tangotiger

You are wrong. Come over to my blog from time to time, and we'll straighten you out.

You have equations. This one: 1.8*OBP2+SLG2 = 1.12 And this one: -OBP2-SLG2 = -.9 Add the first equation to the second equation and you get: 0.8*OBP2 = .22

In 2007, I asked fans from the blogosphere to give me their forecasts for players on their 40-man rosters. I also had in hand the forecasts for Marcel and seven other forecasting systems (PECOTA, ZiPS, Shandler, Bill James, Pete Palmer, MGL, Chone). The Marcels finished right in the middle, right alongside the Community Forecasts. And there was very little separating these two from the leaders (Chone, MGL, PECOTA I believe). I reported all results and published all data on my blog. To the extent that you want to call Marcel useless for being boring, you have to call all other forecasting systems useless for being {insert adjective}. Maybe some systems are too extreme, and others are too biased against young players. Etc, etc. The main point of Marcel is strictly to provide the minimum competence level against which all other forecasts (including those by the fine people of the blogosphere through Community Forecasts) are judged. And to that end, the results of all forecasting sysems are very dull.

The 2008 balloting is now underway. All participation appreciated. Same link as usual for the main page.

The project: http://www.hardballtimes.com/main/article/with-the-game-on-the-line-i-want/ The players: http://www.fangraphs.com/clutch.aspx

As long as you read things in context, there's no issue. I really can't help it if some people try to read things intentionally out of context!

This is my standard set of words. I think it's apparent by the ordering that "fair" is between "average" and "poor". Furthermore, I have GREAT and POOR as uppercase, so it again is apparent that those two are the boundaries. No one will mistake "fair" as somehow being between Average and Good, or being below Poor, especially the way I have it laid out.

Time to put up or shut up: http://www.tangotiger.net/clutch/

Fans' Scouting Report – Results http://www.tangotiger.net/scouting/scoutResults2007.html Thanks to all who participated.

Current ballot counts: http://www.insidethebook.com/ee/index.php/site/comments/the_2007_scouting_report_by_the_fans_for_the_fans/ Cub support is good, but a bit lower than I expected. If anyone's been waiting to participate, it'd be much appreciated.

Thanks for the support. Much appreciated.

(This post approved by Tim.) ======================================= I am now running Community Forecasts, and participation among hardcore fans is appreciated. http://www.tangotiger.net/community/ I've already taken an initial look at Redsox, Jays: http://www.insidethebook.com/ee/index.php/site/comments/community_forecasts1/ As soon as I get 20 ballots for the Cubs, I'll do likewise.

Yes.

Oh, one of the issues that I'm addressing is the validity of the Fans' opinions (as this has been discussed on my site at length). What I'll be doing is canvassing for "superfans", fans that would be considered amateur scouts, dedicated to baseball, and spending the extra time to collect information. This will form a basis against which the hardcore fans will be compared to. *** I've seen professional scouting reports from MLB teams. One thing is clear is that scouts are overworked, providing limited useful information, when they focus on players for jsut a game or two. For one player, the scout changed his opinion on Francoeur's arm from "average" to "great" in the span of 3 months. Other reports that I saw also had that kind of inconsistency. A more focused look, like the superfan being told to just focus on the Cub infielders for 30 games, would be much more insightful. *** Any suggestions you'd like to present to improve the project, I'd be interested in.

Goal: I've already answered that question. So, let me repeat my answers, and expand some: 1. Collect data from hardcore fans, as a way to represent how a hardcore fan thinks 2. Use this data as a historical reference point 3. Use this data to find similar players on other teams 4. Use this data to compare players across positions 5. Use this data to see how opposing fans think of their own players 6. Compare this scouting data to performance data (like UZR and Dewan) The last one is interesting. UZR correlates year-to-year with an r=.50 (r-squared = .25). Fans to UZR correlates at r=.35 (r-squared of .12). When I run UZR+Fans against next year's UZR, I get an r of .60 (r-squared of .36). If the Fans and UZR were completely independent (that is, Fans are not influenced by the data as UZR sees it), we would have expected an r-squared of .37. If Fans were completely dependent on UZR, it would have remained at .25. We got .36. The Fans are providing additional information that UZR would need so that it has a better population to regress to. As for your points, they are valid. I don't disagree with them. *** 99%+ of the respondents evaluated only one team.