The obvious problem, IMO, with the AL projections this year are that the winners of the Central and West have only 83 and 84 wins. That's much too low to be realistic. The East will be excellent, clearly, but who wants to bet me that both the Central and West won't have winners with at least 87 or 88 victories? Someone almost always emerges, and the "sellers" get REALLY bad in the second half and provide more easy wins to the contenders.

 

Projections are fun to talk about, but are mostly worthless.

 

 

The obvious problem, IMO, with the AL projections this year are that the winners of the Central and West have only 83 and 84 wins. That's much too low to be realistic. The East will be excellent, clearly, but who wants to bet me that both the Central and West won't have winners with at least 87 or 88 victories? Someone almost always emerges, and the "sellers" get REALLY bad in the second half and provide more easy wins to the contenders.

 

Projections are fun to talk about, but are mostly worthless.

 

Exactly.

 

I just want everyone to notice that in the last four years, BP has yet to correctly predict the AL Central winner. So consider it a blessing that they didn't pick us to finish 1st.

Combined projection totals for the Sox and Cubs over the last 4 years:

 

Sox 38 Cubs -45

 

So they have a history of decidedly underestimating the Sox and overestimating the Cubs.

 

I'm so shocked!!!

Like I said on another board, I'd rather wipe my ass with something quilted and two-ply.

Looks like that 5th place prediction is contagious... the guy from NBC sports did a team-by-team analysis and places us dead last, behind even the Royals.

 

http://nbcsports.msnbc.com/id/6841778/?pg=..._Spring_preview

 

Fine with me...

Combined projection totals for the Sox and Cubs over the last 4 years:

 

Sox 38 Cubs -45

 

So they have a history of decidedly underestimating the Sox and overestimating the Cubs.

 

I'm so shocked!!!

 

Not only that, but they have yet to predict the AL Central champion correctly in the last four years.

 

BP is a good baseball publication, but the media attention it attracts for being a crapshoot is a complete joke.

 

Not only that, but they have yet to predict the AL Central champion correctly in the last four years.

 

BP is a good baseball publication, but the media attention it attracts for being a crapshoot is a complete joke.

The thing I find interesting about this model is that it doesn't seem to learn/be adapted from year to year based on results. If a team is constantly missed, or hell, if players are constantly missed, then you'd think adapting the model might make sense. You could publish PECOTA's original model and a PECOTA v. 2.0 where the previous year's performance was compared to expectations and used to recalculate things, and that might be a very interesting exercise.

The thing I find interesting about this model is that it doesn't seem to learn/be adapted from year to year based on results. If a team is constantly missed, or hell, if players are constantly missed, then you'd think adapting the model might make sense. You could publish PECOTA's original model and a PECOTA v. 2.0 where the previous year's performance was compared to expectations and used to recalculate things, and that might be a very interesting exercise.

 

What makes it even stranger is that a basic "updating" feature can be done somewhat simply (e.g., a smoothing forecast) and is probably the simpliest part of what is already a complicated model.