2012 projections - Wisdom of the crowds

[Akston]

I have been doing this for a few years, figured I would post it in case anyone finds it valuable.

Essentially, the premise is that no one "expert" projections will be as valuable as if it you aggregate and average the projections from multiple "expert" sources. So I downloaded a bunch of free projections from various non-schmoe sources, averaged them, and poof - a new projection source.

So far, I have 6 sources on this: https://docs.google.com/spreadsheet/ccc ... VdCVjhMRFE

I still need to add ESPN, CBS, full ZIPS projections, and Marcel (if they ever get released - for some reason I thought I heard that he wasn't doing Marcel's anymore), and update some of the older ones I downloaded. I'll update it periodically over the next month as the projection systems come out.

And I wouldn't put much weight on the listed "value" metric. It's something that I use for my own specific league (which uses OPS and has some other wrinkles), but I kept it in there just so I had another way to order the list beyond just alphabetically or by PA/IP.

[Grey]

Interesting... Thanks for doing this!

[Paper Tiger]

I'd love to see the finished product. I did a weighted average of multiple projections one season but it was a ton of work. I did, however, win with it.

[Beardawg]

this website does all of that work for you

http://www.fantasypros.com/mlb/rankings/overall.php

Grey, why aren't they including your rankings?

I like that it lets you select which of their experts you use in the composite.

[Akston]

this website does all of that work for you

http://www.fantasypros.com/mlb/rankings/overall.php

Grey, why aren't they including your rankings?

I like that it lets you select which of their experts you use in the composite.

That's an awesome site, but it does rankings, not projections. What I did was aggregate and average statistical projections. Though this aggregate ranking is pretty valuable too.

[Grey]

Yes, his has projections.

[Paper Tiger]

I don't give props very often but I used your "value" metric for a standard 5x5 roto mock I did last night. It wasn't weighted for a standard league but it did amazingly well overall. More importantly, it allowed me to better focus on players I might otherwise overlook in the later rounds as true value picks. I then took my team and applied my personal projections to them after the fact. My average came out a little low (which is to be expected), but I projected for more runs, homers, ribbies, and steals than any other draft I've done so far (this was my 15th mock of the year).

It wasn't a tremendously competitive draft, but I believe there is real value here. I hate to even ask given the amount of work you've already done (and are still doing), but could you create a value metric specific to standard 5x5 roto? I'd love to take it for a spin on some mocks and see what kind of returns I get.

[Akston]

Not that hard. Just running an average & standard dev calculation vs. batting average instead of OPS.

Next time I post an update (once the full ZIPS projections come out), I'll add a column for "5x5 value"

[Paper Tiger]

Thanks!

[Akston]

edited to include CBS, ESPN, and Davenport projections (now up to 9 sources). And also included the 5x5 "value" metric.

Still need to update ZIPS once Dan gets around to doing the Pirate projections.