# FanGraphs Baseball

RSS feed for comments on this post.

1. The point at which a stat “stabilizes” is the point at which it might be telling us anything at all.

I see this over and over again on FG, and this is simply incorrect and a vast misstatement of the original research.

The point at which a stat “stabilizes” is the point at which the sample is sufficiently big that you can predict future performance by equally weighting sample performance and league average; no more and no less.

“Before” a stat stabilizes, you weight league average more (“regress more”) to account for the greater variance in a smaller sample, and you have wider error bars/less precision in your prediction; “after” a stat stabilizes, you weight league average less and have more confidence. But that’s a quantitative difference, not a qualitative difference.

Even a small sample size does have some meaning; otherwise, you couldn’t add up a whole bunch of 0s from individual data points and get non-zero meaning from the sum, or have a discontinuity from “no meaning” at 199 PA to “meaning” at 200.

So you can predict, even at this point, that Perez’s BABIP will be higher than league average — just not as much higher than league average as you could/should predict if you had a much bigger sample. [I’m ignoring his minor league BABIP — if you add that to your data set, you might well get a different answer, but that’s not relevant here.]

So the fact that the data you have about Perez says far more about his Swing% than his BABIP doesn’t mean that it says nothing about his BABIP, let alone that it says “‘less’ than nothing” about it.

Comment by Chris — July 6, 2012 @ 2:28 pm

2. Thank for the comment.

I may have been hyperbolic in some of my originally statements above (I’ve toned it down a bit to hopefully lessen confusion), but as for your main point about regression/prediction/stability, I asked about the distinction between the point of “stability” and the point of where we weight league average for regression. Pizza Cutter (the original researcher) and Tango have different approaches, but unless I’m misreading them, they both seem to make a distinction (although in different ways) between the point at which a stat “stabilizes” in Pizza Cutter’s sense and the number of PA at which one weights observed individual PA equally against the same number of league average PA of league average. See their responses to my question along those lines (a long time ago) here:

http://www.fangraphs.com/blogs/index.php/when-samples-become-reliable/#comment-76819

I’m not saying that I’ve necessarily got things exactly right here, but they are both making a distinction between Pizza’s point about “stability” and Tango’s way of using regression to the mean.

Comment by Matt Klaassen — July 6, 2012 @ 3:10 pm

3. Wait. Who wanted Aviles gone for Giavotella to get more time and was ready to give up on Moustakas? I think that Neyer’d tag belongs a few lines lower.

Comment by Gadwin — July 6, 2012 @ 3:40 pm

4. Let’s see if I can clear this up.

The original research looked tried to identify at what number of PA various stats crossed a reliability of .70 (using a split-half correlation method). The idea is that at r = .70, r-squared is .50 (I know, I rounded.) At that point, we conclude that more than half the variance is endogenous to the player himself. That was my definition of stable. Feel free to disagree with it. You also correctly point out that like any dichotomy, there’s a lot lost in the middle.

Tom Tango often likes to see the point where reliability hits .50, because this is the point where he projects at half player performance and half league average. It works for his purposes.

The issue of whether a player’s performance after X PA is predictive of future results is one that I often see mentioned. I plan to write something on this in the near future, but I think that saying that once a stat “stablizes” that it is necessarily predictive is a logical error.

Comment by Pizza Cutter — July 6, 2012 @ 3:46 pm

5. Essentially, Pizza is saying that “stability” has to do with measuring a *change* in baseline — if you see an uptick in Swing% over a small sample, it may suggest that the player’s actual skillset has changed.

Tango is saying that “stability” is just measuring the current skill — once you reach a “stability point,” you can statistically expect that X% of the player’s deviation from average is repeatable. It has nothing to do with change.

In this post, it’s a distinction without a difference. If you’re starting from scratch, then the “change” is from your expected baseline, which means both definitions wind up with exactly the same outcome.

The only time you would really distinguish between the two is if you were looking for evidence that a player’s skill has actually changed — at which point you have a very complex problem ahead of you in determining how much to “age” old data so that it carries less weight.

That’s the crux of Tango’s issue with Pizza — he thinks that Pizza is asserting that a lower “stability point” indicates a skill that actually changes more rapidly than a skill with a higher “stability point,” but that the research doesn’t actually show that — instead, the research just shows that some skills are easier to identify with a smaller data set than others without giving any indication as to which skills might quickly change (let alone how you would age old data to properly incorporate it).

And I agree. I think it’s entirely possible that some skills do change more quickly than others, but I don’t see any evidence to support that. Instead, I see good research on how big of a data set we need before we can make a specific prediction about a specific skill, and how that varies based on the skill. Which is still meaningful but very different.

But in either case, neither Tango nor Pizza is saying that “stability point” is a discontinuity where your sample suddenly goes from meaningless to meaningful. It’s just crossing a line and reaching a certain “amount” of meaning.

Comment by Chris — July 6, 2012 @ 3:51 pm

6. Or I could have waited for Pizza to clear it up himself. Thanks!

Comment by Chris — July 6, 2012 @ 4:05 pm

7. And I may well have Pizza’s position on what “stability” means (i.e., I might be doing exactly the same thing that I just accused the author of doing in attributing more to Pizza’s work than he said, only disagreeing with the position rather than accepting it), but I will let him correct me himself…

Comment by Chris — July 6, 2012 @ 4:08 pm

8. LOL at bringing up Aviles’ numbers as if they mean anything. Aviles was 27 as a rookie.

Sal Perez just turned 22.

Comment by Marc — July 6, 2012 @ 4:16 pm

9. 40 ABs notwithstanding, Perez has more home runs (3) than strikeouts (2).

Again, I know it’s a SSS, but that is just extraordinary and unheard of in any kind of sample size.

Comment by Marc — July 6, 2012 @ 4:19 pm

10. And?

Comment by maguro — July 6, 2012 @ 4:19 pm

11. And it’s irrelevant to the discussion. Comparing a 21 year-old rookie’s numbers to a 27 year-old rookie’s numbers is pointless. A 21-year old’s numbers (even in a small sample) is a hell of a lot more predictive when it comes to future production than a 27-year old’s numbers. I thought this was common sense.

Comment by Dale — July 6, 2012 @ 4:31 pm

12. Count on Fangraphs to overanalyze what has been a thoroughly impressive start to his Major League career and reach the conclusion that Salvador Perez, the best young catcher in baseball, is the next Mike Aviles.

Comment by Tom — July 6, 2012 @ 4:32 pm

13. Yeah, and it’s kind of surprising to keep seeing this type of thing since for a while now the cultural leader of the site, Dave Cameron, has been provocative in the other direction regarding outliers.

No disrespect to the statisticians, but when it comes to outliers like Sal Perez, simply plugging them into a “regressor” or “de-lucker” is just boring as hell. Have somebody like Mike Newman come in and say, “The sabrists have ID’d so-and-so as an outlier. Here’s why he’s not, or here’s why he’s for real (within reason, because we all know he’s not going to hit .350 this year).

I’ve watched quite a bit of Perez and I have to say that his minor league scouting report was mostly rubbish, save Greg Schaum. When he first came up he looked like he had a decent idea at the plate, but was clearly a defense first guy. Within a week Seizter had made a couple adjustments and he just took off. He has exceptional hand-eye coordination, is damn smart, really instinctive, and is a sponge. He may not be the next big thing, but I wouldn’t rule it out either.

One other thing about the numbers above. Mike Moustakas has the worst first 200 ABs of any MLB player I’ve ever seen. But once he got it, he looked like he’d been there for ten years. He’s easily their best hitter now, has great AB’s, is really mature, a natural leader, and like Perez incredibly instinctive. None of that came through in the first 200 AB’s because he looked like he didn’t even have a trigger. Point being that when we’re talking about young players in their first exposure to MLB, not only do we have SSS issues, we have huge bias issues relating to org, makeup, etc. There are definite limits to statistical analysis under these circumstances.

Comment by Paul — July 6, 2012 @ 5:18 pm

14. “In other words, from his first 198 plate appearances, going by the numbers, we may have learned a tiny bit about Perez’s swing rate, contact rate, strikeout rate, and maybe his walk rate. On the other hand, his major league numbers basically tell us practically nothing, so far, about his power, for example.”

In those 198 PAs, he has a .176 ISO and a 12.8% HR/FB rate. This is somewhat higher than one would have expected from his minor league record. To do that, plus to maintain a high contact rate and a high line-drive rate (27%) and a low IFFB rate (6.4%) is surely telling us more than “practically nothing”. BABIP doesn’t come close to capturing what we do know about what he has done so far.

ZIPS gets this. It has projected him now at .284/.312/.403 for the rest of the year, which isn’t much different from what he hit in double A last year.

Comment by Mike Green — July 6, 2012 @ 5:32 pm

15. I really wonder what kind of scouting grade his hit tool would get right now. Anywhere close to supporting a career O-Swing of around in the low 40s (against the MLB AVG of about 29-31%), with O-Contact in the low 80s (against the MLB AVG of about 65-68%)? Perhaps not surprisingly, given the no walks and 2 Ks, those numbers are actually more extreme this season so far.

Seems like he might be inviting the Josh Hamilton pitching approach. If so, how Perez responds will be telling.

Comment by ralph — July 6, 2012 @ 5:37 pm

16. “Perez has more home runs than strikeouts… that is just extraordinary and unheard of in any kind of sample size”

Grandal had 3 homers and 2 ks through 14 at bats to start his MLB career.

Hyperbole Fail.

Comment by Matt NW — July 6, 2012 @ 5:41 pm

17. wait a minute… hyperbole win!

Comment by Matt NW — July 6, 2012 @ 5:42 pm

18. Are you suggesting that 200 at bat samples are meaningful for 21 year olds but not meaningful for 27 year olds? Otherwise, I don’t understand your point because the article doesn’t say Sal Perez = Mike Aviles, it just says that you can’t draw conclusions from 200 at bats.

Comment by maguro — July 6, 2012 @ 5:48 pm

19. And Moustakas was 22 last year and apparently 200 ABs wasn’t predictive for him, but we’re supposed to accept that Perez’s 200 ABs are predictive because he’s 21? Why?

Comment by maguro — July 6, 2012 @ 5:51 pm

20. Perez is not the best young catcher in baseball, unless you demand such a strict standard of youth that no one else qualifies. Wieters and Santana are 26; Posey and Avila are 25; Ramos, Mesoraco, Grandal, Norris, and Rosario are 24 or younger. Maybe you wouldn’t take every one of those players over Perez, but you’d take most of them.

Comment by byron — July 6, 2012 @ 6:06 pm

21. Okay, fine, more HR than K over 25+ AB. Unheard of.

Comment by Marc — July 6, 2012 @ 7:02 pm

22. Come on boss, of course not mad, just giving you a hard time because what you said is so obviously untrue. It took me about 15 seconds to pull up an Albert Pujols game log, and see a 64 plate appearance stretch when he hit 10 homers and struck out 7 times.

If your point was Sal Perez is playing very well, point taken, he is. But to suggest this homers greater than walks is unheard of, that’s just silly… find a power hitter who makes good contact, and you’ll find example after example.

Comment by Matt NW — July 6, 2012 @ 7:10 pm

23. Alright, I smell you dawg…but look at the one dude you found…Albert Pujols aka the best hitter in the last 50 years. That should clue you in as to how incredible Sal Perez has been thus far.

Here’s your next challenge, partner…find me a dude who dropped more taters than strikeouts over a 40 AB stretch at 22 years old or younger.

You won’t find any.

Sal Perez is a 245 pound hombre, 100% USDA man, plus-plus hit tool at a premium position. He is going to be a superstar. Leave the dodes at Baseball Failmerica and Baseball Amateurspectus to leave him off of their top 100s like they always do.

Comment by Marc — July 6, 2012 @ 8:09 pm

24. A.) I didn’t realize Sal Perez was touted as the next Joe Mauer or the next Buster Posey.

B.) Aviles was on a 4A track who made good while Moustakas was one of the Royals top prospect; any struggles by Moustakas would be shrugged off as he is sort of part of the future.

Comment by Larry — July 6, 2012 @ 8:23 pm

25. Sal Perez isn’t Mauer…

…He’s Mauer with Power.

Comment by Marc — July 6, 2012 @ 8:43 pm

26. Well to me, “young” implies that there is growth. Most research has shown that a prime is basically 25-29 with a peak at 27. So really, 25 isn’t “young” in baseball terms. That’s the beginning of your prime, probably will get better and will likely have better years at some point.

Wieters and Santana are in their prime. Posey and Avila are just now getting into it. My personal cut off is 24. All depends on how good you are though. If you are as good at 20 as someone else is at 23, you are both “young” and equally “good” but because one is younger, the implication is that they’ll end up better.

I hope that made sense. Probably didn’t.

Comment by Antonio Bananas — July 6, 2012 @ 8:54 pm

27. think about this, at one point in time, Soto was a “young” (25) catch who had just won the rookie of the year award. Had a few other years with about that sam production, and has dropped off. 25 isn’t really “young”. Unless you mean like, normal first world human age.

Comment by Antonio Bananas — July 6, 2012 @ 8:55 pm

28. I don’t know that there’s a properly “right” position on this. The main question that I had was “What’s the reliability of this stat at X PA?” Whether you want to draw lines in the sand on it (and where) is a matter of what you need it to do.

Comment by Pizza Cutter — July 6, 2012 @ 8:59 pm

29. Might make for an interesting empirical study.

Comment by Pizza Cutter — July 6, 2012 @ 8:59 pm

30. I named 5 catchers who are 24 or younger.

Comment by byron — July 7, 2012 @ 1:31 am

31. “Now, those things is neither very exciting nor surprising if you looked at Perez’s minor league numbers or read what people were saying about him as a prospect.”

Wait, what about his minor league numbers? And what were people saying about him as a prospect pre-2011?

Comment by Bob Loblaw — July 7, 2012 @ 12:21 pm

32. Actually, Pujols was not “the one dude I found”, he was the first and only guy I looked up (obviously, I picked a guy who would likely meet the criteria). And for your most recent challenge, I also IMMEDIATELY met your criteria, first guy I looked up, went straight to his 21 year old season… and again, I went with a very good hitter who didn’t strike out a lot and hit homers: Griffey Jr.

In mid-late May of 1990, he did it. Took me 90 seconds.

So again, I’m not saying Perez isn’t good. I’m saying you’re wrong to say his hot start is without precedent… as I am easily finding precedents.

Enjoy

Comment by Matt NW — July 7, 2012 @ 4:15 pm

33. It’s far from unheard of. Ted Williams in 1941 (and 1948, 1950, 1953… ). Rogers Hornsby had a K/HR of 1 in 1925. Doing 1.0 and lower K/HR and we see Lefty O’Doul (Baker Bowl effect?), George Sisler, Mickey Cochrane, Lou Boudreau….

And probably the hardest guy to strike out, ever, Joe Sewell.

Comment by Nathan Nathan — July 7, 2012 @ 11:25 pm

34. I don’t believe I ever expressly said that, although it sounds like something that I would say if I had thought of it.

When I wrote the original article, nearly 5(!) years ago, I was actually most interested in an empirical way to determine what my cutoff point should be when I said “min X PA” in research. The issue of whether this models change is something that is something that people (including me) have probably inadvertently assumed.

More to come.

Comment by Pizza Cutter — July 8, 2012 @ 4:22 pm

35. Luke Scott had 5 HR and 4 K in 34 at-bats during September 2006… I just randomly picked 2 Oriole hitters, the first being Nick Markakis, who struck out a lot more than I thought he did when he was hitting for power.

Comment by David — July 8, 2012 @ 4:39 pm

36. As recently as August 2011 (i.e. immediately after he got his first callup), this was John Sickels about Perez:

“A 6-3, 230-pound right-handed hitter, Perez is now 21 years old. He has a consistent swing with few mechanical flaws, makes contact, and flashes above average power. However, he seldom draws walks, inhibiting his on-base percentage, and he is one of the slowest runners in professional baseball. He has been young for his leagues, and while his numbers haven’t been exceptional since leaving rookie ball, he should be at least adequate with the bat eventually and should provide occasional power.”

And of course his ML numbers are surprising looking at his minor league history. Right now, his career batting line is .344/.366/.516. All 3 of those stats are higher than *any* single season of Perez’s minor league career, except for a tiny sample of 46 at-bats in 2008 for one of KC’s Rookie League affiliates (a level he was repeating, and would repeat again the following year).

Comment by David — July 8, 2012 @ 4:42 pm

1. Perez has had a very good start to his career.
2. Aviles also had a very good start to his career, and is useful but not a star.
3. Moustakas had a poor start, but is now good.

Context plays an important part in regards to the above though. Aviles was older. No one expected him to maintain his torrid initial pace. In fact, his career has turned out probably how most sensible people would have thought.

Moustakas was on a different level of hype from the other two. While fans may have worried, a top prospect of his ilk is going to be given more than 200 PA to prove himself.

Most people probably don’t expect Perez to maintain this level. Still it’s an encouraging sign from a very young player at a premium defensive position. Even if (and when) he regresses, he’ll still provide value.

Perez was a sleeper prospect who performed well in an initial sample. Comparing him to Moustakas (much more hype) or Aviles (much older prospect) is completely irrelevant to how his career turns out.

Comment by jmarsh — July 8, 2012 @ 4:54 pm

38. @Dale, excellent job just pulling a random statement with no support from that hole in your hind. Well done Sir.

Comment by Colin — July 8, 2012 @ 7:02 pm

39. “Perez is not the best young catcher in baseball, unless you demand such a strict standard of youth that no one else qualifies. Wieters and Santana are 26; Posey and Avila are 25; Ramos, Mesoraco, Grandal, Norris, and Rosario are 24 or younger. Maybe you wouldn’t take every one of those players over Perez, but you’d take most of them.”

“I named 5 catchers who are 24 or younger.”

And I’ll take Salvador Perez over all five. Ramos has suffered a serious knee injury, and there’s no telling whether or not his pre-injury career trajectory will be re-established once healthy. Mesoraco likely has a solid future but he’s two years older than Perez and has had abysmal results throughout his first 185 PAs in the majors while playing below-average defense. Grandal, Rosario and Norris are excellent hitters but are not nearly as advanced defensively as Perez is. So, yes, you managed to list some fine young catchers, but you failed to make a complelling argument (or even an argument, for that matter) as to why you feel any of them deserve to be vaulted ahead of Perez for title of “best young catcher in the majors.”

Comment by Tom — July 8, 2012 @ 8:46 pm

40. Having an OBP and BA the same is always the sign of a quality player.

Comment by shthar — July 8, 2012 @ 9:23 pm

41. its unheard of for a hitter to have more home runs than strikeouts in any kind of sample size? what is it with the hyperbole on FG lately?

Comment by jim — July 9, 2012 @ 12:59 am

42. …and no walks.

Comment by shthar — July 9, 2012 @ 1:27 pm

43. One thing it suggests is a young player with young player skills who has the chance to grow as a hitter as this player is only 21. In a way for a young player it is a good thing to be succeeding with very few walks. If you come to the big leagues with fully developed plate discipline It’s hard to improve upon. George Brett walked 21 times in 486 plate apps as a 21 year oldin 1974. He walked 103 times in 665 Plate apps in 1985 at age 32. So it’s not all bad.

Comment by whatever — July 9, 2012 @ 6:19 pm

44. One interesting thing about such extreme contact as Perez is showing thus far this year (2 Ks in 47 PA to go along with the 0 BBs) is that by adding in power like he has so far, his AVG will be higher than his BABIP (right now his .383 AVG is supported by a not entirely unreasonable .341 BABIP). I highly doubt he’ll keep up such an extreme contact rate and he obviously won’t keep up his 4 HR/47 PA rate. However, if he ends up being one of the rare BABIP wizards, he could still maintain a respectable OBP without ever taking many walks.

Or he could just improve, as has been known to happen from time to time for players who get significant MLB ABs at such a young age.

Comment by ralph — July 9, 2012 @ 7:51 pm

Line and paragraph breaks automatic, e-mail address never displayed, HTML allowed: `<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong> `