Most Accurate Player Forecasts?

[Neilage]

This very well has been covered in the forums, and I believe it was discussed on the old Rotojunkie forum.

In the past, I have used the "Baseball Prospectus PFM", "Baseball HQ", "RotoWire Draft Guide", "TG Fantasy Baseball", "Fangraphs FANS/ZiPS/Steamer", etc.

Has there been a post-season analysis to see which of the subscription and software options has been the most accurate?

[Sharky]

Pretty sure it's BP (they've studied as well) but not an enormous difference.

[bryanbutler]

it's been studied to death, and was part of the motivation for marcel the monkey - tom tango's point was that all the fancy prediction algorithms didn't do statistically any better than extrapolation of the past few years with a regression to the mean component.

now, if memory serves, it's been shown that good projections do a bit better job than marcel, but it's not by much.

i think fangraphs had a good article comparing systems last year. i'll see if i can dig it up. or todd will certainly have such analyses at his fingertips...

[bryanbutler]

some links:

short descriptions of various projection systems: http://www.fangraphs.com/library/pri...s/projections/

an analysis of 2014: http://www.hardballtimes.com/evaluat...ction-systems/

an analysis of 2007-2010: http://www.insidethebook.com/ee/inde...icial_results/

note that PECOTA = BP.