3 innings.

Jack: I think you're overreacting. I hear he looked great in a bullpen session.

Nice, maybe he'll bring that Spring ERA down a few notches, to merely "not respectable" from his current "stratospheric."

I hear that may or may not be related to how well he does in the regular season.

"Mitchell may do neither great, but he may be the Goldilocks choice."

KAS, love it!!! Great analogy.

Only on BL.

rolo: "To amplify my previous point: I'm more worried about Roy Halladay than Cook or Lopez."

Hey rolo, this link is just for you

http://mbacookie.files.wordpress.com/2009/10/panic-button.jpg

Jack: A stomach bug != arm falling off.

He was spotted losing his breakfast after the first inning on his last start, and he's supposedly lost a lot of weight (I assume via heaving it out of his guts) in the past couple of days.

When you've got an illness that severe, you're not going to work back into "game-ready" shape in only a couple of days. The idea of him not starting the 2nd game of the season was - I assume - because he's expected to get a ST start too close to opening day to be able to feasibly start again in game #2.

Truth be told, if Halladay's not right, they could just DL him for the first 15 days and only have to use a fill-in (Cook, Cloyd, etc.) for a single start (4/6 against KC) before activating Halladay for 4/16 against the Reds. That would also give them a couple of weeks with 11 pitchers, which gives them some time to sort through the roster crunch.

Not saying I'd rather see Doc on the DL, but if he's not 100% now, it might be a good option.

Man, KK doing it on both sides of the ball today.

CT, I think they'd be wise to just DL him and let him work in extended spring training.

lorecore- do you think Jordan Zimmerman is a better pitcher than Cliff Lee? And do you think Gio will see any regression from last season's improvements, or will he repeat last year's success?

Don't know if anyone mentioned this....

http://www.philly.com/philly/blogs/phillies/Bulletin-board-material-for-Phillies.html

"But yesterday a new item appeared [on the Phillies clubhouse bulletin board]. It is a photocopy of a graphic from a publication -- it looks like Baseball America, although I'm not positive -- that breaks down the major league batting averages on balls in play on swings in certain counts. The graphic illustrates what every hitting coach preaches: get the count in your favor, and you've got a better chance of getting a pitch that you can turn into a pitch. The moral: work the count.

I don't know who posted the photocopy, but I don't remember anything like it ever being posted before.

Just an interesting little tidbit for a team that has been criticized for its approach at the plate."

The times they are a'changin'.

Isn't Sarge a little old for the backward hat look?

Yankees get their second base runner... each time on a throwing error to first. (The second one was on a bunt that likely would have been a base hit.)

lorecore- to follow up, I'll just say what I'm getting at: I would take Lee over Zimmerman and Hamels over Gio this year, though the Hamels/Gio thing is close. I think it's the difference between Halladay/Strasburg that will tell who has the better top of the rotation.

Yankees get their first real hit of the game when Ichiro doubles into the LCF gap. Pitch got too much of the plate and Ichiro did what Ichiro does.

1-0 Yankees.

Every once in a while this spring, my MLB.tv has dropped the audio.

I am not complaining.

It generally only happens to me at work, but I've lost the video more than once.

Can't really blame MLB because I don't know if it's just everybody else here chewing through the bandwidth (and/or my terrible workputer).

I have to watch in silence here anyway, so no TMac, Wheels or Sarge for me.

Back to back K's to end the 4th inning. Kyle pitching well.

DOM.

BROWN.

But spring training doesn't matter...

DOMONATED

Think we should go back to the ol' "Four Aces" routine? 33, 34, 35, 38 has a nicer ring to it than 33-35 and 44.

Inside pitch, down this time, quick swing, dropped the bat on the ball and drilled it over the wall.

Dom Brown just looks like a different player.

He's just trying to reach BAP's 80 AB, 9 HR benchmark.

DOM plz save some for the reg season.

DOM joins Mike Morse with 6 ST HR. Morse has 6 HR in 35 AB.

I should probably note that I'm not predicting a bad season for Dom Brown. In fact, quite the opposite.

As many people know, I've long been one of Dom's biggest supporters on here, even while others were calling him a bust at some point in the last couple seasons.

I think he'll have a pretty good season this year--but it has very little to do with the fact that he's hot right now in mid-March.

I also think it's funny posts like KAS ironically saying "But spring training doesn't matter," as if that home run somehow proved that it does.

Nix and Lerud. Two names we knew we'd be talking about this year way back in 2011.

6 HR in 62 spring ABs for Dom Brown so far.

Clout's mentioning Ron "Palm-trees" Stone brought back fond memories. Now there was a genuine character. I recall one spring training where he went up in the stands with the fans and started booing his fellow players.
Cracked me up.

Another meaningless HR by Dom.

Was Brown's ball a bomb?

Chamberlain rocking an awesome mustache.

NEPP and KAS: I sense a pretty humorous logical fail going on.

Your argument is (reducing it a lot) that a lot of HRs in Spring = a lot of HRs in the regular season.

Your argument isn't proven right the more Dom fulfills the first part of the equation. The argument is only proven right if he goes on to fulfill the second part of the equation.

No one is arguing that Dom is having a bad spring. Everyone agrees it's fantastic. The question is how much that means to the regular season. Until we get to the regular season, nothing can possibly be proven with this.

I wasn't arguing anything of the sort. I merely mentioned that Brown hit a HR in a ST game which is meaningless.

Why read anything into a statement of fact like that?

Since Gameday ST edition uses the ridiculous assumption of least number of pitches necessary to get the result indicated, does anybody know how many KK has tossed through ~5 IP now?

Ugh, strikeout for Brown that AB. He's probably going to suck this year.

Kid Kendrick is dealin'.

Phillibuster he just finished 6th with 58 pitches

By the way, stick in Revere for Mitchell, and Kratz for Lerud, and make Kratz the 7th hitter and Revere the 8th hitter, and I think this is your opening day lineup (if it's against a righty pitcher).

Mike: Thanks! Yeah, he's seriously dealing - although the Yankees are obviously a bit injury-depleted.

Gameday has the DOM highlight. Another big one. I stick by my estimation of ~420ft average hr distance this spring.

Why is it so hard to listen to a ST game? My XM Internet radio is garbage and it's not on local radio. I have about had it with local adio talking football like it's still football season.

CS: Get AtBat for your smart phone. You get home/away radio access to every game, no blackouts.

$15 for the season's a pretty good price too.

Jack: It was a joke.

If there was ever a reason for the Yankees to lift their facial hair ban, it's Kevin Youkilis' face.

Adams continues to absolutely mow down the competition. He's another one where yeah, it does matter what he looks like in spring. Because it was a question as to whether or not he'd even be ready for opening day.

Especially important seeing him get Ichiro and Hafner out, as one of his big problems last year was getting lefties out.

Ryan Howard with another bomb.

It's nice seeing that swing. We didn't see it much last year.

With the release of the Evil Dead remake, Howard felt compelled to bring out his boomstick.

Howard HR off a (garbage) lefty.

The comments to this entry are closed.

EST. 2005

Top Stories

HardballTalk

Rotoworld News

Follow on Twitter

Follow on Facebook

Contact Weitzel

CSG