I just use the average of all these systems, except for Fangraphs fans because for deep leagues they do not account for players who are close-to or at replacement level.
Really, you can’t go wrong with whatever you choose. The most recent incarnation of Oliver is pretty good, especially with young players, and Marcels actually came off looking pretty darn good considering its simplicity.
Sean got snatched up by a team this off-season, so I believe the common assumption is they want him to keep data like that private. CHONE’s not available anywhere, pay or no pay, which makes me very, very sad.
That said, the other systems are still quite awesome. Marcels is good with established players, and the Oliver system is really the tops with the youngins.
Basically, they don’t give out too many details on how they calculate their projections. All I got from that page was that they take a weighted average of a player’s performance over the past three years, try to adjust minor league stats for players with less than 400 major league PAs, and regress players based on their expected LD, GB, and FB rates.
Mike D says:
February 16, 2011 at 4:27 pm
I just use the average of all these systems, except for Fangraphs fans because for deep leagues they do not account for players who are close-to or at replacement level.
ectomobile says:
February 16, 2011 at 4:36 pm
Is there anywhere to find the accuracy of each system from previous years?
Steve Slowinski says:
February 16, 2011 at 4:40 pm
The best you’ll find is the study I linked at the bottom of the article. Tango just did some excellent research: http://www.insidethebook.com/ee/index.php/site/article/testing_the_2007_2010_forecasting_systems_official_results/
Really, you can’t go wrong with whatever you choose. The most recent incarnation of Oliver is pretty good, especially with young players, and Marcels actually came off looking pretty darn good considering its simplicity.
ectomobile says:
February 16, 2011 at 4:48 pm
Great article – thanks Steve!
Justin says:
February 16, 2011 at 8:53 pm
Wonderful article Steve! Why isn’t CHONE available anymore? Does Sean charge for it?
Steve Slowinski says:
February 16, 2011 at 9:11 pm
Thanks! I’m glad you enjoyed it.
Sean got snatched up by a team this off-season, so I believe the common assumption is they want him to keep data like that private. CHONE’s not available anywhere, pay or no pay, which makes me very, very sad.
That said, the other systems are still quite awesome. Marcels is good with established players, and the Oliver system is really the tops with the youngins.
reader says:
February 17, 2011 at 7:42 am
What’s the deal with the Rotochamp projections included on the projections page? How does that system stack up?
Steve Slowinski says:
February 17, 2011 at 7:48 am
Oh geez, I’d forgotten about them. Here’s a link to their full description: http://www.rotochamp.com/Help/ProjectionsFAQ.aspx
Basically, they don’t give out too many details on how they calculate their projections. All I got from that page was that they take a weighted average of a player’s performance over the past three years, try to adjust minor league stats for players with less than 400 major league PAs, and regress players based on their expected LD, GB, and FB rates.
dREaDs Fan says:
February 17, 2011 at 1:51 pm
So is ZIPS the only one then that takes batted ball (e.g. BAPIP) and DIPS theory into account when projecting?
reader says:
February 18, 2011 at 7:24 am
Thanks — doesn’t sound like they’ll add much unique value, but it’s always nice to have more options.
Will Larson says:
February 23, 2011 at 4:11 pm
Shameless self-promotion: See my articles below
http://www.fangraphs.com/community/index.php/comparing-2010-pitcher-forecasts/
http://www.fangraphs.com/community/index.php/comparing-2010-hitter-forecasts-part-1-which-system-is-the-best/
http://www.fangraphs.com/community/index.php/comparing-2010-hitter-forecasts-part-2-creating-better-forecasts/
domki na mazurach says:
February 24, 2011 at 2:11 pm
Awesome!
Ben says:
June 4, 2011 at 8:03 am
Great article! Just one thing, in the Marcel section you accidentally wrote Tom Tiger instead of Tom Tango. Other than that, great job!