CSS Button No Image Css3Menu.com

Baseball Prospectus home
  
  
Click here to log in Click here for forgotten password Click here to subscribe

<< Previous Article
Prospectus Q&A: Tim Pu... (03/15)
<< Previous Column
Premium Article Prospectus Today: Logi... (03/10)
Next Column >>
Premium Article Prospectus Today: The ... (03/17)
Next Article >>
Premium Article On the Beat: Weekend R... (03/15)

March 15, 2009

Prospectus Today

Panorama

by Joe Sheehan

the archives are now free.

All Baseball Prospectus Premium and Fantasy articles more than a year old are now free as a thank you to the entire Internet for making our work possible.

Not a subscriber? Get exclusive content like this delivered hot to your inbox every weekday. Click here for more information on Baseball Prospectus subscriptions or use the buttons to the right to subscribe and get instant access to the best baseball content on the web.

Subscribe for $4.95 per month
Recurring subscription - cancel anytime.


a 33% savings over the monthly price!

Purchase a $39.95 gift subscription
a 33% savings over the monthly price!

Already a subscriber? Click here and use the blue login bar to log in.

Recently, a panel on MLB Network's MLB Tonight show had a discussion about the use of statistics in the evaluation of baseball players. The discussion was preceded by a clip that introduced Nate Silver's PECOTA to the audience, explained what it is, and showed how it attempts to project player performance.

Take that in. It would have been unfathomable just a few years ago that a national television show would be devoting a segment to explaining an externally derived system for projecting player performance, and then discussing it. This is a significant step forward for analysis, and whatever we can say about the details of the discussion that occurred, you have to acknowledge that MLB Network has gone out of their way to present something new to baseball fans. The clip even got PECOTA's acronym correct, which is something I generally can't do even at gunpoint.

The discussion that followed was disappointing. Sean Casey and Barry Larkin paid lip service to the value of performance analysis, and then insisted on evaluating players by anything but their performance. Harold Reynolds was all over the place, dismissing the use of stats entirely because of things that happened in his own career, calling stats "ridiculous," then calling out some of the projections presented without presenting reasons why he disagreed with them. Larkin used the term, "a quality .215 [batting average]," which if nothing else gives me a great name for a fantasy team.

Host Matt Vasgersian-a BP reader who frequently would name-check BP on Padres' telecasts-drove the discussion off course by asking the panel to choose between "the guy with the calculator" and "the guy with the straw hat." That is, of course, a false choice. No team ever has to choose, as Dayn Perry so eloquently explained years ago. All three players chose the scout, after which Vasgersian made the key point that a scout sees a guy three times, while his stats see everything the guy has done. The players nonetheless cited the scout's opinion as being more valuable than what the analysis could provide.

My problem isn't with these opinions, which are to be expected. My problem is with the lack of another viewpoint. Setting Vasgersian up as the opposition is inadequate to the task, not because he's not talented, but because it's not really his role and he's not invested in the argument. Asking three former players about the value of performance analysis isn't going to produce an interesting discussion, because players' opinions on this stuff, with rare exception, fall into a narrow range. A real discussion on this issue would involve other voices, and even if those voices were to be shouted down or marginalized for lack of a baseball-reference page, well, they needed to be heard.

It's entirely possible that statheads undervalue the non-performance aspects of being a baseball player. Even though I've come around on the issue a little bit, to be fair, I'm probably at the outside edge of opinion on this. I think that success, or the lack of it, drives chemistry or the lack of it, and I think what a player does on the field means more than what he is off of it by a factor comfortably above 50-to-1. I've always been the guy talking up Barry Bonds or Manny Ramirez because the runs they create mean more than the annoyances that may follow in their wake, or arguing that you don't roster Luis Sojo or, um, Sean Casey because they're more popular than other choices. There are interesting discussions to be had about these decisions and ones like them, and the relative importance of being a player and being a guy, and I've had them. I've won and lost them, and I imagine I'll continue to do so. I'm searching for the right answer.

The other side of this, however, is that players absolutely value non-performance characteristics much more highly than is sensible. This is understandable, because to them the guys around them are not just baseball players, but co-workers, travel partners, drinking buddies, and confidantes. They are around these people 12 hours a day for two-thirds of the year. They're using a completely different set of criteria to evaluate their worth. It's not an additional set; it's a different one. To the extent that these people evaluate each other on their work product, it's on an extremely micro level that doesn't translate well, if at all, to an overall assessment of performance.

There's a distinction here between "baseball players" and "people." The people who occupy the roster, the uniforms, and the clubhouse have a common goal, but a completely different set of challenges and processes than the people who do not. The traits that people have are very important to this group, whereas those traits are largely invisible to analysts outside of that realm. Analysts are evaluating baseball players, their demonstrated skills, their performance, and how those things contribute to wins. The micro-processes, the personality traits, the preparations that are most visible to, and perhaps most valued by, the people in the clubhouse are lost in the numbers .263/.312/.399.

That this tension exists is inevitable and natural, and it's another reason why I will always say that the last person you want evaluating baseball players and teams is another baseball player. Their viewpoint is too micro for the task at hand, they get caught up in the things that they value and lose sight of the bigger picture. What the discussion needed was that other voice, that vested interest in the value of performance analysis to explain the silliness of "a quality .215" and make the beer-and-tacos argument to the players, pointing out that the work product of a player is far more important than how he actually gets there. Right now, MLB Network doesn't have that voice. They have hired enough ex-players to take a run at the NL Central-or maybe to declare Secaucus a nation and play in the 2013 WBC-but they have no diversity of voices.

This is an unexpected hole in their game. Entities such as Sports Illustrated and SI.com, ESPN the Magazine and ESPN.com, and The New York Times have seen the value of informed-outsider analysis and folded this material into their coverage of the game. Performance analysis is part of baseball now, not something to be ignored or mocked, but a critical part of understanding the game. It's mainstream. So for MLB Network to launch with more good left-handed relievers than the AL West has, but with no one who can make the case for Kevin Youkilis' PECOTA projection, is a surprise.

Players bring a lot to a studio show; Baseball Tonight used Reynolds to great effect, getting him out from behind the desk and having him demonstrate and explain the physical aspects of making plays around the bag at second. He knows the game as he played it, he's extremely engaging and entertaining, and you can build a show around him. I want his voice, and Larkin's, and Casey's, and Al Leiter's, but I want them telling me things that they know that I don't. You wouldn't ask Joe Sheehan about the footwork around second base, or Clay Davenport about the effects of different finger pressures on breaking balls. So don't ask Harold Reynolds what he thinks about a PECOTA projection. You need other perspectives.

If the trends in baseball coverage that we've seen in print and on the Web find their way to television, we're in for a treat. A panel having a discussion of the relative values of performance analysis and skills analysis that featured people steeped in both sides of the debate would be a sight to see. It would be great television that advanced the argument, or at least educated the viewers. MLB Network has a tremendous opportunity here, and having done some great things to start-such as the "30 Teams in 30 Days" series and their coverage of the World Baseball Classic-it's not hard to see them taking a leadership role in bringing other voices to their coverage.

Joe Sheehan is an author of Baseball Prospectus. 
Click here to see Joe's other articles. You can contact Joe by clicking here

Related Content:  Analysis,  Mlb The Show,  The Who

23 comments have been left for this article. (Click to hide comments)

BP Comment Quick Links

ElAngelo
(942)

Matt Vasgersian will always be Armenian Tean Heartthrob to me.

Mar 15, 2009 10:18 AM
rating: 0
 
Fresh Hops

It was good to see an attempt at a discussion of advanced metrics and projections. I wish that they'd tried to be nuanced. Both scouts and computers have their limitations, and I wish the discussion had tried to bring this out. If I want to know likely outcomes for an 18 year old half way through his first year at low A, I'm asking a scout. If I want to know the most likely outcome for a 29 year old entering his fourth season as a full-time player at the MLB level, I'm looking to a computer. Pretending as though a one-size fits all approach to player projection is smart is stupid, and that's exactly the discussion that MLB Network had.

Also, there's no misrepresentation that bugs me more than "stats don't take heart into account." This is just false. What stats won't do is tell you whether heart is 20% or 1% of the game. But they do a good job of identifying guys with the right combination from one year to the next.

I'd love it if a smart, charismatic and not too-smug stat guy got a show on MLB network with the theme of putting your money where you're mouth is, i.e. let's make some predictions about reasonably predictable things and see whose doing a good job of predicting. They could do a segment called "Healthy or Hurt" and have people predict DL time for some players and see how well it does against, say, Will Carrol's system. It would be fun. It would also be nice to see people who actually work hard to get right answers being asked for their opinion, instead of guys who earned the right to be heard for doing something other than trying to know what their talking about.

Mar 15, 2009 11:24 AM
rating: 6
 
Richie

"(S)mart, charismatic and not too-smug stat guy". ARFARFARFARFARFARF!!!

I will show my maturity by delving no further into this gold mine of humor you've provided there. Please appreciate the sacrifice, as I hate showing maturity. So little to spare. Anyways, like Joe said, just putting the acronym PECOTA in front of an MLB Network audience is astounding progress. Kudos all around!

Mar 15, 2009 13:18 PM
rating: -3
 
Brian24

It's pretty strange that they would introduce the topic of PECOTA on a show like that but not invite Nate to comment on it.

Mar 15, 2009 11:37 AM
rating: 7
 
RedsManRick

Obviously you're preaching to the choir, but amen. From a marketing perspective, MLB missed a big opportunity to engage in the analytic friendly fan market. Instead we've been treated to a slightly cleaner version of baseball tonight, sans the seasoned reporters such as Gammons.

Given the airtime they have to fill, there seems to be a golden opportunity to give voice to this corner of the market. After the great success of the Fantasy 411 in driving attention to radio/podcast activities on mlb.com, you'd think they'd have picked up on this.

As far as I'm concerned, give Cory Schwartz an hour to do what he wants and it would instantly be the best daily baseball show on TV.

Mar 15, 2009 12:15 PM
rating: 2
 
Wrigleyviller
(883)

Agreed. Heck, given the sheer amount of airtime on the network, they could easily have multiple shows in this vein. Would it be so crazy to have a show dedicated to baseball statistics for beginners or something? Go over one stat per 30-minute episode, maybe cover the history of statistical measurements for that category, etc?

Actually, as someone versed more in stats than scouting, I'd kill to have a scouting-only show on MLB Network. How great would it be to have real-life scouts go over the basics of scouting, or maybe take one player per episode and break everything down? With really detailed, high-def slo-mo video to accompany the scouts, and maybe even a stats ringer brought in to go over the correlation between their scouting conclusions and data from PitchFX or something...OK now I'm just drooling.

Also, I always thought Harold Reynolds was at his best when he showed how to actually do something on the field of play. He should have a whole show where he and other former players go over every little aspect of how to actually do each little part of the game.

Mar 15, 2009 14:36 PM
rating: 4
 
Cory Schwartz

Thanks!

As always I think Joe has nailed the topic here, but let's be fair to MLBN, which has been on the air for only two-1/2 months and the season hasn't even started yet. I'm willing to give the benefit of the doubt that they will eventually find room for other types of programming and/or voices, but right now I would expect they are focusing on their ability to produce game broadcasts and the 8-hour "MLB Tonight" show that is probably going to result in many, many divorces.

So while I think everything that's been said here is valid, let's walk before we run. Remember, three months ago there was no MLB Network at all... we've come a long way baby!

Mar 15, 2009 20:53 PM
rating: 1
 
Hawkeye

I enjoyed your article Joe, but disagree about Reynolds and making a show around him unless it's a show with him only showing things in the field and not doing any analysis.

I can't watch him so I'll take your word for it that he was "all over the place." I think he has an inate desire to be liked so he'll agree with whatever the last person said even if the last person said "Harold Reynolds doesn't have a clue." Plus in his eagerness he chimes in so fast I wonder if he gave what he said an thought.

I have not watched much of the alleged analysis on MLB.com but I love the games.

Mar 15, 2009 12:51 PM
rating: 2
 
Lou Doench

"a quality .215 [batting average],which if nothing else gives me a great name for a fantasy team."

When does Hacking Mass start, because there will be a line forming up for that name....

Mar 15, 2009 13:07 PM
rating: 2
 
Richie

Actually, there is such a thing as "a quality .215" batting average. See Deer, Rob. We all just know that that's the exact opposite of what Larkin meant by it.

Mar 15, 2009 14:20 PM
rating: 4
 
Pietaster07

There is .215 Adam Dunn and .215 Mario Mendoza

Larkin wants Mr. Mendoza I'm guessing

Mar 15, 2009 20:37 PM
rating: 0
 
ccasin

You really buried the lede here. I almost gave up 1/3 of the way in because it seemed like it was just going to be another "why don't people love stats" rant (not that I don't agree!). Your suggestion that they get more stat people is good, though.

Mar 15, 2009 17:06 PM
rating: 0
 
krusch

While I agreed with you 99% of the way, I think there is one part that statistics miss.

We assume that every player is playing to the best of his ability at all times; that Cristian Guzman's .265 is because he's inable to do better. ("take a pitch or two" notwithstanding.) And as such, when a co-worker fails, you know it's because he's doing the best he can and isn't succeeding.

And if Bonds or Jeff Kent is a jerk, well, he's still out there getting his hits, and you can put aside the fact you can't stand him.

However, if the rumors are true that Manny doesn't try hard, and he's faking his leg injuries, etc, or taking three strikes to show up the manager, I really think that *would* be detrimental to the team. I mean, that's not just "poor chemistry", but that's insubordination. And that leads to guys deciding not to dive for balls, not legging out ground balls, etc.

Anyway, Bonds and Manny aren't just piles of stats, they're people, and while the numbers we use to estimate their performances are pretty good, they're not everything. Usually the "not everything" is just a little bit around the margins, but I think sometimes it's a lot more than that.

Mar 15, 2009 17:29 PM
rating: 1
 
Brian Kopec

I've heard this argument time and time again...especially from the hate-Manny-first crowd. I can't believe anyone thinks that Dustin Pedroia gets to fewer ground balls or Mike Lowell swings at more sliders in the dirt or Josh Beckett loses 2 MPH off his fastball because Manny Ramirez is standing in left field. It just doesn't work that way. If there are players who would be spooked by playing alongside Manny, they probably didn't have the personal drive to make it in the bigs in the first place. If you can't hit a curveball because you don't get along with a co-worker, then I suggest you are the one with the problem, not the co-worker.

Mar 16, 2009 07:06 AM
rating: 2
 
cdoyle31

One always get the sense that the false choice presented in these conversations isn't "stats vs. scouts," but "food on the table vs. abject poverty" for the former major league marginalia who are paid to pontificate. Someone should pull Harold Reynolds aside and show him your article as proof that the stat geeks aren't gunning for his job. Let's hug it out, H.R.

Mar 15, 2009 17:32 PM
rating: 0
 
JoeSky60

Great article Joe! I saw the video segment from MLB Network, and was heartened by the setup, and disappointed in the commentary. As you point out, though, what else can you expect from ex-players. What I'd really like to see, and I'm sure it will eventually happen, is a dove-tailing of both sides of the argument, like "Wrigleyviller" suggests, rather than a debate.

Mar 15, 2009 18:35 PM
rating: 0
 
mrflak

This all reminds me of Galileo's battle against the Catholic Church. Galileo is declaring these revolutionary ideas that totally go against the scripture ("The Book" if you will) and the Pope just cannot compromise his version of reality with what Galileo is proposing. After the segment, I can't help but think that Harold Reynolds is the reincarnate of Pope Urban.

Regarding scouts v. stats (AKA: faith v. science) I see it as a false choice. Each gives us some perspective on a player, but left on a desert island where the only way I can get off is to accurately project player performance, I will gladly take a calculator over a straw hat.

What if we extended this argument to other sectors? Perhaps we should direct our stimulus package to the auto industry because they're a "scrappy" group of workers. Maybe we should invest in companies that show "good makeup" despite what their balance sheet reads. This "stats can't tell me nothin'" approach is like judging your stock portfolio on whether or not you like the CEO of the company.

Yes, I'm preaching to the converted, but I love to hear myself type.

Mar 15, 2009 19:51 PM
rating: 0
 
Pietaster07

True, but wouldn't you like to know how honest that CEO is? Or whether those auto workers will work harder to make a better car when under financial pressure?

We need stats AND scouts.

(and IMO we need both science AND faith)

Mar 15, 2009 20:50 PM
rating: 0
 
Bob

I don't see scouts v. stats as a "faith v. science" debate, it's more like qualitative research v. quantitative research. Just as in any other social science, both are important and both have their limitations. But I do agree with you, left with a choice between one or the other, I'll take stats (though for real social sciences, I'd take qualitative research any day of the week).

Mar 16, 2009 05:01 AM
rating: 1
 
Bob

That's great that the MLB network featured PECOTA.

Now if only Nate Silver would make an appearance on Baseball Prospectus...

Mar 16, 2009 04:55 AM
rating: 6
 
Ira

Just to chime in and add another voice to the choir. I also saw that segment as an attempt to try and both introduce something to the general public, namely the idea that baseball seasons can be predicted using numerical methods, and then get the opinions of some ex-players on the subject.

As far as what the ex-players said, the answer is that they said what they've always said: That numbers help you determine what kind of salary you'll get but have nothing to do with winning a ball-game.

What MLBN could have done was, instead of getting 3 ex-players, gotten one ex-player (Reynolds or Eric Davis), an ex-manager (if there are any around) and an ex-GM. That way the ex-player could say how he judges his own game by feel, the ex-manager could explain how he balances scouting reports and statistical analysis to decide playing time and strategy, and the ex-GM could note how he uses both scouting and stats to determine which players to acquire (or get rid of) and mostly stats in salary negotiations.

At the Major league level, Stats will always tell you how someone did, and can be used to predict how you think someone will do. Scouting is used the same way, but much more of it is tilted towards prediction. A statbook can note that a shortstop turned a groundball into an out, and with advanced metrics can even say where that ball was and how hard it was hit, but the scout can tell you whether that player made the play as if it were routine, or at maximum effort. He can gauge how far the player went to get the ball, and how quickly. He can tell you if the player's first step (his reaction time) contributed to his getting to the ball, or if he froze and used his speed to make up for it. In other words, a scout might be more likely to predict how that fielder might do on a ball hit 2 feet farther away than the ball he saw. That info might be important to the organization. But scouts aren't perfect. The player could have a cold that day, or a twisted ankle that he's hiding, or hung over. A really good scout (you would think) would find that out.

The problem with hitting is that I don't think that the nuances of hitting are such that a scout will tell you that much that a batting line won't. I've seen too many guys with UGLY swings who were successful at the major league level, and too many guys with either extremely fast, or extremely fluid swings who just couldn't hit. Thats the kind of thing that can fool a scout. Also the pitcher with the 100 MPH fastball that everyone can hit. Scouts can get fooled by that too. In those cases performance metrics can help.

Mar 16, 2009 09:09 AM
rating: 0
 
wonkothesane1

MLB Network does have an ex-GM and I'm surprised he wasn't part of the discussion. I might be misinterpreting some of his comments, but John Hart seems to be a guy that understands the evaluation of players by quantitative and qualitative means. I'm pretty sure he could have brought a little more substance to the conversation while also eliciting respect from the ex-MLBers in the room.

Mar 16, 2009 16:54 PM
rating: 0
 
Richard Bergstrom

I think, in some aspects, it is a bit unfair to judge the players' reaction so harshly. Even those of us who come here to BP on a daily basis have no clue as to how PECOTA works besides a general understanding of _some_ of the concepts it uses in its projections. A good portion of that has to do with the proprietary nature of PECOTA, but if I was a baseball player and someone asked me about a system that I knew nothing about and whether it was better than a system that I knew something about i.e. scouting, especially with a lack of information on the minutae of how that system works, then I'd still most likely choose scouting because I'm more familiar with it. Even though PECOTA is coming more into the mainstream, you don't have a Peter Gammons or even a Bill James writing often about it, nor can anyone "look under the hood" to see if the theories it bases its predictions on hold merit.

Take the Alex Rodriguez prediction... a relatively unique player in terms of position, ability and performance. PECOTA's projection of Alex Rodriguez, which I read about in Marc Normadin's column on third basemen, seemed unusually low. Soon enough, he goes down with a hip injury. Did PECOTA just get lucky in predicting a severe downturn this year, or was there some statistical indicator suggesting a decline? It's not like I saw Albert Belle or Bo Jackson among his comparables. The way BP comes up with its fielding ratings also merit some concern, especially when some of the comments in BP2009 directly contradict numbers showing a player as being a good or bad fielder. How do we critique and analyze systems that are not open to analysis? If I and others sure can't, how can we believe people not familiar with BP's history, rigorous testing, and methodology will accept PECOTA and other such systems?

On the flipside, Kevin Goldstein in his Future Shock articles has talked about how portions of his analysis of prospects rely on experience, observation, and other qualitative means. Maybe I can't agree with his perspective, especially when the numbers disagree (or in the case of someone like Michael Ynoa who has no numbers), but he does explain what he sees and why he sees it, and to a baseball player more used to watching baseball than reading the backs of baseball cards, a scouting viewpoint is probably more accessible.

Mar 16, 2009 18:18 PM
rating: 2
 
You must be a Premium subscriber to post a comment.
Not a subscriber? Sign up today!
<< Previous Article
Prospectus Q&A: Tim Pu... (03/15)
<< Previous Column
Premium Article Prospectus Today: Logi... (03/10)
Next Column >>
Premium Article Prospectus Today: The ... (03/17)
Next Article >>
Premium Article On the Beat: Weekend R... (03/15)

RECENTLY AT BASEBALL PROSPECTUS
Premium Article League Preview Series
Every Team's Moneyball: Minnesota Twins: Reb...
Premium Article Skewed Left: History Repeats Itself
Premium Article League Preview Series
Premium Article Pitching Backward: Why Relievers Get A Free ...
Premium Article Spring Training Notebook: Cactus League
Prospectus Feature: How the Astros do Spring...

MORE FROM MARCH 15, 2009
Premium Article On the Beat: Weekend Roundup
Prospectus Q&A: Tim Purpura
Prospectus Q&A: Shiraz Rehman

MORE BY JOE SHEEHAN
2009-03-24 - Premium Article Prospectus Today: Curt Schilling Bows Out
2009-03-22 - Premium Article Prospectus Today: Atlanta's All-Anderson Aff...
2009-03-17 - Premium Article Prospectus Today: The Motor City Money Pit
2009-03-15 - Premium Article Prospectus Today: Panorama
2009-03-10 - Premium Article Prospectus Today: Logistics
2009-03-06 - Premium Article Prospectus Today: There's a Breach
2009-03-05 - Premium Article Prospectus Today: The Manny Has Landed
More...

MORE PROSPECTUS TODAY
2009-03-24 - Premium Article Prospectus Today: Curt Schilling Bows Out
2009-03-22 - Premium Article Prospectus Today: Atlanta's All-Anderson Aff...
2009-03-17 - Premium Article Prospectus Today: The Motor City Money Pit
2009-03-15 - Premium Article Prospectus Today: Panorama
2009-03-10 - Premium Article Prospectus Today: Logistics
2009-03-06 - Premium Article Prospectus Today: There's a Breach
2009-03-05 - Premium Article Prospectus Today: The Manny Has Landed
More...