Fangraphs article on why their projections hate us

[24fps]

14 minutes ago, Can_of_corn said:

They aren't?

Are you suggesting that ignoring their internal projection and ab-libbing the O's higher would make them more credible?

I'm suggesting that common sense indicates that their internal projections might warrant a good, long look in light of results.  That and at least half the smoke and a good two-thirds of the mirrors be jettisoned if keeping an audience engaged long enough to show off the authors genius happens to be a goal.  Devising highly complicated theories these days is a tough sell in an environment where attention spans are measure in milliseconds.  Why on Earth would anybody devote the time figure out why there might be some value in this analysis when actual result already undermine it?

[Flash- bd]

Fangraphs strikes me as the kind of predictions and statistical analysis that Nassim Taleb would absolutely destroy. Random thought, I know. 

[Pruke]

Or it fair to say you can’t project clutch hitting without looking at actual historical statistics? Seems they just don’t have the data with the exception of the past few years and a group of young kids who are proving them wrong. Maybe they value higher in the future 

[Hallas]

Not really sure why Fangraphs doesn't like our starting pitching.  I don't think their projections are fully incorporating their pitch-modeling stats (stuff+, control+, pitching+)  My best guess is that they consider our pitching to be untested, so they're regressing to the mean.  Which isn't all that fair.

[dystopia]

Fangraphs hates the Orioles. Just accept it. 

[dystopia]

7 hours ago, Can_of_corn said:

They aren't?

Are you suggesting that ignoring their internal projection and ab-libbing the O's higher would make them more credible?

Ah yes, the resident Fangraphs fanboy chimes in…

You’d think FG embarrassing themselves like this would get them to rethink their model, not “make exceptions”. 

[Daddy-O's]

I read it and got lost a couple of times, but I think the gist of the article is Fangraphs is right and the Orioles ar wrong.

[btdart20]

8 hours ago, Flash- bd said:

Fangraphs strikes me as the kind of predictions and statistical analysis that Nassim Taleb would absolutely destroy. Random thought, I know. 

It's much easier to destroy than to create.  Even moreso with the benefit of hindsight.  Nearly every model will have flaws.  Economic, political, batting average...  

The article doesn't seem to have any clue as to why the model and reality were so far apart.  That doesn't bode well for being tweaked in the offseason. 

I suspect the SigBot has a better model.  Maybe he found the black swan in the low K%?

[Flash- bd]

6 minutes ago, btdart20 said:

It's much easier to destroy than to create.  Even moreso with the benefit of hindsight.  Nearly every model will have flaws.  Economic, political, batting average...  

 

Very true, that's a very good response to Taleb in general! 

[btdart20]

4 minutes ago, Flash- bd said:

Very true, that's a very good response to Taleb in general! 

Honestly I don’t mind Taleb.  He’s thought provoking.

[Flash- bd]

2 minutes ago, btdart20 said:

Honestly I don’t mind Taleb.  He’s thought provoking.

Yeah, I quite like him, but you're right, that's a good response, and it goes along with similar qualms I have with him. Just because a model has flaws doesn't mean it's worthless, and as you say, it's easier to destroy than create. 

[ledzepp8]

10 hours ago, Can_of_corn said:

They aren't?

Are you suggesting that ignoring their internal projection and ab-libbing the O's higher would make them more credible?

They don't seem to have any reason for how they sort these teams into tiers, or at least they don't explain it. At this point, they're set in their ways. They project things. Things don't go the way they project them. They don't deviate and continue to project those things. They're right and the actual performance of the team is wrong.

I mean, they can do what they want but it's hard to take this kind of stuff seriously. 

[Can_of_corn]

21 minutes ago, ledzepp8 said:

They don't seem to have any reason for how they sort these teams into tiers, or at least they don't explain it. At this point, they're set in their ways. They project things. Things don't go the way they project them. They don't deviate and continue to project those things. They're right and the actual performance of the team is wrong.

I mean, they can do what they want but it's hard to take this kind of stuff seriously. 

Then don't take it seriously.

 

[Say O!]

9 hours ago, Hallas said:

Not really sure why Fangraphs doesn't like our starting pitching.  I don't think their projections are fully incorporating their pitch-modeling stats (stuff+, control+, pitching+)  My best guess is that they consider our pitching to be untested, so they're regressing to the mean.  Which isn't all that fair.

I suspect this is the reason as well, particularly if zoomed in on the second half of the season.  Those pitching metrics would place GRod and Bradish among top SPs in baseball and therefore conclude that pitching is strength not weakness, esp in postseason where rotation is shortened up. 

[Can_of_corn]

Here you go you crybabies. 

https://blogs.fangraphs.com/heres-whos-maybe-going-to-win-the-2023-world-series/

Orioles finish with the second most support, behind the Braves.