THE BOOK--Playing The Percentages In Baseball

Tango -
Thanks for all your work on this challenge!  Nice to see our Razzball entry finished in the top half this year.

Three quick comments:
1) On the ‘Official Competition’, Razzball is in 8th place for Value but is 5th in points_ct (ahead of 3rd place Marcel and 5th place CHONE). 

So would you attribute CHONE and Marcel’s lower point totals vs. Razzball but higher values (assuming because of more wins and/or 2nd-6th place finishes) to being better or to luck?

2) How come ZiPS isn’t part of the contest?

3) What was the playing time source for CHONE and MARCEL?  Will you be doing the analysis like last year for determining the best source of playing time estimates.  I stuck with BaseballProspectus again this year but will switch to Fantistics or FanGraphs community if they prove better again.

Thanks!
Rudy

I used these as Playing Time:
http://www.tangotiger.net/survey/

(See last link for all players.)

I know Chone used either mine or Fangraphs.  There was huge agreement between Fangraphs and mine, so either would have worked out the same I guess. 

I really can’t imagine NOT using the crowd for playing time.  I mean, those guys are so in tune with their baseball teams, that they know all the ins-and-outs and shifts and depth charts.  Why in the world would I spent 2 or 4 hours trying to figure out anything, and trying to do that every day, when with just the few minutes it takes me to create the survey, I get everyone else to do my job?  This is like the best possible crowd-sourcing project possible.

I *love* the crowd.

***

ZiPS, MGL, PECOTA, Oliver all had some issues with their ballots, or expected ballots. 

ZiPS/MGL only provided forecasts, not lists, and those forecasts were missing key things (saves or rbis) that I couldn’t try to convert that to a list.  I mean I COULD, but I didn’t want to spend more time on this like I did last year.

PECOTA had some serious ID mapping issues.  And those couldn’t get resolved in time.

Oliver never had a ballot submitted, and I think Brian said he had his time limits.

***

As for the details of wins and points, I wouldn’t try to read much into it.  After all, the different competition setups resulted in different orders, and, I think it’s too nuanced to try to figure out why the wins and points didn’t match up, and whether it was luck or not.

I think all you readers should DEMAND that whatever system you are considering using or buying should enter this project.

And, as you can see, I’m open to trying out different setups.  The only constraint is that everyone submit an ordered list by MLBAM ID.  After that, I have no contraints.  I’ll code whatever it is that people want me to code.

I’m all for learning more, rather than looking to crown anyone.  Though, seeing Hanson win this thing twice is definitely interesting.  I think getting the right depth charts is rule #1.

Yes, I like the crowdsourcing too.  Only challenge with it is timing - our site visitors are generally barking for rankings in late February for March drafting where I imagine community forecasts don’t finish up until near opening day. 

And it would be great if all the other projection systems participated.  It doesn’t take much work on their end.

PECOTA had some serious ID mapping issues.  And those couldn’t get resolved in time.

Would it be possible to resolve those now and see how PECOTA ranked with the others?

If it’s feasible and simply a matter of somebody spending sometime unraveling the ID mapping, I’d be willing to try to tackle that.

Right, last March I asked if someone wanted to act as a trustee to one of the systems, I’d be happy to hand that off. 

For example, Dan (ZiPS) and MGL gave me their forecasts.  The trustee would come in, fill in the blanks (Saves for Dan, RBIs, Saves, Wins for MGL), calculate the points, map to MLBAM and give me an ordered list.  And, I did that for several systems last year (2009).  I took great care in doing this.  For example, to estimate RBIs, I would use a player’s historical RBI-HR per Hit, and then apply that to MGL’s forecasted number of hits, and add in HR to get an RBI estimate.  (Actually, since MGL didn’t forecast playing time, I would first need to get the community numbers.) I did some sort of trustee actions for five systems last year.  This year, I decided that I would only do one (Fangraphs Community).

Anyway, no one stepped up for the four systems I mentioned.  If someone wants to do it, I have the submissions from MGL, ZiPS, and PECOTA on my flash drive.

Rules are here:
http://www.tangotiger.net/forecast/rules.html

Damn! There’s always next year.

Does Hanson share his projections publicly?

I’m not interested in doing the work for ZiPS or MGL, but I’d be willing to work on PECOTA if you would send it to me.

Ok, send me an email at tom~tangotiger~net

Hanson: he said it would be posted at Fangraphs.  So, Fangraphs is going to host Hanson (tentatively called Slapshot), Chone, and Marcel, as well as ZiPS, Community, and Bill James.

Tango,
Thanks for doing this.  Also, it’s great to hear that Hanson/Slapshot will be at Fangraphs. 

It would be interesting to participate in a fantasy league in which every owner uses the same forecasts or even the same dollar values.  Then differences in draft strategy, plus trades, injuries, and in-season add/drop moves would determine success (plus luck).

Yes, thanks for doing this Tango.  Peter and Dash, although now in college, are set on making improvements to Steamer and coming back strong in 2011.  In Steamer’s defense, the list was optimized for the official competition (making sure not to draft guys sooner than we’d need to in order to pick them before the typical fan) which may not have worked as well in other match-ups.

I’ll be doing the same analysis as last year to try to see where the strengths and weaknesses of each system lie.  I should have chone, marcel, zips, pecota, steamer, oliver, cairo, baseballhq, cbssportsline, Bill James, MGL and fantistics included.

JEH, if you’re up for sending them I’d love to get Slapshot in the mix as well.  That goes Razzball and all other projectors as well.  Projections lined up with MLBAMs are much preferred, of course.  Hopefully, including MLBAMs with projections becomes the industry standard.

Note: I have ZiPS lined up with MLBAMs if that’s helpful to anyone.

Mike Fast, if you get pecota lined up with MLBAM’s and are willing to send them to me I’d be much obliged.

@11/J.Cross

I don’t have a set of projections that I can guarantee predate Opening Day or you would be welcome to them. 

I do plan on sharing my 2011 projections on Fangraphs, so they’ll be available for a postmortem next fall.

J. Cross -
I could send over the projections for Razzball.  Do you want the actual projections for each hitting/pitching category or just the projected points used in Tango’s contest?

Also, what’s your e-mail address?

Rudy, that’s excellent, the actual projections.  I’m going to see who had the most success projecting playing time and each rate stat. 

jfcross4~gmail~com