The stupidest thing I've ever heard.

The stupidest thing I've ever heard.

A buddy who now smokes a pack a day told me he started during his first "winter" in Florida because he missed seeing his breath in the air. But it's close.

I do things quietly, don’t want anybody to find out, because it’s the only way to have the players’ trust.

You know, when I don't want people to know things, the first thing I always do is talk to a reporter on the record. It's foolproof.

In 2012, the Red Sox massively underperformed both their individual player projections and their underlying expected runs created / runs allowed. This followed a 2011 where the team also significantly underperformed expected wins.

If you want to explain what went wrong with the Boston Red Sox in the past few years, you can throw up your hands and talk about bad luck, or you can start looking at issues of clubhouse culture and management and other problems that could cause such severe underperformance.

I hear the Fenway peanut vendors and the Fenway hot dog vendors aren't even on speaking terms, and the giant foam finger vendors take 25 different cabs at the end of a game. No wonder the team fell apart last year.

I wonder if calling the owner to have meeting without the manager impacts clubhouse chemistry?

In 2012, the Red Sox massively underperformed both their individual player projections and their underlying expected runs created / runs allowed. This followed a 2011 where the team also significantly underperformed expected wins.

Maybe their players just weren't that good. Projections systems aren't truth.

If you want to look for a cause, I'd look at systematic talent mis-evaluation by the front office, and a horrible injury management/treatment.

@4, yes, but this is a pretty self-serving point of view. I think that, more than anything, is the problem people have with it.

It is true that the 2012 management team did not have great chemistry. You had an ownership group install a GM who then got overruled by its pick for his manager, and then proceeded to throw that manager under the bus when it turned bad.

AGon is correct. But also: injuries, bad luck, some player pinheadedness.

Maybe their players just weren't that good. Projections systems aren't truth.

If you want to look for a cause, I'd look at systematic talent mis-evaluation by the front office, and a horrible injury management/treatment.

Projection systems aren't truth, but one wouldn't expect that projection systems would fail on one team in particular. It's not like ZiPS knows who the Red Sox front office is. If the front office is mis-evaluating talent, it's in a mysterious way that projection systems didn't pick up on either, and that few thought was the case a priori. The buck stops at the front office, so they certainly deserve blame for whatever happened (especially so with regards to injuries), but I don't think many people thought this wasn't one of the better teams in the league going into 2011, and even 2012.

Do you think that ZiPS, CAIRO and Marcel all systematically mis-evaluate talent, and if so, how?

He's all yours, L.A. 42 walks a year and a slugging percentage in the mid .400's for just $22 million. All that and a whiny attitude, too. Sure hope the Dodgers don't play too many Sunday night games.

#11 - Oddly, I just posted this piece in another thread earlier today. Basically, while it would be nice to have more data, there's a pretty clear suggestion that PECOTA (probably along with other projection systems) tends to overrate the stat-oriented teams, suggesting that both are missing something that's important in putting together a good team, like scouting information or something to do with "chemistry."

The brief case of Adrian. (Relax with Truedi)

That intro should come with the voice of Charlie Minn, slathered with FM echo, intoning "Deeeeep cuts..."

He's all yours, L.A. 42 walks a year and a slugging percentage in the mid .400's for just $22 million. All that and a whiny attitude, too. Sure hope the Dodgers don't play too many Sunday night games.

Yeah, who needs that bum when we got Mike "Iron Man" Napoli.

Even if it was poorly articulated and overly sincere, I don't give a damn-- Gonzalez turning around the criticism he saw there and saying that the Red Sox organization are deceitful wolves is awesome.

Projection systems aren't truth, but one wouldn't expect that projection systems would fail on one team in particular. It's not like ZiPS knows who the Red Sox front office is.

ZiPS doesn't know who the Red Sox are, but the Red Sox know about projection systems. If the Red Sox are believers in ZiPS or something ZiPS-like and uses it, they will suffer from any bias in the system.

A team with other beliefs ("we like strong facial expressions!") will have other biases, but they won't be correlated with the published projections.

Well, top management gave him $20M a year to play baseball. Sounds like an OK relationship to me -- he wants a hug, too? Or maybe he is saying it was Valentine's fault that he stunk last year?

Basically, while it would be nice to have more data, there's a pretty clear suggestion that PECOTA (probably along with other projection systems) tends to overrate the stat-oriented teams

This is interesting, but the Red Sox have underperformed projections by a lot more than a couple games a year recently. The Sox were stat-oriented from 2003 to 2009, too. What's happened in the last few years is more than this.

To be clear, I think that the three largest causes of the recent unpleasantness are (1) talent mis-evaluation, (2) misdiagnosis and poor treatment of injuries, and (3) bad luck. But I think the problem is big enough to include a (4) bad clubhouse culture which has led to player underperformance.

It's also true that you can't easily disaggregate these issues - for example, the bad clubhouse culture was partly a function of dysfunction higher up the chain (see, eg, the Valentine hiring debacle), which was one of the causes of the talent mis-evaluation. As we learned today, the medical and training staffs were at each other's throats, with players caught in the middle, and the front office not only failed to intervene effectively but took sides as well.

The two major causes of the bad outcomes on injuries are the dysfunction on the training staff, and a screwed-up clubhouse culture in which players hide injuries or take them only to favored members of the training staff. It got all ###### up.

I do think Gonzalez is being self-serving, but I would bet that when he thinks of "poor chemistry", he's thinking of a lot of stuff like the above, which I am confident really did cause the Sox to lose more games than they should have.

"Maybe their players just weren't that good. Projections systems aren't truth.

If you want to look for a cause, I'd look at systematic talent mis-evaluation by the front office, and a horrible injury management/treatment. "

Sure. But why did they systematically misevaluate talent, and / or badly manage injuries? Incompetence? Or perhaps, infighting, no means of resolving disagreements, amongst those whose jobs were managing injuries? IOW, management had poor chemistry.

F.O. disagreements about what players to sign, about what doctor to use, whether thoroughly the MDs or the trainer has chief medical authority, thoroughly botched managerial hiring, a rookie GM with no authority, and LL looking to up the 'sexiness quotient' of the product.

I haven't the slightest idea what AGon is talking about.

F.O. disagreements about what players to sign, about what doctor to use, whether thoroughly the MDs or the trainer has chief medical authority, thoroughly botched managerial hiring, a rookie GM with no authority, and LL looking to up the 'sexiness quotient' of the product.

I haven't the slightest idea what AGon is talking about.

You forgot having a pitching coach who did not communicate with the manager or the other pitching coach.

Reportedly Gonzalez himself was ferrying messages between the manager and different coaches who weren't on speaking terms.

Projection systems aren't truth, but one wouldn't expect that projection systems would fail on one team in particular. It's not like ZiPS knows who the Red Sox front office is. If the front office is mis-evaluating talent, it's in a mysterious way that projection systems didn't pick up on either, and that few thought was the case a priori. The buck stops at the front office, so they certainly deserve blame for whatever happened (especially so with regards to injuries), but I don't think many people thought this wasn't one of the better teams in the league going into 2011, and even 2012.

Do you think that ZiPS, CAIRO and Marcel all systematically mis-evaluate talent, and if so, how?

I think ZiPs, CAIRO and Marcel only include what they include. They're 100% based on what happened on the field. They don't have scouting input, medical diagnosis input, etc. ZiPs would have thought Lou Gehrig was a pretty good pickup for 1939.

If a team was bad at scouting, I would expect them to systematically under-perform projections. If a team was bad at evaluating players' health, or keeping their players healthy, I would expect them to systematically under-perform projections.

But the Red Sox underperformance of projections was caused at least as much by poor play by guys who had been on the club in previous years. If all that went wrong was free agents / new acquisitions sucking, the Sox would have made the playoffs in 2010 and 2011. Mis-evaluation doesn't get you close to all the way there.

It's a significant piece of the puzzle, but it's only one piece.

Sure. But why did they systematically misevaluate talent, and / or badly manage injuries? Incompetence? Or perhaps, infighting, no means of resolving disagreements, amongst those whose jobs were managing injuries? IOW, management had poor chemistry.

Either way it's incompetence; either incompetence in evaluating players, or incompetence in managing people. The front office isn't like the MLB roster; you don't have to live with people that can't get along and work together.

Reportedly Gonzalez himself was ferrying messages between the manager and different coaches who weren't on speaking terms.

I have a hard time understanding how any serious adult could behave like this. I can certainly understand avoiding someone with whom you'd rather not interact. But when communication is necessary, I would feel completely ridiculous using a messenger in this teenage drama queen way.

I have a hard time understanding how any serious adult could behave like this.

Yup.

I have a hard time understanding how any serious adult could behave like this. I can certainly understand avoiding someone with whom you'd rather not interact. But when communication is necessary, I would feel completely ridiculous using a messenger in this teenage drama queen way.

I have a hard time understanding how the GM doesn't fire them both immediately upon hearing about it.

Yup.

there's a pretty clear suggestion that PECOTA (probably along with other projection systems) tends to overrate the stat-oriented teams, suggesting that both are missing something that's important in putting together a good team, like scouting information or something to do with "chemistry."

It could be that the stat oriented teams are just rating players very similarly to PECOTA, so that you're both off in the same way.

Suppose you have 100 players and three rating systems. Both systems predict that 50 players will be above average, 50 below. Systems A and B agree about which 50 players will be best, while System C agrees on 25 players, but chooses another set of 25 to be above average. For the sake of argument, let all three systems be equally accurate.

If System A is used to rate teams run according to Systems B and C, it will naturally tend to overrate team B and underrate team C. It's not a question of accuracy, just similarity of methods.

Do you think that ZiPS, CAIRO and Marcel all systematically mis-evaluate talent, and if so, how?

Not directly related, but I think many would have more faith in projection systems if they did more to report on their accuracy. Granted, any of us can compare the projections with the actual end-of-season results, but that seems like real work and I don't do much of that anymore. Besides, it's an easier task for the folks that already have their spreadsheets loaded. On an anecdotal level, everyone knows which players clearly exceeded or failed to live up to expectations, but I haven't seen anything noting how well projections systems did for specific players, teams, ages or years, which would be interesting, and also helpful in evaluating the systems, and perhaps even the players. If that type of reporting is out there, it doesn't seem to have made it to BBTF.

Not directly related, but I think many would have more faith in projection systems if they did more to report on their accuracy. Granted, any of us can compare the projections with the actual end-of-season results, but that seems like real work and I don't do much of that anymore. Besides, it's an easier task for the folks that already have their spreadsheets loaded. On an anecdotal level, everyone knows which players clearly exceeded or failed to live up to expectations, but I haven't seen anything noting how well projections systems did for specific players, teams, ages or years, which would be interesting, and also helpful in evaluating the systems, and perhaps even the players. If that type of reporting is out there, it doesn't seem to have made it to BBTF.

I imagine Dan does a fair amount of this that never really sees the light of day, but I agree it would be interesting to see. I'm still kind of interested in what exactly people think the systems are systematically getting wrong that would effect the Red Sox more than say, the Yankees or other statsy teams, other than "something". I mean, Marcel is basically a regressed weighted average. It's hard to imagine a reason why 2011-2012 Red Sox players specifically wouldn't follow the same general patterns every other player has ever followed. Obviously injuries are a big part of it, but I'm kind of swayed by MCA's argument that chemistry should't be totally ruled out.

I'm still kind of interested in what exactly people think the systems are systematically getting wrong that would effect the Red Sox more than say, the Yankees or other statsy teams, other than "something". I mean, Marcel is basically a regressed weighted average. It's hard to imagine a reason why 2011-2012 Red Sox players specifically wouldn't follow the same general patterns every other player has ever followed.

The system isn't getting anything "wrong". The hypothetical incompetent front office is just assembling a group of players more likely to under-perform than average.

Assume a team with no scouts, who only selected players based on projection systems. They will systematically select players who are more likely to under-perform and less likely to over-perform based on scouting evaluations. Unless scouting has zero value, they have to under-perform the rest of the league who is using both projections and scouts.

The system isn't getting anything "wrong". The hypothetical incompetent front office is just assembling a group of players more likely to under-perform than average.

I think it's a big blind spot for the sabermetric community that it basically ignores player development. This seems to be a huge factor in whether players underperform or overperform - the Cardinals, for one example, appear to be an organization that really gets the most out of its talent.

Most of the time this is a factor for young players arriving at the major leagues, but the Red Sox seem intent on showing that even veteran major leaguers can underperform if you establish a culture where it's perfectly fine to spend games sitting in the clubhouse eating Popeyes and drinking beer.

#11 - Oddly, I just posted this piece in another thread earlier today. Basically, while it would be nice to have more data, there's a pretty clear suggestion that PECOTA (probably along with other projection systems) tends to overrate the stat-oriented teams, suggesting that both are missing something that's important in putting together a good team, like scouting information or something to do with "chemistry."

Interesting. I'm not sure we have the ability to separate saber teams from not saber teams, and even if we do I'm not sure how that says anything about stats or chemistry or any specific cause. It could just mean the Indians are bad sabermetricians.

The article you linked is also only looking at a certain small sample of data and there will naturally be teams who overperform and underperform by pure chance. Also none of that stuff applies to 2011 and 2012 which makes me think it's just random variation. Check it:

http://www.rlyw.net/index.php/RLYW/comments/the_2011_diamond_mind_projection_blowout_-_pecota_edition
http://www.rlyw.net/index.php/RLYW/comments/the_2012_mlb_projection_blowout_-_pecota_edition

Rays we're projected to win 87 games in 2011 and 88 games in 2012. They actually won 91 and 90. A's? Projected to win 83 and 77, actually won 74 and 94 (so over-performance in the aggregate). Twins? Projected to win 83 and 71, actually won 63 and 66.

I think it's a big blind spot for the sabermetric community that it basically ignores player development. This seems to be a huge factor in whether players underperform or overperform - the Cardinals, for one example, appear to be an organization that really gets the most out of its talent.

Not sure where any of this is coming from. Sabermetrics hardly ignores player development. One of the tennants of sabermetric leaning analysts is that cheap young players are way more valuable than expensive veterans. It's true that scouting is much more valuable for young players as they have smaller sample sizes and play in weird conditions, but nothing about a sabermetric leaning front office precludes a team from also having a good scouting deparment.

Cardinals? You mean the team who's player development was headed by Jeff Luhnow (who is now the saberyest gm in all of sabertown in Houston) for the past few years? I swear you guys are just making #### up.

The sabermetric community ignores player development?
Isn't it more likely that as an older team the Red Sox were more likely to have players get injured and thereby underperform?

Assume a team with no scouts, who only selected players based on projection systems. They will systematically select players who are more likely to under-perform and less likely to over-perform based on scouting evaluations. Unless scouting has zero value, they have to under-perform the rest of the league who is using both projections and scouts.

I think I get what you're saying, although I think you're putting more faith into scouting than I would, necessarily. Are you saying that a teams' players outperforming of their projections is due to scouting alone? I'm not sure I'd agree with that.

I think I get what you're saying, although I think you're putting more faith into scouting than I would, necessarily. Are you saying that a teams' players outperforming of their projections is due to scouting alone? I'm not sure I'd agree with that.

No. Scouting is just one factor, along with medical care, player development, etc.

All I'm saying is that if your org. is bad at talent evaluation, I would expect you to under-perform projections, b/c there is some value to scouting.

Sabermetrics hardly ignores player development. One of the tennants of sabermetric leaning analysts is that cheap young players are way more valuable than expensive veterans. It's true that scouting is much more valuable for young players as they have smaller sample sizes and play in weird conditions, but nothing about a sabermetric leaning front office precludes a team from also having a good scouting deparment.

Scouting departments and a desire for young players have nothing to do with player development. You're kind of proving my point, since you don't seem to grasp what player development might entail. I'm talking about taking raw talent and turning it into productive major league players, which some organizations are clearly more skilled at than others.

If there has been any sabermetric analysis of this sort of thing, I've missed it, but I'd love to see it.