I, for one, blame Leo Mazzone.

The "problem" is that he had 271 PA's of great hitting in the majors and 356 terrible PA's (MLE's) in the minors and did not project to be even an average major league corner outfielder in 06. That is why we use projection algorithms and that is why we use minor league MLE's and that is why we use regression toward the mean as part of the projection algorithms. That is the "problem..."

DiSar is still inexplicably loved in some parts out west...

mgl, I don't know if you read the linked entry but he presents his hypothesis for Francoeur's shockingly good performance last year. He's claiming there's a window for a prospect/rookie after he's called up because minor league scouts look for different things than advance scouts. I'm not sure I buy it, but the "problem" he's talking about is the time it takes for advance scouts to figure out a hitter. Once all teams have a scouting report, a player benefiting from lack of advance scouting would presumably return to their minor league MLE performance. Perhaps conveniently, his hypothesis would only explain overperformance by toolsy, undisciplined minor leaguers who get get called up unexpectedly and immediately play full-time.

I'm a Braves fan and I've probably watched the majority of Francoeur's at-bats on television. For what it's worth, here's my take. Francoeur had a lot of success when he first came up because he hits low fastballs really well. When a rookie debuts in the league, teams generally first see if he can hit the fastball, and most of today's pitchers prefer to work the low end of the strike zone. Hence, when Francoeur first came up, teams were pitching to his strength. Once they figured out that he couldn't hit the breaking pitch, and that he would swing at just about anything that he thought he could get his bat on, the party was over. I don't think he can learn patience because he quite simply can't tell whether or not most pitches are going to be strikes or not. More specifically, he acts like he can't see the spin on breaking pitches. I saw that even in a few of his minor-league at-bats last year. He was facing a guy who had a good "12-6" curveball and he kept swinging over the top of it. It was apparent that he didn't recognize it.

No I hadn't read the article, but I doubt that his hypothesis or Chuck's is true, but I am certainly not sure by any means. Since all you hear from announcers is that a all young players can hit the fastball but they typically have trouble with the breaking stuff, it would be odd (if not insane) if teams "generally first see if they can hit the fastball" as Chuck suggests. I mean if young, inexperienced players typically can hit fastballs but not breaking pitches (and I don't know if that conventional wisdom is even true), why not throw lots of breaking pitches to most if not all rookies?

My guess is that Francouer was simply lucky in those 271 PA and that people as usual are trying to make some "sense" out of randomness. I highly doubt that any particular subset of rookies has early success and then a large dropoff. But I am open minded on this. I welcome anyone to provide some real evidence that any paricular type of rookie shows an atypical pattern of performance in his first/second year. All I am hearing/seeing herein and in this article is speculation and conjecture.

In fact, after reading this article, his whole permise is silly and misguided. The fatal error is assuming that his performance in those 271 PA means anything at all - that poor performance following good performance means that "pitchers have figured something out." Maybe they have and maybe they haven't, but without literally looking at each PA on video there is simply no way of knowing how much of any player's good or bad performance in 271 PA is good or bad luck.

This particualr comment is ludicrous:

Scouts saw his poor plate discipline and just reported, “he’s not ready yet, fire it in there.”

If scouts "saw (or knew about) his poor plate discipline (which I have no doubt they did)," wouldn't they tell the pitchers the exact opposite, or, "Don't fire it right in there! Give him little to hit since he'll swing at anything anyway." Isn't the first rule of pitching that if a player swings at bad pitches, don't throw him a strike, and vice versa? I'm sorry, I don't buy any of this...

"Ludicrous?" "silly and misguided?" I guess I'm just an idiot and I clearly don't know anything about randomness. It's a good thing that BTF has some great referees to catch all the ludicrous stuff I put out there.

Francoeur did get lucky an awful lot last year. I remember one at bat where a popup behind the shortstop landed in between three defenders and resulted in a double.

I'm not sure what to make of Jeff's start, because on the one hand, how could a big league team not even talk to their Southern League affiliate about a player? I'm no scout, but I could tell anyone who asks about Jeff's trouble hitting pitches outside the zone. It seems hard to believe that professional scouts in AA, watching the game for whatever purpose, wouldn't notice that.

However, mgl, I don't necessarily agree with you either. Suppose that a team had no scouting report on Francoeur, just his AA line stats. I don't see instructing your pitcher to throw a recently-called-up minor leaguer a diet of breaking pitches out of the zone; the guy wasn't hitting that great in AA, so why put him on base when you can get him out if you challenge him? Of course, the assumption that teams had no scouting report is not a good one.

It's a good thing that BTF has some great referees

JCB,
"BBTF" doesn't *have* the commentators (referees) on your work. We provide an open forum.

Sorry it isn't all back-slapping.

I'm too lazy to google last year's projections that Francoeur was going to be great like Juan Gonzalez.

JCB : So you stand by the idea that scouts recommend throwing strikes to guys known to swing at balls?

G.W.O.: I'm not standing by anything. I pondered---notice the post wasn't titled "explaining"---if minor league scouts scout double-A scout players differently than major league advance scouts. I did not run any tests, I made conjecture that may or not be correct. The statement in question:

My guess is that Francoeur’s jump surprised everyone, and that scouts were not scouting him like advance scouts typically do. Instead, they focused on his raw ability and promise. Scouts saw his poor plate discipline and just reported, “he’s not ready yet, fire it in there.” (italics added)

In the last sentence I proposed a hypothetical one-line scouting report. I have no idea what scouts tell teams. I don't pretend to know. All I was trying to give an example of the type of scouting report a minor league scout might give. A throwaway line that I put about five seconds of thought into writing, yet it has somehow become the focus of stat-head-centric board. I don't see anything of substance in mgl's post that I did not address in my original post. Could this all be random chance? Absolutely, but I think there might be more to it. I might be wrong, but nothing I said should be characterized as ludicrous or silly. Misguided, perhaps.

If anyone has any more for me, please discuss it on my blog. I clearly bring out the worst in people here, so let's keep this stuff off of this board.

JCB:

My only comment is that I believe the Braves CAN send him down to Richmond without any long-term negative consequences. Jeff has to understand that right now he is overmatched. And nobody is better at communicating with players, especially young players, then Bobby Cox.

Which is the only thing that lends me pause in this whole episode. I hold Bobby Cox in very high esteem when it comes recognizing and developing talent. His record is almost without peer. He looked at Andy Marte and must have given his ok to trade him. He looked at JF for almost 300 plate appearances and decided he was ready to contribute in 2006. (I had thought Jeff was thrown into the breech but would be sent back down to begin this season.)

Of course, it is also possible that Bobby knows what we don't in that JF will "break out" at some point.

JCB vs mgl..battle of the crumudgeons!!!

That is why we use projection algorithms and that is why we use minor league MLE's and that is why we use regression toward the mean as part of the projection algorithms.

I haven't been to a game yet this year. I sure miss the crack of the regression toward the mean and the smell of projection algorithms.

Could this all be random chance? Absolutely, but I think there might be more to it. I might be wrong, but nothing I said should be characterized as ludicrous or silly. Misguided, perhaps.

I know, JCB. This is the kitchen, and that is heat. Check the interactions when I wrote about Glavine's first season with the Mets. I got the same comments.

If anyone has any more for me, please discuss it on my blog. I clearly bring out the worst in people here, so let's keep this stuff off of this board.

Gonna have to be tougher than that to make it as a researcher.

I highly doubt that any particular subset of rookies has early success and then a large dropoff.

Jerome Walton and Dwight Smith tip their caps...

I saw Francouer play yesterday and he looked overmatched, but then again Sheets was pitching. Beautiful day at Miller Park. I've got a sunburn.

And Bob Hamelin.

I highly doubt that any particular subset of rookies has early success and then a large dropoff.

Jerome Walton and Dwight Smith tip their caps...

And Bob Hamelin.



and Kevin Maas. And Joe Charoneau. Etc, etc, etc...

I don't think that's what MGL meant. By subset, I take him to mean a group of rookies who share certain charcteristics, age, hitting style, handedness, speed, tools...

If you can show how Smith, Walton and Hamelin were all truly similar, and identify others in that group, that would be one think.

But listing a group of random rookies who performed well and then, pffft, well, that's not evidence of anything.

Charboneau got hurt - I excused him.

I think Smith and Walton were pretty similar actually, but I'd have to look at the numbers. They did finish 1 and 2 in the ROY voting in '89 and their toolsy skill sets seem similar, but its been a long time since '89 and my memory may be lying.

Walton is a very good comp, IIRC. Smith is a LHB as was Hamelin (and he was big and slow).

Walton was a slightly built CF with good defense.

I pondered---notice the post wasn't titled "explaining"---if minor league scouts scout double-A scout players differently than major league advance scouts.

Scouts are all looking for the same things, regardless of level. The problem is that scouting reports for minor leaguers are based on incomplete data. Scouts don't see every player play every game - teams don't have that many scouts, they have to cover more than one team, and there isn't as much video of minor league games as there is of major league games, so you can't always go back and look at tape. You make judgments based on what you see, when you see it. A minor league scout might have seen Francoeur play 10-15 games, tops, against widely variable competition. There's much more room for error when evaluating a minor leaguer (just as there is when evaluating his statistics).

-- MWE

Walton is a very good comp, IIRC. Smith is a LHB as was Hamelin (and he was big and slow).

Walton was a slightly built CF with good defense.

Yes, but a counter-example to Walton would be Marquis Grissom.

A counter example to Smith would be Rusty Greer

A counter example to Hamelin would be Steve Balboni or Travis Hafner

I don't think you could look at all three sets of matches as rookies and predict with any authority that Walton, Smith, and Hamelin would bomb, while Grissom, Greer, and Hafner would thrive.

After reading in another thread how 30 ABs was enough to write off Bonds, I'm surprised to see that 271 PA don't mean anything at all...

(not that same person said both things, obviously)

I think Smith and Walton were pretty similar actually, but I'd have to look at the numbers. They did finish 1 and 2 in the ROY voting in '89 and their toolsy skill sets seem similar, but its been a long time since '89 and my memory may be lying.

Not really. Walton was a pretty good CF. Smith was a born DH. Smith was a far better hitter. Walton was worthless at the plate if he didn't hit .300.

Smith was fast, but a pretty poor baserunner. Walton I recall was an excellent baserunner.

I watched Dwight Smith and to this DAY cannot figure out what the heck happened. Scouts told me the guy had one swing, akin to Dan Pasqua, and that if the ball wasn't where he swung the bat Dwight was hosed.

The guy sure looked like a hitter. He could run. Seemed to know the strike zone. But Miserlou is right in that the only defensive spot was left field and THAT was iffy. He made Lonnie look adept.

I wrote Jerome Walton off as a legit player two months into the '89 season. You could tell THAT was luck. He was just a guy. Dwight Smith, totally different. I figured that guy for 10-12 years and maybe an All-Star game or two just on his hitting. Good one Harv..........

Gonna have to be tougher than that to make it as a researcher.

Taking isn't the problem. My feelings aren't hurt, and I certainly don't lose any sleep over any BTF comments. In fact, so many BTF comments are excellent, which is why I consider myself a primate. However, there is no reason to put forth comments in such a manner. I respond to intellectual bullies, because they prevent people with good comments from making them. I'm trying to point out that such behavior is unnecessary. Over my research career, I have received hundreds comments, some good and bad. If I couldn't take it, I would have been weeded out long ago. Rejection comes with the business, it's the tone I have a problem with.

Rejection comes with the business, it's the tone I have a problem with.

I understand completely, JC. Did you get the chance to look at my Glavine piece? It really has the same comments (fromt eh same poster).

Assuming this site reaches a wider audience than your blog, having hte discussion here is *likely* to get more good evaluation. Not necessarily, but potentially.

I still think wistfully about Jeff Stone: 70 for his first 189 for the Phillies in the mid-1980s (.370!!), one of the fastest guys I ever saw on a ballfield. How could he have missed?

Smith was fast, but a pretty poor baserunner.

I know we're getting off the topic, but this may be the understatement of the year. Jimmy Piersall said that Dwight Smith had one baserunning strategy--Run with your head down until somebody tags you.

Smith suffered because the Cubs, for at least the last 20 years, have had abysmal player-developement success. Their players rarely improve as they come to the majors. The Braves seem to have a better track record in this area. I think that bodes well for Francoeur.

With respect to some of the comments above, JC demonstrated that Francoeur's MLB stats last year were partly a function of "luck" or "randomness." JC's proprietary PrOPS shows that he hit at a higher rate than expected. To wit, "his PrOPS for 2005 (.270/.311/.506/.816) are much lower than his actual line (.300/.336/.549/.884). That is an over-performance of 68 OPS points, about 8%. He was a bit hit lucky."

JC also detailed that Francoeur's MLB #s were better than his MiL #s. "The funny thing is that his performance in the minors was slightly worse than is PrOPS numbers in the majors."

As far as "looking at each PA on video," JC is a Braves fan and I bet he has seen the vast majority of Francoeur's at-bats, both last year and this year.

I, for one, thought his article and accompanying analysis were balanced and insightful.

Would it be such a bad idea to send Francoeur down if he's overmatched? What happened to Kelly Johnson? I figured Cox would plug him in there if Francoeur or Langerhans struggled.

I wrote Jerome Walton off as a legit player two months into the '89 season. You could tell THAT was luck. He was just a guy.

IIRC, Bill James totally disagreed with you. In his 1990 book (they're all at home right now), I remember him raving about Walton. It gave me warm fuzzies as a Cub fan. 1989 turned out to be a complete fluke for much of that team (except for Maddux and Moyer).

I understand completely, JC. Did you get the chance to look at my Glavine piece? It really has the same comments (fromt eh same poster).


Do you mean this One? There were some doozies, like "I'll address this post to Sam M., because I don't think that Chris is either able or willing to understand it." Compared to that Mitchell has been sweetness and light here.

Bob Dernier Cri,

if you didn't catch it the first time around, you might enjoy this excellent article about Jeff Stone and why he didn't make it.

Link

Kelly Johnson hurt his elbow and hasn't played at all yet, minors or majors, AFAIK. As recently as this month he was unable to throw the ball at all, according to anecdotal reports from people who have seen him during batting practice.

Would it be such a bad idea to send Francoeur down if he's overmatched? What happened to Kelly Johnson? I figured Cox would plug him in there if Francoeur or Langerhans struggled.


He's hurt. The question is whether Cox / Schuerholz will cut bait with Brian Jordan when Johnson is healthy, send Matt Diaz to the minors, or leave Johnson in Richmond.

Jeff francoeur did _not_ have 271 PA of great performance. He had about 120 PA of insane performance, and has since followed that up with 240+ PA of absolute sucking:

Prior to drawing first major league walk:
379/394/734, 10 HR in 120+ PA
Remainder of last season after that (including postseason):
228/281/376, 4 HR, 11BB in about 160 PA
Total line since then including postseason and this season (thru Tuesday):
213/250/354, 7 HR, 11 BB in about 240 PA

My cutoff of when he drew the first walk is coincidence; I eyeballed his game log from last year to try to identify when he really started to plummet. He peaked a few weeks earlier, but this looked to be the point where he really started to suck.

There's some question as to whether Kelly Johnson will be back at all this year.

Johnson's injury is a pretty big blow to the Braves in terms of OF depth...they still have Jordan and Diaz to possibly 'spell' Francoeur, but they're pretty much bench players and short-term solutions at best.

I don't know what the Braves should do at this point. Like most, I expected a regression, but this is beyond the pale - to the point where you just feel for the kid. I think he has too much natural talent not to be a decent player in this league - but he's going to be *very* up-and-down until he (hopefully) gets better pitch recognition and some semblance of the strike zone. Now whether he should learn in AAA or the majors...

JC demonstrated that Francoeur's MLB stats last year were partly a function of "luck" or "randomness."

All that you can demonstrate is that his actual performance was an outlier, of indeterminate cause, and that all else being equal the PrOPS line was likely to be closer to his actual 2006 performance than his real 2005 performance would be.

I am really of the opinion that we need to stop talking about "luck" and "randomness" and throwing around terms used by statisticians, if we want our work to be more broadly accepted, and start investigating outliers more closely before dismissing them as "just" outliers. Sometimes individual players don't fit the broad statistical models.

-- MWE

Thanks for that link, Guapo. I hadn't seen that Jeff Stone story. He is obviously a smart, hard-working guy, though the Philadelphia press used to portray him as a fool.

Thank God for Mike Emeigh. They should put you on sabermetric posters :-)

I am really of the opinion that we need to stop talking about "luck" and "randomness" and throwing around terms used by statisticians, if we want our work to be more broadly accepted, and start investigating outliers more closely before dismissing them as "just" outliers. Sometimes individual players don't fit the broad statistical models.

I agree with Mike on this one - luck and randomness might account for the change in some players, but its not the only reasonable explanation. There sometimes are, as Mike indicated, outliers.

Its possible that Francoeur's early performance was simply luck, but its also certainly a reasonable possibility that scouts and pitchers figured something out about him. Both are logical explanations, and there is no real evidence regarding which is the primary factor here.

In fact, after reading this article, his whole permise is silly and misguided.

This is really an awful statement by MGL. Not because the reason suggested in the original article is neccesarily true (I personally lead more toward the randomness explanation), but because the original article is well thought out and certainly a reasonable explanation and should not be discarded with such arrogance.

and Bob Hamelin

I'd caution against drawing any conclusions from Hamelin. He had serious eye and back problems.

also add Al Bumbry, and Pat (your father's) Listach

I mean, Bumbry was never a BAD player, just never anything like his rookie season of 400 or so plate appearances

Listach, on the other hand, was always a bad player

Suppose that a team had no scouting report on Francoeur, just his AA line stats.

I could see this being the case with the Bo Harts of the world, but didn't everyone know about Francouer's plate discipline issues? For $25 you could probably have found this out from BA's prospect book.

I wouldn't put Bumbry in this group. His OPS never matched his rookie year, but he did have a hell of a year in 1980.

I am really of the opinion that we need to stop talking about "luck" and "randomness" and throwing around terms used by statisticians, if we want our work to be more broadly accepted, and start investigating outliers more closely before dismissing them as "just" outliers. Sometimes individual players don't fit the broad statistical models.

Luck and Randomness seem to be the new gods. In the old days when people couldn't explain stuff they'd blame it on the gods. Now, I'm a theist, so this isn't necessarily a bad thing....

My assumption is that Franceuor is the type of guy who can really hurt you if the pitcher challenge him- but is in a lot of trouble if a pitcher of at least moderate mlb ability tries to work the plate and count.

I also suspect that the majority of mlb pitchers don't care what a scouting report says on some guy fresh out minors- they are going to challenge him until he "proves" something.

Or it just could be randomness- Franceuor's career mlb stats NOW are not that far off his translated stats...

Sometimes individual players don't fit the broad statistical models.

And sometimes guys that hack and flail their way to proficiency in their first 30 games or so predictably come crashing back to earth. Sure, on the whole, outliers should be examined more closely but let's not complicate the Francouer case.

Luck and Randomness seem to be the new gods.

Yeah, I just read "Against the Gods", and oddly enough Keynes is quoted as saying that the problem with economics is that it can tell you that the ship will sail smoothly overall, but it won't help you with the storm dead ahead.

also add Al Bumbry, andPat (your father's) Listach


Listach, on the other hand, was always a bad player

And you know who had a remarkably similar (but inferior) rookie year to Listach? Julio Franco.

Listach .290/.352/.349 54/18 SB 99 OPS+

Franco .273/.306/.388 32/12 SB 87 OPS+

Franco had better power, and a little better plate discipline despite walking half as much, 50 K to 124 K in about the same AB's. But there's noting about those two lines to make you say the top guy will be out of the game by age 30 and the bottom guy will become the oldest player ever to hit a HR.

OK so everyone railing against luck and randomness, get to work. Tell us what's going on.

</blockquote>Mike Emeigh Posted: April 27, 2006 at 09:44 AM (#1992410)

I pondered---notice the post wasn't titled "explaining"---if minor league scouts scout double-A scout players differently than major league advance scouts.


Scouts are all looking for the same things, regardless of level. The problem is that scouting reports for minor leaguers are based on incomplete data. Scouts don't see every player play every game - teams don't have that many scouts, they have to cover more than one team, and there isn't as much video of minor league games as there is of major league games, so you can't always go back and look at tape. You make judgments based on what you see, when you see it. A minor league scout might have seen Francoeur play 10-15 games, tops, against widely variable competition. There's much more room for error when evaluating a minor leaguer (just as there is when evaluating his statistics).

-- MWE


I rarely post here... but Mike's above comment here couldn't be further from the truth... A Minor League Scout evaulating a AA player looks for TOTALLY different things than a Major League Advance Scout. It is a different job, requiring a different skillset (which is why teams employ different individuals in these roles).

A Major League Scout is evaluating a AA player based on his skills as they are (his 5 tools) and how his tools project to play in the future... Grading him on both how he would fare at the major league level now and how he would perform in the future. A scout will not only review his tools, but his makeup and other non-quantifiable factors. Certainly a scout will include in his report items such as "tendency to chase sliders off the plate" or other weaknesses... but in a less precise and definitive way than a MLB advance scout. A Major League scout at a major league game will evaluate major league players in the same way as a minor league player.

A MLB advance scout will frequently chart every pitch thrown to a batter (both pitch type and location) and will include the results of the pitch (i.e. taken, swinging, put in play etc.).... a MLB advance scout will produce a hit chart and identify specific pitching patterns that a hitter is vulernable to. The information is much more detailed than the work performed by a Major league scout at the AA level.

From the teams that I am familiar with some of their internal processes, they use different computer systems to record the reports of the scouts scouting the Majors and minors than the use for the reports of the advance scouts (which is often further supplemented from information received from a Team's Video Department).

Joe

I wouldn't put Bumbry in this group.

OK--Rich Coggins,then; he & Bumbry were the Smith/Walton of their day



so after 21G, 85 PA's, Francoeur is at .190/.207/.303

in his first 85 PA's after his ROY award, Bumbry was .198/.247/.296

They did finish 1 and 2 in the ROY voting in '89 and their toolsy skill sets seem similar, but its been a long time since '89 and my memory may be lying.

Smith batted lefty, Walton right, and Smith had a much broader skill set--he had more power than Walton, better discipline, and at least as much speed. I thought he was going to be a hell of a player; I always thought Zimmer gave him the shaft.

Aw Miser, Julio Franco had a MUCH superior minor league record compared to Pat. As a Brewer fan, I would love to be able to agree with you but Franco was just a more polished hitter then Listach. And hitting in Cleveland back then was very different then the 1992 County Stadium.

I know it's easy to claim NOW so clearly I will accept folks rolling their eyes. But any Brewer fan in attendance then and now will agree that Pat just had one of those years. He was a .250 hitter in the minors, in hitting environments, who did manage to draw some walks. But you could see that pitchers could knock the bat out of his hands. Garner running him into the ground didn't help matters as Pat didn't have the body to hold up to the pounding of running the bases willy nilly.

As a totally separate note, all of the Brewers who ran amok in 1992 saw a downturn the following season. Would make for an interesting study to analyze the before and after of a team strategy of running wild one season.

Harvey is correct, however, that Walton was a much better outfielder than Smith. Smith's positions were LF and DH.

JCB, I'm a big fan of the work you've done (Mazzone, Props) but I do think you have a tendency to overreact to criticism. The criticism might or might not be constructive or warranted but I don't think you should it take it as a personal assault.

Also, JCB, I don't think you bring out the worst in people here. It's always like this.

Looking at Francouer's game log, what Colin missed is that after the game where Jeff drew his first walk for the next month he was playing against teams that had already seen him. By the time he saw the Marlins for the first time on Sep, 23rd his average had dropped 64 points in a month. The real drop off is between the first 10 or so at bats against a team and everything after that.

This is the traditional once around the league and then fall off the cliff pattern.

Smith batted lefty, Walton right, and Smith had a much broader skill set--he had more power than Walton, better discipline, and at least as much speed. I thought he was going to be a hell of a player;

and actually, their rookie seasons weren't THAT comparable (offensively)

Smith was .324 .382 .493 OPS+ of 142
while Walton was .293 .335 .385 OPS+ of 100


there's a tendency to group them together because they happened to be rookies on the same team

I have to disagree with some of the above comments. I watched an awful lot of Braves games. I saw the kinds of pitches that Francoeur was getting earlier on and they were fastball strikes. I know it doesn't make sense, but it has been my observation over the years that clubs do tend to throw a lot fastballs to guys who first come up. JPWF13 is right too, Francoeur is the kind of guy who can hurt you if you challenge him because he drives the ball well. That is mostly why he did so well at first. It wasn't simple randomness.

At this point, I don't think Francoeur can learn to handle major-league pitching. He seems to be missing some of the requisite visual skills that make it possible for him to effectively handle breaking and off-speed pitches. Those visual skills are congenital. You either have them or you don't. If that is the case, asking him to learn patience and be selective is like asking someone who is color blind to learn to discern colors. It just isn't going to happen.

I know it doesn't make sense, but it has been my observation over the years that clubs do tend to throw a lot fastballs to guys who first come up


yes, they do, and, no, it doesn't really make sense


but that's the way it's been since I can remember


I guess it's to see if they're man enuff to be up in the Bigs

Good posts while I was at lunch.

Yes, dlf, that's hte one.

Nice link Guapo.

Very cool, Colin - I didn't realize it was so dramatic.

Nice to see you, Joe. Hope things are good.

I almost mentioned Listach as well.

A Major League scout at a major league game will evaluate major league players in the same way as a minor league player.

I've talked to a number of scouts who do both major league and minor league games when they pass through my neck of the woods, and typically they evaluate players in the same way regardless of level, and capture the same set of information. That said, I accept Joe's comment that advance scouts at the major league level do something different, and I should not have used the general word "scout" to refer to both sets of individuals.

-- MWE

One thing that was obvious from watching him play in July was that since he was a right hand batter pitchers would go down and in on him and get crushed. Usually its lefties you avoid goingdown and in on. As the season wore on you saw him getting less and less of that.

Also, i really dont understand how you can write off 10 HRs in 100 ABs as luck. This isn't singles falling in to boost a BA

ok...doing a quick check

through two starts against a particular team: Jeff was 41/97 with 7 2b, 1 3b, 8 hr, 3 bb, 19 k, for a .423 average and .763 Slugging. after his second start against a team he 36/160 with 13 2b 6 hr, 8 bb, 39 k, for a .225 average and .419 slugging.

Also, i really dont understand how you can write off 10 HRs in 100 ABs as luck. This isn't singles falling in to boost a BA

Kevin Maas says "hi". So does Victor Diaz.

And Mike Jacobs.

It's not "luck" to be *able* to do that, but many prospects will hit 10 HRs in a *given* 100 ABs. That he did it in his first 100 ABs is teh "chance" part.

Also, i really dont understand how you can write off 10 HRs in 100 ABs as luck

Kevin Maas says hi

p.s. pattern holds for series against houston (who had not seen him in the regular season), first two games he was 3/7 with a a triple, a walk and a strikeout, the next two games he was 1-10 with a 2b, a walk and three k's.

My guess is that Francouer was simply lucky in those 271 PA and that people as usual are trying to make some "sense" out of randomness.

...

The fatal error is assuming that his performance in those 271 PA means anything at all - that poor performance following good performance means that "pitchers have figured something out."

Isn't it possible that a factor in Francouer's being lucky last season was that he was getting more than his usual share of hittable pitches? Of course it is.

Does his good luck prove that he was seeing more hittable pitches? Of course it does not.

"[P]oor performance following good performance" does not necessarily mean that "pitchers have figured something out." But it doesn't mean that it's impossible that the pitchers figured out something, either.

I don't know thing one about Jeff Francouer, but the inherent randomness of player performance doesn't preclude the possibility that sometimes there may be explanations for why players hit above or below their true talent level for a period of time.

scooped you, jmac!

Good comp: Angel Berroa
poor zone control, good rookie season, sux ever since.

Worth noting that although the mythology says that Charbonneau didn't hit in the minors, he actually chewed them up. He was just very old at some of the stops. 23 years old in A ball due to his being suspended for almost a full season. As such, not really on radar as a prospect, but that's a very different thing. (For a guy who didn't hit a lick in the minors and went on to have a really good career see Bobby Bonilla. I know he was hurt most of the time, but Bonilla never truly mastered A ball)

He hit the majors in 1980 and had hit .350 in 1978 and .352 in 1979 ( BA titles both years) and had hit 59 doubles and 39 HR in 240 games in those years. Don't have his walks but that's a pretty fair package even if he didn't walk much. And he wasn't a hacker in 1980.

I am really of the opinion that we need to stop talking about "luck" and "randomness" and throwing around terms used by statisticians, if we want our work to be more broadly accepted, and start investigating outliers more closely before dismissing them as "just" outliers. Sometimes individual players don't fit the broad statistical models.

Quite.

In fact, I'd go further - in cases like this, attributing a thing to "luck" is not an explanation. It's a refusal to give an explanation, an indication that you don't care about the matter. Which is fine, but then you shouldn't insult people who do find it interesting. To take another popular Primer discussion, I believe that Beltre's 2004 was a fluke in that he is extremely unlikely ever to replicate those numbers in his career. But it most certainly wasn't a fluke in the sense that he rolled double sixes every time he came to the plate. He was seeing the ball and hitting it with authority - 48 HRs, not 48 seeing-eye grounders. And now, of course, he's back to where he was. If you're interested in this from a purely statistical perspective, it makes sense to write this off as an anomaly, and move on to something else. But from a mechanics-of-baseball point of view, I find it fascinating.

The same applies - although to a much lesser extent - with Francouer. It seems that mgl is essentially saying that just because Francouer did very well in his first 271 MLB PAs, when that is taken into account with the rest of his professional career we had little reason to think he will do well in 2006. And, although he says it in a nasty way, I wouldn't begin to argue with him. But the question JCB raises is why Francouer did so well in those 271 PAs, which mgl doesn't address. I found his theories interesting, although I don't agree with all of them.

Perhaps what this comes down to is a difference in perspective. As a fan, I simply want to understand the mechanics of baseball, and how to appraise past performance. Trying to project the future is of little interest me. From the point of view of mgl, involved in professional analysis and so on, it's a different question. Statistical projections are what matter because that's what he can offer - and I respect him a great deal for it. I just wish he wouldn't dump all over this thread.

Good comp: Angel Berroa
poor zone control, good rookie season, sux ever since.

And what's the point of all these random examples? No one has claimed that rookies who have good seasons then tank are rare; it happens all the time. But is there something that these guys have in common, aside from their sophomore year struggles? I don't see it. For every example of a flop, there are examples; same age, same position/defensive ability, similar batting lines, similar speed indicators, etc, who were successful.

What is it about Jerome Walton that made him a bet to flop, while Marquis Grissom did not? Or Bob Hamelin/Travis Hafner? that's what I take mgl's " I highly doubt that any particular subset of rookies has early success and then a large dropoffs " to mean.

That said, I do think there's one; The 26-27 year old DH types. Look at Hamelin's age 27 comps. A sad list. Bye Bye Balboni's the best of them. Guys like Hafner are the exception.

Also, i really dont understand how you can write off 10 HRs in 100 ABs as luck. This isn't singles falling in to boost a BA

As it happens I ran a fairly large study a few years back of player batting stats in consecutive month. Limited it to players who had 90+ PAs in consecutive months -- making it very much on point here. (study had 1471 player/month combinations. I tried to cull all of the players who had been traded. May have missed a few but I doubt it'll make a big difference). Rates to largely eliminate the issue of injury.

Standard deviations of various stats in the study.

BA OBP SLG IWR ISO OPS
.049 .057 .117 .034 .090 .160

(IWR is OBP-BA. ISO is SLG-BA.)

Largest swing in the study: Brett Boone

Boone hit .367/.417/.755 on 107 PAs in July 1995 Followed it up with a .152/.252/.212 in 112 PAs in August.

Point being that 100 or so PAs don't tell much about a player's ability level. The expected range of performance is huge.

Also, i really dont understand how you can write off 10 HRs in 100 ABs as luck.

Shane Spencer scoffs at such feeble production.

What is it about Jerome Walton that made him a bet to flop, while Marquis Grissom did not?

Oh, man, that one's easy. I'm no sabermetrician and I can tell you: Jerome Walton was a centerfielder for the Chicago Cubs.

What is it about Jerome Walton that made him a bet to flop, while Marquis Grissom did not? Or Bob Hamelin/Travis Hafner?

Statistical analysis - at least at the macro level that we typically have to evaluate due to the level of data available to us - is not sufficiently granular to give us that answer. "How" a player achieves his numbers - the shape of his performance - is (normally) more important than the numbers themselves. As more granular data becomes widely available, we might be able to get a better feel for this.

We also don't know the extent to which external factors - injuries, personal stresses, conditioning or lack thereof, drug usage (not talking steroids here, FWIW) - might have affected a player's performances.

-- MWE

Johnson's injury is a pretty big blow to the Braves in terms of OF depth...they still have Jordan and Diaz to possibly 'spell' Francoeur, but they're pretty much bench players and short-term solutions at best.


Yes. Francouer would not be with the Braves right now if Johnson were healthy. It's been a huge blow. Huge.

Oh, man, that one's easy. I'm no sabermetrician and I can tell you: Jerome Walton was a centerfielder for the Chicago Cubs.

Hmmm, Adolpho Phillips, Walton, Corey Patterson. You may have something there. Is it too early to write off Felix Pie?

At least Andy Pafko made good.

I wasnt saying Francouers homer per 10 AB would be easily repeatable. I was just trying to say that it wasnt merely "random luck", quite a bit of skill was needed to do what he did. He was red hot (of course thats not sabr approved either) and playing great. The luck he got was from pitchers not knowing his tendencies but more than anyrthing his inital success was a result of skill, not luck

Well, maybe his new deal with Delta will help him relax.

I wonder if they were able to get an escape clause just in case he gets sent down and never makes it back up, or if the braves are able to package him in a trade for a serviceable outfielder.

"luck" and "randomness" as we talk about them in baseball take on many identities. It is a misunderstood concept. It does not mean that there is not a logical explanation. It is not only about "bloop hits that fall in" or line drives that are caught. It is about getting more than your fare share of hittable pitches. It is about HR's that just make it over the wall or don't. It is about wind blown home runs and home runs that are blown back in. It is even about daily fluctuations in hitting talent or the fact that a player's contact lenses were bothering him on a particular AB.

And yes, as I said in my original post, it is certainly possible that JCB's premise is true and I am certain that that secenario happens occasionally, maybe even more than occasionally. My main contentions were that in the absence of anything (evidence) other than speculation and conjecture, it is likely that Francouer's pattern of performance is not much more than random fluctuation. Or at the very least, that is always the default theory we must fall back on until and unless we have evidence that allows us to make inferences otherwise. I see no evidence of anything in this article. assume it was not meant as a "research study" by any means, so that's fine.

I also stand by my statement that teams would be throwing "right at" a hitter who is known to have little plate discipline (JCB is the one who stated that that was the scouting report on him) is a ridiculous assertion. Do they throw fastballs at guys who are good at hitting the fastball and curveballs at hitters who are good at off-speed pitches? Are teams living in a "bizzaro world" to paraphrase Rob Neyer. My guess is that JCB now knows this (he is a real smart guy), but does not want to back down, as most people are loathe to do. My guess is also that rookies receive far more breaking balls than fastballs, as compared to veterans, again, for obvious reasons, but again, I do not know this and would not be shocked if I were wrong.

And BTW, I have trmendous respect for JCB. His work on Mazzone was brilliant - one of the best revelations in a long time. However, if I don't like something that someone writes, I am going to say so even if it might hurt their feelings. I don't recall saying anything personal about Mr. Bradbury in any of my posts.

Then maybe you should change the damn word to something else. "Luck is a misunderstood concept". Every person who walked the earth has an understanding of what luck is. Sabers have twisted it to mean something else. If I started calling cheese 'bread' and mocked everyone who didn't understand my new paradigm, I'm pretty sure I'd take a lot of flak. Lucky is lucky. Randomness is not luck. Change your damn word.

I think that labelling francouer a bust based on 100 PAs this season is just as ridiculous as labelling him the messiah based on his 100 PAs last July.
Most hitters when in a slump look awful (and amazing in the midst of a hot streak), and any number of armchair scouts can point out what they are doing wrong. He is not using the whole field, he is jumping at the pitches, he is swinging early in the count ( all problems currently plaguing Francouer ). Hitting is a dubious skill to quantify, and IMO, lot of it has to with personal confidence. thats why the famous "sophomore jinx". How do young players with no proven track record adjust when they hit a big slump for the first time? If Francouer is unwilling to make any adjustments, thats when you have a "problem".

In response to Mike's post, what is the purpose of quantifying luck? There is no true talent level at any point of time, there are only confidence intervals, which we can project to based on available data. more data, less noise. While it might be interesting if someone could quantify luck, I think it would be an exercise in futility. Outliers are people at the edges of the confidence interval, who due to rather propitios conditions, achieved their stats. Didn't PECOTA have Ryan Howard projected for 70 homers with a 5% chance this year?

As for the article itself, I thought it was a fairly intelligent comment. As to why minor league pitchers might have been more successful, I think one would also have to think that being minor leaguers, they have less control, and are not around the strike zone all the time. I can see type of pitchers giving francouer fits, yet not totally shutting him down.

no proven track record

as an addendum, i know MLEs exist, but not in the hitter's mind. Every level is a new challenge, which you wonder if you can overcome...

In response to Mike's post, what is the purpose of quantifying luck?

Because I don't think that is what Mike said or what he is talking about.

It isn't about trying to develop an entropy measure. It isn't about trying to improve the calculation for the confidence interval in Classical Statistics.

What Emeigh is talking about is with improved instrumentation, you should be able to get improved metrics.

Daly hits on the problem. Depending on the context, the sabes use luck to either describe variables they randomize in a projection, or for prior event spaces, those factors that they consider to be external causal agents.

Most of the rest of the world uses luck to mean someone's fortune or fate.

When the scientists mix with the fanboys, they then create a new "God" This external thing called "luck" then creates causality into the event space.

What Emeigh is saying is that with more instrumentation:

(1) We don't need to randomize everything, with better instrumentation, we can predict some things that we currently project with random variables. AND
(2) Not everything that is treated as an external causal agent is totally external. We should be able to exhibit influence on some things.

He could also be saying:

(3) Even if there are external agents that we randomize in projection, better plans can still be developed to deal with those things for which we have partial control.

The problem is, particularly among certain populations, is people will just throw their hands up and say "it was luck."

I'm no great fan of Cadbury, but he is at least formulating a hypothesis on repeatable criteria that can be seen among certain populations. Some of it sounds plausible.

mgl opines that you cannot create any type of selection algorithm to define that set. That is also valid, although I think it might be obtainable.

But to answer your question and to also deal with this "luck" concept, that is why the word is dangerous and it impacts conversation.

Francoeur did exhibit the performance output over these last two years. The other examples of Walton, Smith, etc. did exhibit those output lines in those years.

Over the set of all MLB players that make and play in the major leagues, the MLE's do converge to whatever error rate mgl cited.

There should be no arguing these conclusions. The only question is whether the whole set that is measured by MLE's exhibit two distinct curves or one distinct curve, each curve having its own statistical properties.

If you answer that question, the next question would be is there any causal agent you can introduce that will push the "slumpers" into a more stable development pattern.

That isn't measuring luck. That is searching for what would be a valueable tool in personnel selection.

You know i love statistics and all stat tools...however...if there is one thing that i wish people like mgl would acknowledge it is that:

Whether the results are random or not...the causes are never random...been to a lot of games and never seen a pitcher or hitter rolling dice yet...

All the people involved in a play make concious choices...and those combine to create a "psuedo-random" result. But that apparent randomness is only because of the limits to our knowledge of the process. With every increase in knowledge we decrease the level of randomness in the results.

If we had stats breaking down how Jeff F. hit based on the type of pitches thrown, where they were in or out of the strikezone etc...we would be having a much different arguement. Without those kinds of numbers arguing for "randomness" is exactly the same as old timers who felt that there was no skill at all involved with drawing a walk, it of course was purely random, or only "luck".<b></b>

I agree pretty much entirely with BL's post, with one comment:

If you answer that question, the next question would be is there any causal agent you can introduce that will push the "slumpers" into a more stable development pattern.

That isn't measuring luck. That is searching for what would be a valueable tool in personnel selection.

And even if you can't identify a causal agent that will help the "slumpers", it would be extremely valuable if you can at least identify whether a player is a slumper early on. If you knew in October '06 that Francouer would fall appart, or Angel Berroa after '03 or Dwight Smith after '89, then you don't make plans around them or give them big contracts.

I think MGL is positing that you can't even identify that. I would say we currently are unable to do so given the data available to us outsiders, but wouldn't go so far as to say that something like that is impossible. Which I guess is what Emeigh and BL are saying.

I also stand by my statement that teams would be throwing "right at" a hitter who is known to have little plate discipline (JCB is the one who stated that that was the scouting report on him) is a ridiculous assertion. Do they throw fastballs at guys who are good at hitting the fastball and curveballs at hitters who are good at off-speed pitches? Are teams living in a "bizzaro world" to paraphrase Rob Neyer. My guess is that JCB now knows this (he is a real smart guy), but does not want to back down, as most people are loathe to do.

I believe that you might be taking this a different way than was intended. From what I read of the article, it seemed to me that he was making a conjecture of why Francoeur was doing well rather than making an assertion that scouts must be telling the MLB pitchers should just chuck it in.

People do have the tendency to be loathe to back down, but I believe, in this case, it's miscommunication (and perhaps mgl using words which are cantankerous, at best).

I also agree with what Backlasher is clarifying in this conversation. Emeigh, IMO, is talking about better instrumentation and not trying to "quantify luck" but rather trying to see if you can get more data that is relevant to future predictions.

I think I would disagree with Godot that there IS a true talent level at any point in time. However, I DO think that it's nigh impossible to figure out, and that's where the confidence level issues come in.

Whether that data will lead to enough improved projection to cover that confidence level is questionable, but not necessarily futile.

Also, I, too, believe that "luck" should probably be discarded for a different term like "outlier" or "randomness." Too often, people get stuck on the semantics and therefore shut down their brain and refuse to discuss when they see results attributed to "luck," and what's the point in using a term which will repeatedly get misconstrued?

Whether the results are random or not...the causes are never random...been to a lot of games and never seen a pitcher or hitter rolling dice yet...

All the people involved in a play make concious choices...and those combine to create a "psuedo-random" result. But that apparent randomness is only because of the limits to our knowledge of the process. With every increase in knowledge we decrease the level of randomness in the results.

While this is, in part, true, in the end, EVERYTHING comes down to randomness (if you believe in Heisenberg's uncertainty principle). So, in a sense, you DO have randomness rearing it's head even if you have access to perfect data beforehand. The question is how much does this randomness affect the predictibility of events.

F

Human nature, in general, has a very hard time accepting anything as chance. I think that's one of the most annoying things about people - Astrology, superstitions, seeing patterns where there are none, etc.

I guess to some, what seems to them to be a movement to explain away everything as chance is equally annoying. I don't see it that way, but hey, could understand how some people get that idea.

through two starts against a particular team: Jeff was 41/97 with 7 2b, 1 3b, 8 hr, 3 bb, 19 k, for a .423 average and .763 Slugging. after his second start against a team he 36/160 with 13 2b 6 hr, 8 bb, 39 k, for a .225 average and .419 slugging.

You might have something. He sucked the last month and a half he was in the majors though. Did he suck because teams were developing a scouting report on him, or did he suck the more he faced teams because he was just sucking in general.

We ain't talkin' bout the game. We talking bout luck. We talkin bout luck man.

</Mitchell G Iverson>

You might have something. He sucked the last month and a half he was in the majors though. Did he suck because teams were developing a scouting report on him, or did he suck the more he faced teams because he was just sucking in general.

well....test this from both ends...the first couple of months when he wasn't sucking, he batted .200 when facing teams for the third game or later. In the last couple of weeks and the playoffs when he faced three teams for the first time he batted .381 with walks and power in the first two games against those teams, and .058 after the second game against those teams.

In the last couple of weeks and the playoffs when he faced three teams for the first time he batted .381 with walks and power in the first two games against those teams, and .058 after the second game against those teams.

You're getting incredibly small sample sizes at that point though.

Whether the results are random or not...the causes are never random...been to a lot of games and never seen a pitcher or hitter rolling dice yet...

Rolling dice is not any more random than hitting. Just like hitting, the results are determined by forces that are measureable. If you are going to define one of them as being luck based, then it follows that the other is as well.

Also, I, too, believe that "luck" should probably be discarded for a different term like "outlier" or "randomness."

Uncertainty?

I think "luck" is often used as shorthand for "that which we aren't measuring/predicting/adjusting for".