How’s the East look now?

[Sports Guy]

27 minutes ago, foxfield said:

I do think it is fair to criticize projections as having some “Yankee” bias. But it should be easy to understand. Their players are big name players who have put up numbers. Repeating those numbers within a certain range makes sense. 
 

Judge for me is an interesting case study. He is amazing when healthy there is no doubt. But he hasn’t been durable and the Yankees insist on playing him in CF which seems to me to speed his decline. 
 

I guess I am saying I don’t think projections are biased, but that they are more likely to inflate the high salary teams because of past performances. 
 

The same is true in reverse for the Orioles with so many young players. The Orioles are more likely to be under rated because they lack the past to project.  Their success in areas like 1 run games should not be repeatable but it’s hard to project player growth for a team that is very young. 
 

I am fine with where the O’s are projected. 

The algorithm doesn’t know if the players are big name guys.

[Moose Milligan]

Some fans are so thirsty for their team to be respected they’re about to duke it out with a computer. 

[dystopia]

14 minutes ago, Moose Milligan said:

Some fans are so thirsty for their team to be respected they’re about to duke it out with a computer. 

I admit I laughed out loud at this post. 

[dystopia]

PECOTA had the O's at 74 wins last year. 27 games off. They were 10+ games off on multiple other teams. These projections are a waste of bandwidth. 

[dystopia]

1 minute ago, dystopia said:

PECOTA had the O's at 74 wins last year. 27 games off. They were 10+ games off on multiple other teams. These projections are a waste of bandwidth. 

Oh and they had the MFY at 98 wins. Lol

[Moose Milligan]

26 minutes ago, dystopia said:

I admit I laughed out loud at this post. 

Good, you should. 

[Frobby]

25 minutes ago, dystopia said:

Oh and they had the MFY at 98 wins. Lol

Perhaps they didn’t count on the Yankees losing 1,750 games to injury.  

[Can_of_corn]

32 minutes ago, dystopia said:

PECOTA had the O's at 74 wins last year. 27 games off. They were 10+ games off on multiple other teams. These projections are a waste of bandwidth. 

Do you realize this weakens your argument that they have a bias against the Orioles?

You've never explained to me why they would have an incentive to purposefully low-ball the Oriole win totals.

[Pickles]

1 hour ago, Sports Guy said:

The algorithm doesn’t know if the players are big name guys.

Well, it kind of does, and that's ironically what people are arguing about here with quite using the right terms.

It's not about "bias."  And it certainly isn't directed at franchises.

It is about how the mathematical formula is written to make these projections.

The formulas universally assume a "regression to the mean."

Well, that phrase has a whole different meaning to different players at different stages in their careers.

A guy like Stanton for instance, if you assume he's just his "average" self for totality of his career, well that's probably pretty rosy imo, but that's a regression to his mean.

Older guys with track record, even as they're breaking down, will have higher projections than younger guys with more upside who haven't performed yet, or have only been performing for a short time.

Guess which team has a lot of older guys with longer track records (who make more money)?

Guess which team has a younger team with less track records?

The projections aren't "biased."  They're just flawed.  But hardly imperfect.

I'd love to bet an over on the O's at 87 wins.

[Sports Guy]

4 minutes ago, Pickles said:

Well, it kind of does, and that's ironically what people are arguing about here with quite using the right terms.

It's not about "bias."  And it certainly isn't directed at franchises.

It is about how the mathematical formula is written to make these projections.

The formulas universally assume a "regression to the mean."

Well, that phrase has a whole different meaning to different players at different stages in their careers.

A guy like Stanton for instance, if you assume he's just his "average" self for totality of his career, well that's probably pretty rosy imo, but that's a regression to his mean.

Older guys with track record, even as they're breaking down, will have higher projections than younger guys with more upside who haven't performed yet, or have only been performing for a short time.

Guess which team has a lot of older guys with longer track records (who make more money)?

Guess which team has a younger team with less track records?

The projections aren't "biased."  They're just flawed.  But hardly imperfect.

I'd love to bet an over on the O's at 87 wins.

Stanton’s ZIPs projection is .9 WAR. 767 OPS.

[Frobby]

32 minutes ago, Pickles said:

Well, it kind of does, and that's ironically what people are arguing about here with quite using the right terms.

It's not about "bias."  And it certainly isn't directed at franchises.

It is about how the mathematical formula is written to make these projections.

The formulas universally assume a "regression to the mean."

Well, that phrase has a whole different meaning to different players at different stages in their careers.

A guy like Stanton for instance, if you assume he's just his "average" self for totality of his career, well that's probably pretty rosy imo, but that's a regression to his mean.

Older guys with track record, even as they're breaking down, will have higher projections than younger guys with more upside who haven't performed yet, or have only been performing for a short time.

Guess which team has a lot of older guys with longer track records (who make more money)?

Guess which team has a younger team with less track records?

The projections aren't "biased."  They're just flawed.  But hardly imperfect.

I'd love to bet an over on the O's at 87 wins.

I think you are making some assumptions on how those systems work that are incorrect.  These systems f don’t just take an average, they look at age, trend line, comparable past players and other data.  For example, ZiPS’ methodology is described here.

[Il BuonO]

5 hours ago, dystopia said:

Pretty much everyone who isn’t a blind Yankee homer has the O’s as the favorites for the division at this point. 
 

The O’s will never be respected by national media or these so-called “projection” outfits. It’s just the way it is. 

False. The Yankees are the odds on favorite to win the East right now at +150.
 

The Orioles are +225, the Blue Jays are +400. 

[foxfield]

1 hour ago, Frobby said:

I think you are making some assumptions on how those systems work that are incorrect.  These systems f don’t just take an average, they look at age, trend line, comparable past players and other data.  For example, ZiPS’ methodology is described here.

This is all I was referring to regarding the Yankees. I don’t think there is a bias, but their players are usually more established with history of numbers. 

[Pickles]

2 hours ago, Sports Guy said:

Stanton’s ZIPs projection is .9 WAR. 767 OPS.

You caught me in a lazy citation, but I believe my larger point holds true.

Especially among pitchers.

[Pickles]

1 hour ago, Frobby said:

I think you are making some assumptions on how those systems work that are incorrect.  These systems f don’t just take an average, they look at age, trend line, comparable past players and other data.  For example, ZiPS’ methodology is described here.

Oh I didn't mean a literal "average."  I just used that as shorthand.

The projections are going to "favor" guys who have a "track record."  That's why it's going to skew winning projections to teams with more "established" players.

And if anyone doubts it, I posit the simple proposition: I'll take the over on the O's winning 87 games.  How much do you want to bet?