You, too, can be a scout
4
|
The Hardball Times found this 1/28/2008 on www.hardballtimes.com [flag] |
Tags:
MLB
Comments (8)
-
-
walshj58 thanks, Chris.
Agree that it would be interesting to see how the Fan's Scouting Report varies from year to year. I'm guessing "not much", but it one should have a look.
Certaintly there is y-t-y statistical noise in my analysis, there's no getting around that given the fairly low number of throwing opps.
-
-
mlentzner Just a quick note on a small detail in your article.
Arm strength and accuracy are strongly correlated in a physical sense. It is much easier to hit a target with a fast moving projectile than a slow moving one. The greater arc of the slower moving projectile is a lot more sensitive to mistakes in the estimate of distance than the fast proj.
Regards,
Matt
-
tango tiger Agreed, great stuff.
I wouldn't be surprised if the Fans would outdistance the actual results in a prediction test. That is, if you take the actual data of 2004-2006, and use the Fans evaluations of 2007 (or 2006, but it doesn't matter which one, since the two Fans datasets will agree so highly to each other) to correlate against sample data of 2007, the Fans may likely beat out prior data. Certainly, if it was only the 2006 sample data, I think the Fans would trounce the data.
Tom
-
walshj58 Thanks, Tom.
You know what? I'd bet on the Fans to beat the analysis in that contest, as well. I haven't figured out the statistical uncertainty on the 1-year arm numbers, but I'm guessing they are larger than what the fans would give you.
-
constancio That's what I was alluding to, as well.
Let's say you had a limited set of data (say, one or two seasons) to assess the value of a player's arm. Would you be better off predicting future runs200 value with the fans' assessment or the existing runs200 information?
I would also bet on the fans.
-
-
JinAZ Really neat to see how tightly the correlations match up at high numbers of opportunities, especially considering the differences in what the two stats try to measure (forgive me if this was mentioned elsewhere, but I didn't see it):
I usually think of the FSR arm ratings as a "pure arm" rating, assessing actual arm strength and accuracy. In contrast, the arm ratings that you (Walsh) provide are measures of how well outfielders prevent baserunners from advancing.
Obviously, those things are bound to be correlated. But the latter has two components: 1) getting to the ball, and 2) chucking it to the right spot on the field as quickly as possible. Arm strength, as measured by the FSR, is only influencing the second of these, while the first is probably more closely tied to one's outfield range.
I think that's why you see some of the fast guys with high residuals in your last figure (Taveras, Soriano), and slow guys with large negative residuals (e.g. Griffey, Anderson). I don't want to overstate this (it doesn't explain all the residuals), but it's likely to be a factor.
Ultimately, of course, what we're usually most interested in with an "arm" rating is precisely what Walsh's arm stat tells us--not so much how strong is the arm, but how good is the player at preventing folks from advancing around the bases. That's what matters on the ballfield... -j
-
baseball chick very interesting
one thing thought that might could matter is that not that many fans evaluate each player - could this be the dreaded small sample size thingy goin on?
also, i personally only evaluate the players in the NL central because they are the players i watch often enough to feel i can be reasonably accurate judging. and i don't know about the other fans, but i am pretty honest about each guy's strength and weakness as i happen to see it, no matter how much i like or dislike a player.
i wouldn't think that there are real too many fanboys/grrrls who are doing the ooooooh jeterroolz type thing here.
Links (4)
Roundup - Great Writing Abounds
Published 1/28/2008 by Dave Rouleau at Baseball Digest Daily
... his 23 major league seasons. The report also compares variations in Clemens' career with those of Randy Johnson, Curt Schilling and Nolan Ryan, and maintains slumps often can be correlated with injuries. "Of the six years that feature Clemens' best ERA margins, two occurred in Boston, after he had been in the major leagues for several years; two occurred in his two years in Toronto; and two occurred after he switched leagues and pitched for the Houston Astros,'' the report said." - You, too, can be a scout . - Mark Townsend, at Bugs&Cranks, ...
Analyzing the Fans’ Scouting Report: Arms
Published 1/28/2008 by Tangotiger (tangotiger@yahoo.com) at THE BOOK--Playing The Percentages In Baseball
... John Walsh gives the Fans’ Scouting Report a once and twice over to see how well the Fans line up to his OF Arms system. Great job by John, and I’d love for someone to take on the Fans’ Speed numbers against their own Speed Score system. ...
And now batting leadoff for the Chicago White Sox ...
Published 1/28/2008 by thewizardsofoz <info@southsidesox.com> at South Side Sox: Front Page Posts
... Mariners MLB.com reporter says "The White Sox and Angels have expressed interest in the veteran slugger [Carl Everett], but he remains unsigned." Everett currently is on the roster of the Long Island Ducks of the Atlantic League of Professional Baseball. Seems Kenny can't let go! The latest on the Erik Bedard to Seattle, Adam Jones to Baltimore trade. Mike Pindelski on 2008 Paulie. THT's John Walsh looks at Tango's 'Fan Scouting report'. And Tango's Sabermetric Archive Links (like BaseBoogle (aka ...
The Fans Speak (Through Tangotiger)
Published 1/29/2008 by Cubnut at A Hundred Next Years
... Cub players received grades. The most graded player was Jason Kendall (rated by 43 fans); only 11 fans completed ballots for Daryle Ward. Felix Pie, Derrek Lee, and Ryan Theriot got the highest marks, while the aforementioned Ward, Michael Barrett, Cliff Floyd, and Matt Murton were down at the bottom. Overall, I would say this fan consensus squares very much with how I would have rated the players. Interesting stuff, as tangotiger's always is. (Also, right here, The Hardball Times compares the Fan Scouting Reports with some of their own ...
Really great work, John!
For what it's worth, I would be even more interested in whether or not the scouting scores actually predict performance. For example, how well do the 2006 scouting reports predict 2007 performances, and so on? Otherwise the interpretation of the correlations are not clear to me - it's possible the scouting scores reflect observed performance rather than the actual arm 'tool'.