How often do articles grow from Facebook status updates?

I'll cut-and-paste my comment.

- The sabermetric explosion and Moneyball have resulted in a lot more people wanting to make a name in sabermetrics than were in the past.

- The sudden increase in types of data available look to be propelling sabermetrics into a new age, similar to that of the 80s. I'd probably describe the 1990-2008 (give or take) as the period of classical sabermetrics, in which the forms were set and most work was done in sorting out and refining concepts. Error bars were part of the established language and even new things like DIPS were heavily indebted to previous ideas (it's essentially, at its core, a spin, though a novel one, of component era). ... See More

Now we have all sorts of new things coming in, pitch speed, location, detailed batted ball data, and there just hasn't been that long sorting out period in which the validity and use of the information. There's going to be a flurry of new measure and acronyms that we aren't used to and it's going to be a while we really get a handle on it. Kind of a neoromantic era.

- There are no doubt some generational issues related to the first two points. The so-called classical era of sabermetrics had a fairly small orbit, with most of the work deriving from the fairly small group of usenet denizens and STATS employees back in the mid-90s. That group is getting older (I'm probably the youngest of this group and I'm in my 30s. and struggles of generation are perfectly normal human reactions. For example, I still greatly enjoy young indie rock bands, but it gets more awkward every year to be hanging around a bunch of 21-22 year-olds at a bar.

I guess the market will correct this somewhat over time. If there are too many people and not enough breakthroughs, a lot of people would lose interest. And, I imagine, the new types of data and measures available will become more mature over time.

To some extent, I agree. Some of the work I see lends very little new insight. But, isn't this the way a lot of science works? We make a little progress here and there, and a lot of it is kinda boring, incremental stuff, if it makes any contribution at all. But then, every once in a while, somebody makes a cool argument. I'm impressed by some of the things to capture player's speed contributions, and some of the recent Pitch F/x stuff on strategy and pitch-by-pitch analysis. I think there's a lot on defense and pitching to learn.n Sure, the new metrics may offer little new and much metric. But, there's a lotta cool baseball analysis out there.

I can sort of see where he's coming from. Way back in the day there were the old numbers, and then people found new ways to distill new insights from them. But after a point it started to go past insight (ie "OBP and SLG are better than average") and move towards precision that fans had to read rather than figure out themselves (ie linear weights). Then there seemed to be a bit of a plateau in new thought, as the Retrosheet data started becoming complete and before in-game numbers became available.

Since the wide-spread availability of batted ball data there has been a wave of new thought and neat stuff, but if you're not a math guy you can feel the numbers coming to start the move past the observations and into, well, math, faster than ever. Plots of pitch locations for one guy make sense, but the multiplication of minor league park factors (by area of the OF, by pitching hand, by batting hand, etc) or a whole raft of new BPro numbers can make this hard for casual fans to keep straight.

I can see a guy like Sickles, a liberal arts scouty dude, ending up like the old geezers at Baseball Digest at some point. I think he's proven he CAN keep up, but it looks like someone has to prove to him it's worth keeping up again. I like John's work a lot, so I hope he can come to some point of rest and rejuvenation where this is a joy. Baseball should bring joy.

I can see a guy like Sickles, a liberal arts scouty dude, ending up like the old geezers at Baseball Digest at some point. I think he's proven he CAN keep up, but it looks like someone has to prove to him it's worth keeping up again. I like John's work a lot, so I hope he can come to some point of rest and rejuvenation where this is a joy. Baseball should bring joy.

Well, but if he's a "geezer" who's utilizing FIP, xFIP, BaBIP, and UZR stuff why would that be bad? Minor league stuff is chock full of short seasons or short stints at a level, field quality issues and park factors out the wazoo, bad defenses, pitchers working on a certain pitch, etc. etc. etc.

Some of the newest stats have only marginal value over what Sickels is already using for his particular area. And he doesn't sound burnt out on his work, just in a wait and see mode on the latest stuff.

For example, I still greatly enjoy young indie rock bands, but it gets more awkward every year to be hanging around a bunch of 21-22 year-olds at a bar.

####, I got over this 15 years ago. Just f'ing go, don't pretend it makes you cool and don't make a fool of yourself too often hitting on some 21-year-old and you'll be fine.

Otherwise Dan is on target I think. There are two "competing" streams in sabermetrics right now. There's the finer granularity that Dan mentions and there is a chance this will lead to a more detailed understanding of why things work the way they do which would be cool even if it only ends up telling us what we already knew. Then there's the battle for the ultimate uber-stat and this I do find tiring. That stream is mainly about minor changes in evaluating offense, defense, baserunning, etc., then adding them all up, then converting them into wins, and then into dollars ... and then pretending there's a lot more precision around that than there is. (the variance gets pretty massive)

An underlying problem is that effect sizes in baseball are small. 20 points of OBP is an important difference but that's just reaching base 2% more/less often. Think about your standard political poll which is usually 1,000 people which produces a standard error of about 3%. A "poll" of 600 PA would produce a standard error of about 4% -- we're trying to detect effects the equivalent of 1/2 of a seasonal standard deviation. (which is why you need about three years of data to begin to be able to say anything.) If you want to compare two players, your standard error is 1.4 times larger, assuming they are independent.

And given we're in the projection time of year, what annoys me is that so little attention is paid to the uncertainty of these estimates -- despite the best efforts of ZiPS and PECOTA (and maybe others) to present that variance in their output. Take the recent Carlos Lee projection. He's got an 8% chance of being "poor", putting the standard error on an OPS+ projection around 15-20 points. Hunter Pence has a 19% chance of being poor which means that his projection probably isn't significantly different than replacement level.

We can really say very little with confidence. Pujols is definitely a significantly better hitter than Hunter Pence; I'm not sure that we can say the same for Berkman vs. Pence and maybe not even for Pujols vs. Berkman.

OK, off my soapbox.

...and don't make a fool of yourself too often hitting on some 21-year-old and you'll be fine.

This is going to be a problem.

...and don't make a fool of yourself too often hitting on some 21-year-old and you'll be fine.

Loophole, yay!

And he doesn't sound burnt out on his work, just in a wait and see mode on the latest stuff.

Actually it is a little of both. I'm at least somewhat burned out right now.

Philosophically, I'm not sure I get Sickels' angle. Well, I guess I do, but I think he's framing it in an unproductive manner.

First off, work sucks. Can you hear your barber saying, "I'm going to keep trimming your head until it's no longer fun"?

Instead of "love," I'd ask myself: Am I good at this? Do I find meaning in doing this work? If the answer to both are "yes," then you're probably in the lucky five percent.

I'm not sure why Sickels doesn't just frame the issue like this: I'm going to skim and scroll sabermetrics and only focus on those studies and concepts that I find to be meaningful. I'm sure he does this anyway. There is no field where every statistical study contributes to someone's "love" of it.

As a fellow humanities person, I think he's fallen into the trap that many of our type fall into: it's all about us and our satisfactions. Like, history is supposed to be fulfilling for us or something. History is what it is, and it doesn't give a rip about anyone's happiness. If it takes looking at 18th century basket weaving to make a case history for larger historical currents, or to learn the skills needed for a larger, later project, then you just have to do it. It's not like you'll be stuck on basket weaving your whole life.

I think Sickels forgets, or maybe ignores, that ultimately he's providing a product, a service, that he's wanting people to pay for. As in any business, he needs to offer something in particular to set his product and service apart from his competitors.

On the scouting side, he's outflanked by BA and by Goldstein at BP; both seem to have better contacts, and both do a much better job of funneling the info from those contacts to their product (very rarely do you get an insightful scouting quote or comp from Sickels). Sickels' book does offer comprehensiveness; he should thank his lucky stars, though, that BA doesn't have a decent search function at their site, or else he might already be obsolete, at least in my view (I mostly used Sickels as a way to get the conventional wisdom on a multitude of guys that I wasn't familiar with yet; the increase in blogs and sites like Scouting the Sally has minimized this aspect of Sickels' appeal).

On the stats side, Sickels is outflanked by fangraphs and minorleaguesplits and countless other free resources.

In my view, thinking just in terms of bizness, what Sickels can offer (besides customer loyalty) to expand or keep his sales: 1) integration of stats & scouting; 2) the actual quality/interest of the writing in the book (a la Bill James).

As his self-appointed consultant, I advise Sickels to focus less on whether known big time prospects are B+ or A-, and focus more on: 1) giving fuller, more interesting & detailed analysis and reportage on a wider range of prospects than anyone else; 2) use his integrative approach and background to focus even more energy on finding diamonds in the rough (either low level guys with special potential, or higher level guys who have slipped under the radar).

This would give his book something in particular for prospect hounds, casual minors followers, and folks who are looking for players for their deep keeper leagues.

I would also suggest Sickels focus less on the grades of the upper level, known blue chip prospects, and focus more on his write-up; go deeper into a guy's background, give specific quotes from scouts on the guy's game, identify aspects that may be an Achilles heel in the majors, a sense of the prospect's personality or approach, etc. Here, Sickels could use his strong writing skills and holistic humanities p.o.v. to give a deeper, fuller portrait of a prospect than either BA or Goldstein can in their scouting tally sheets.

Philosophically, I'm not sure I get Sickels' angle. Well, I guess I do, but I think he's framing it in an unproductive manner.

I thought his comments were very similar to the ones Sam and I and a few others made in the Pierce thread the other day, though obviously from the vantage point of someone actively involved in the game.

Oh, I can totally see a critical skepticism regarding the proliferation of stats. I suppose I just don't see its relevance to one's enjoyment or love of baseball. Maybe everyone's just cranky because it's winter and we're all looking out an existential window, waiting for spring training.

I suppose I just don't see its relevance to one's enjoyment or love of baseball.

I didn't get the sense the proliferation of new stats is damaging his love of baseball, but how it's affecting how he feels about his job and his interest in staying abreast of new developments on the saber front.

The man feels burnt out, he feels burnt out. Just because your attitude might be different doesn't change that.

Just because your attitude might be different doesn't change that.

You don't say? Well, by golly.

We all own our own personal feelings, obviously. I'm just basically saying that dissatisfaction/burnout is probably inevitable w/ the approach to a field of study or work that Sickels outlines.

I am suggesting taking it out of the realm of the personal, and to look at the heuristics involved that lead to certain personal feelings.

Sickels is framing this as a problem w/ the field itself (like with history). I don't know the guy, obviously, but I don't know of many professions that are going to hold up to what appear to be his standards of enjoyment. I wouldn't be so interested in this framework if it were just Sickels' personal one; it seems to be a standard stance t/w work that is very American, and very historically conditioned (it seems like something post-60s). It's an interesting phenomenon to analyze.

I thought his comments were very similar to the ones Sam and I and a few others made in the Pierce thread the other day, though obviously from the vantage point of someone actively involved in the game.

Ayup. Sickels is three years older than me. I think this is probably a generational thing where we're just self-selecting out. I'm still perfectly open to new analysis, but you have to show me the value of the thing before I start replacing slash lines. My interest in baseball is about this year's pennant chase and how the Braves can maximize their chances of winning. If your new stat lines up well for evaluation real, in-season roster/game decisions, I'll look. But if you're just fine tuning OPS+ one more time, meh. Okay. If it floats your boat.

And given we're in the projection time of year, what annoys me is that so little attention is paid to the uncertainty of these estimates -- despite the best efforts of ZiPS and PECOTA (and maybe others) to present that variance in their output. Take the recent Carlos Lee projection. He's got an 8% chance of being "poor", putting the standard error on an OPS+ projection around 15-20 points. Hunter Pence has a 19% chance of being poor which means that his projection probably isn't significantly different than replacement level.

We can really say very little with confidence. Pujols is definitely a significantly better hitter than Hunter Pence; I'm not sure that we can say the same for Berkman vs. Pence and maybe not even for Pujols vs. Berkman.

This sums up my feelings very well. I have a lot of respect for the guys who run these projections, but how much are we really learning? I think someone noted the other day that PECOTA hasn't done very well in recent years, and the best systems have been ZIPS and MARCEL. But isn't MARCEL just a weighted average of the last three years? If that's one of the best systems, what's the benefit of the added complexity in the other systems? Also, I seem to recall that last year someone studied the accuracy of all of the systems over the last 5 years, and found that, while overall they predicted hitters much more accurately than pitchers, the projection systems still didn't do a great job of predicting the hitters.

As Walt suggests, maybe the really useful part of systems like ZIPs are the estimates of the variance for each player. In fact, maybe Pecota and ZIPS would be more useful if they simply didn't provide the mean projection line.

On the scouting side, he's outflanked by BA and by Goldstein at BP; both seem to have better contacts, and both do a much better job of funneling the info from those contacts to their product (very rarely do you get an insightful scouting quote or comp from Sickels). Sickels' book does offer comprehensiveness; he should thank his lucky stars, though, that BA doesn't have a decent search function at their site, or else he might already be obsolete, at least in my view (I mostly used Sickels as a way to get the conventional wisdom on a multitude of guys that I wasn't familiar with yet; the increase in blogs and sites like Scouting the Sally has minimized this aspect of Sickels' appeal).

On the stats side, Sickels is outflanked by fangraphs and minorleaguesplits and countless other free resources.

I don't agree. Scouting and analyzing minor leaguers is still far more art than science, and I think most people want as many knowledgeable people providing input as possible. I suppose it's possible that at some point in the future the market will be saturated and people like Sickels will need to provide something really unusual in order to attract readers, but I don't think we're close to that point. IOW, SIckels isn't outflanked by BA or BP because most fans welcome another voice and don't seem the different services as redundant.

As long as we're in Andy Rooney mode, another thing that's started to bug me is the current "sabermetric" view of team construction and strategy. It seems like the BPro/success cycle stuff has become fairly mainstream, and is now repeated mindlessly lots of fans and writers. Constantly citing the $ values from WAR is part of that phenomenon (and I'm probably guilty of that as well). Another interesting trend is that many fans seem to have done a 180 on team spending. 10 years ago (maybe even 5), most fans only wanted their teams to spend as much money as possible on players; more money = a better team. And the stated or unstated assumption was "hey, it's not my money." Now it seems like every fan (an exaggeration, I know) is constantly bemoaning the fact that their team spent $5M on their new 2b when they could have gotten another guy for $3.5M. Of course it's ridiculous to demand that your team spend millions on players, regardless of the long term consequences or the team's profitability, but it's just as ridiculous to fret over a few million on a short term contract.

I am suggesting taking it out of the realm of the personal, and to look at the heuristics involved that lead to certain personal feelings.

It sounded to me as though you were suggesting that he didn't have a right to his fatigue. Guess I misread it.

A "poll" of 600 PA would produce a standard error of about 4% -- we're trying to detect effects the equivalent of 1/2 of a seasonal standard deviation. (which is why you need about three years of data to begin to be able to say anything.)

This is the crux of the matter for me. We (or they, if you're not one of us) will debate endlessly whether a team will get good value out of a player's contract based on a bunch of stats that need 3 years' worth of data to be significant....but in 3 years, that player's talent level will change. In my mind, we end up arguing things that we trust as knowable, yet really aren't (and yea, I'll quote WAR from fangraphs as much as the next guy. Doesn't mean I'm completely comfortable with it).

As Walt suggests, maybe the really useful part of systems like ZIPs are the estimates of the variance for each player. In fact, maybe Pecota and ZIPS would be more useful if they simply didn't provide the mean projection line.

Something I'd like to see in reviews of projections systems is not just how far off the mean projections were, but whether their collapse/breakout rates were on target (i.e. did 10% of players who were given a 10% breakout projection really have such a season?). For some reason, I'm reminded of the 2005 BP which gave Zack Greinke a 0% collapse rate.

I went through the same thing Sickles did in grad school. I had my MA in history almost finished, with just the thesis to write when I quit because I had gotten sick of the topic and of academia in general. It took a few years away from school and a career in sales to make me fall back in love with history. Sometimes a sabbatical is not the worst thing to allow to take a breather.

Now it seems like every fan (an exaggeration, I know) is constantly bemoaning the fact that their team spent $5M on their new 2b when they could have gotten another guy for $3.5M. Of course it's ridiculous to demand that your team spend millions on players, regardless of the long term consequences or the team's profitability, but it's just as ridiculous to fret over a few million on a short term contract.

I agree with this, especially when there's the suggestion that we can predict with such accuracy. I don't believe anyone can properly assess the dollar value of players with an accuracy of $1-2M.

I also can't seem to trust defensive metrics as much as offensive ones. Move a hitter to a different team in the same league, and his matchups are still similar... he may face Tim Hudson a few more times in the NL East and Brandon Webb a few more times in the NL West, but there are plenty of the same match-ups. Move a fielder to a different team, and he's defending a set of entirely different pitchers, and sharing the responsibilities with the same men on either side.

Luis Castillo is a pretty good example of a player that I feel is more harshly punished than is reasonable for his poor defense, in part because of the nature of the staff the Mets have. Santana, Maine, and Perez all have significantly lower ground-ball to fly-ball ratios than league-average. Castillo would be a much more serious defensive liability on, say, the Braves with Hudson and Lowe (who induce far more ground-ball outs). On the other hand, the value of his on-base percentage, good baserunning, and poor slugging are going to translate more similarly (not perfectly, of course, but still more similarly) no matter which lineup he's hitting in.

What Walt says:

In other words (with my WAG's)

Variation due to randomness: 40%

Improved predictive value of new stat: 1%

Relative value 40:1

That said the pitch Fx work is groundbreaking and it will be interesting to see if teams use it, or more precisely, if managers and players accept it.

I got disillusioned with academia a few years ago, but I realized (or at least decided; one's view can differ) that "academia" as a construct is far less important than the work I do: the research, writing, and teaching, which I love. What I dislike is the politics, the overabundance of humanities scholars (which fuels the toxicity of the politics), and the impotence of humanities in the face of budget cuts (exacerbated by the overabundance of humanities scholars). None of that has anything to do with the fact that I love the actual work, so I separate the realm of academia as a construct from the actual work, and I can live with that as long as they let me.

That said, in the realm of sabermetrics, I've gotten to the point at which my eyes gloss over at the latest acronymed stat that slightly tweaks or refines measurement of offense or pitching. I remain very interested, however, in the improvement of defensive evaluation, and excited about the batted-ball and pitch speed/trajectory data that's becoming available. If anything, sabermetrics is moving in a direction that's more comprehensive to the humanities-heavy, mathematics-light observer, since those things are purely observations of what exactly occurred on the field, and conclusions drawn from them are as much qualitative as quantitative in character.

That said the pitch Fx work is groundbreaking and it will be interesting to see if teams use it, or more precisely, if managers and players accept it.

What makes you think teams didn't already have that information available to them long before Pitch f/x made it to the public eye? And what makes you think that the Pitch f/x work is groundbreaking? What are you learning that changes the way that you think about the game?

Yes, those are serious questions.

-- MWE

Just wanted to say that I thought posts 8 and 16 were both pretty insightful. I certainly fall into the thought patterns 16 describes.

And what makes you think that the Pitch f/x work is groundbreaking? What are you learning that changes the way that you think about the game?

The fundamental reason it's groundbreaking is that it lets us evaluate players on process instead of results, in exactly the same way that judging a pitcher based on K and BB rates is more reliable than judging him on his hits and runs allowed. Ideally, we can (for instance) get a more finely-grained measurement tool for control and command than BB rate allows. And much in the same way that one can scan lists of FIP-ERA to get a decent idea of candidates for a breakout (buckets of salt here, of course), we may be able to use the new data to identify pitchers who haven't gotten the K/BB rates their stuff deserves.

I'm not sure that it's the sort of thing that's going to change the way we think about the game, but certainly it will change how we evaluate individual players going forward.

And re: teams already using this sort of data -- that doesn't strike me as a useful question to ask, given that nobody here can really know that. Once the PitchFX has been worked through and is more easily interpretable, however, we may be able to identify front offices that clearly don't use it (always picking up players who have fluky K and BB rates, for instance).

What makes you think teams didn't already have that information available to them long before Pitch f/x made it to the public eye? And what makes you think that the Pitch f/x work is groundbreaking? What are you learning that changes the way that you think about the game?

The short answer to your three questions is that (1) I don't think pitch Fx has been around all that long so I don't think teams have had a lot of time to work on it and pitch Fx is being improved each season. (2) Because it is new, longer answer below, and (3) nothing other than wondering if science will take some of the randomness out of the game.

My longer answer is that my thoughts when I wrote those few sentences were that teams now have statistical information available to them which might replace anecdotal information. Ten years ago teams relied on advance scouts saying "this guy can't hit the outside pitch". Five years ago an intern sitting in front of a TV/Video combo charted those pitches to come up with the same answer. Now pitch Fx could say he can't hit an outside pitch over 92 mph but he kills outside pitches at 88mph. And the information is more accurate than the advance scout or the intern. In my opinion accurate statistical information is more valuable than anecdotal information and I don't know, but I wonder, what smart minds can do with better information.

I don't know what exactly front offices are doing with the information but I worry that if every team has similar information will it take some of the uniqueness out of the game? I don't think about the game differently today but I hope the game doesn't develop into thirty teams with the same information and approach. Was baseball better when teams didn't all have the latest information and analysis? There was more variability, in my opinion, in the abilities of the front offices and managers twenty years ago.

Actually it is a little of both. I'm at least somewhat burned out right now.

Well, you did just finish the book, I'd be surprised if you weren't a little tired right now.

My longer answer is that my thoughts when I wrote those few sentences were that teams now have statistical information available to them which might replace anecdotal information. Ten years ago teams relied on advance scouts saying "this guy can't hit the outside pitch". Five years ago an intern sitting in front of a TV/Video combo charted those pitches to come up with the same answer. Now pitch Fx could say he can't hit an outside pitch over 92 mph but he kills outside pitches at 88mph.

My point is that teams didn't need Pitch f/x to tell them that. I think that, like most people, you believe that just because you doesn't have the information, no one has it. That's untrue; teams have access to a lot of information that we don't, and I think you seriously underestimate the information that scouts provide their teams. It's not just stopwatches and radar guns.

-- MWE

My point is that teams didn't need Pitch f/x to tell them that.

I don't have strong feelings about f/x one way or the other, but who cares if the teams had access to this information years ago? The fans didn't, and having access to it now potentially enhances our understanding of the game. The fact that the information already existed somewhere else in the universe is irrelevant.

Mike - if you are arguing that Pitch f/x data, or stuff similar to that, doesn't have uses in terms of evaluating players and that teams aren't using it, I'm pretty sure you are incorrect. Why not talk to Mike Fast, Josh Kalk and many of the other Pitch f/x specialists providing information to teams?

I think someone noted the other day that PECOTA hasn't done very well in recent years, and the best systems have been ZIPS and MARCEL.

Tango has them as ZiPS and CHONE.

I encourage folks between the ages of 35-45 to note the following:

--persistence in all matters will aid in managing through those periods in which you find yourself disengaged
--you are not bored, you have become resistant to learning. See above
--and most importantly, beware the easier path, it never is

Given that most of you are with those unwilling to smack you the side of the head this post will have to suffice.

Somebody put the old man back in his closet.

Somebody put the old man back in his closet.

Why? It's not like I am a threat to stab anyone.......

You would if you could lift your arm that high, and you know it. Plus, violent attacks on umpires notwithstanding, I don't smell constantly of menthol.