Turn Off Ads?
Page 9 of 9 FirstFirst ... 56789
Results 121 to 123 of 123

Thread: The Value of our 2007 Cincinnati Reds

  1. #121
    Five Tool Fool jojo's Avatar
    Join Date
    Nov 2006
    Posts
    19,139

    Re: The Value of our 2007 Cincinnati Reds

    Quote Originally Posted by SteelSD View Post
    Considering that the following link is a Hardball Times article, I suspect you've read it:

    http://www.hardballtimes.com/main/ar...un-estimation/

    I knew this was heading in the direction of Linear Weights. Here's where a metric like Pete Palmer's Batting Runs (BR) holds an advantage over Runs Created (RC) from 2000 to 2005:

    ...<crickets>...

    Here's where RC comes out ahead:

    Correlation Coefficient: 0.9604 (BR: 0.9539)
    Mean Error: 6.52 (BR: 6.96)
    Mean Absolute Error: 19.2 (BR: 20.1)
    Standard Deviation: 14.68 (BR: 16.71)
    Root Mean Square: 24.14 (BR: 26.15)

    In fact, the only Linear Weights metric that outperforms Runs Created at any point is BaseRuns- and only at the ME level (1.78) and the MAE level (18.8).

    Now, I'm not master mathematician (that's my sister), but when Runs Created produces a higher correlation, a lower standard deviation, and a lower rate of error, and a lower generalized or "power" mean than accepted Linear Weights measurements, that's meaningful. The only Linear Weights metric I've found to allow for close approximation to Runs Created is BaseRuns (BsR) and the only place BaseRuns appears to be more accurate is at the low and high extremes of performance (i.e. where baseball isn't played).

    Over the sample noted, BaseRuns produced the highest percentage of expected Run values within ten Runs of actual output, but BsR also produced the highest pecentage of expected Runs that were off by 60 or more Runs. Double-edged sword there.

    The result is that we have a bunch of metrics that get real close to approximating actual Runs Scored. While I appreciate the work done by the Linear Weights movement, I remain entirely unconvinced that their methodology better represents run value versus Runs Created.

    Different? Yes. Better? No.
    Steel, I'm much more concerned with the run estimation of individuals (as are the second generation run estimators like baseruns and LWTS). I've never argued that RC doesn't work well for team estimation. I've argued that when focus turns to individuals, RC moves to the back of the line while LWTS-derived metrics moves to the head of the class (as Dan's article graphically illustrates). Is the difference dramatic? It depends on which RC formula is used and where the player sits (two factors never really considered when people pull RC data from a website). There are three very popular websites which each quote very different RC values (obviously there are many different versions of RC in popular use). For instance, Dunn has been reported to have accounted for anywhere from 101 to 117 RC in 2007. Obviously a 1.6 win difference is a big deal to begin with and this is before considering which value was derived from the formula that has a lower MSE (which in and of itself is a chore).

    Generally, for quick and dirty, it probably doesn't matter provided the RC formula with the best MSE is used and the player is closer to the league mean than the extreme. I often speak in terms of RC for the sake of discussion. Why is this distinction important then? Well generally speaking because it's tough to justify continuing to use a less accurate method when better ones exist (and as sabermetricians are gravitating toward the better estimators our pet websites, though lagging,will have to follow in order to remain relevant). Also-and probably most important for this discussion- quite often advanced defensive metrics like UZR are criticized for their LWTS component. It's simply not a valid criticism from a run estimation standpoint.
    Last edited by jojo; 11-13-2007 at 06:29 AM.
    "This isnít stats vs scouts - this is stats and scouts working together, building an organization that blends the best of both worlds. This is the blueprint for how a baseball organization should be run. And, whether the baseball men of the 20th century like it or not, this is where baseball is going."---Dave Cameron, U.S.S. Mariner

  2. Turn Off Ads?
  3. #122
    Five Tool Fool jojo's Avatar
    Join Date
    Nov 2006
    Posts
    19,139

    Re: The Value of our 2007 Cincinnati Reds

    Quote Originally Posted by M2 View Post
    That's 91 on top of whatever average is supposed to be. How much that accounts for is anyone's guess. Don't even try to parse through how much the "average" takes from the runs scored and run prevention piles. You'd need a supercollider to figure that one out.
    If the baseline for player evaluation is either replacement level or league average then really it's a moot point-no supercollider required.
    "This isnít stats vs scouts - this is stats and scouts working together, building an organization that blends the best of both worlds. This is the blueprint for how a baseball organization should be run. And, whether the baseball men of the 20th century like it or not, this is where baseball is going."---Dave Cameron, U.S.S. Mariner

  4. #123
    Posting in Dynarama M2's Avatar
    Join Date
    Sep 2000
    Location
    Boston
    Posts
    28,450

    Re: The Value of our 2007 Cincinnati Reds

    Quote Originally Posted by jojo View Post
    If the baseline for player evaluation is either replacement level or league average then really it's a moot point-no supercollider required.
    Not if you're talking about how it relates to total runs scored and what offsets need to be made for the figures you're generating. "Replacement level" is not an alternate universe. Actually, that's sort of the problem with "replacement level" systems. They act like these things exist in a vacuum and that they aren't intimately tied to other facets of the game.

    What is the baseline for replacement level defense? How does that contribute to runs scored? What is the effect of all of the adjustments being made on the plus and minus side? These are straightforward, fundamental, if-you-don't-have-an-answer-you-haven't-done-nearly-enough-work questions.

    If you've got a model for replacement level hitting which skews the game toward hitting, another for replacement level defense which skews the game toward defense and a third for replacement level pitching which skews the game toward pitching, then you're dealing with three alternate universes, none of them connected to the one we supposedly care about. And when you try to combine them, then you're back to the supercollider.
    Baseball isn't a magic trick ... it doesn't get spoiled if you figure out how it works. - gonelong

    I'm witchcrafting everybody.


Turn Off Ads?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

Board Moderators may, at their discretion and judgment, delete and/or edit any messages that violate any of the following guidelines: 1. Explicit references to alleged illegal or unlawful acts. 2. Graphic sexual descriptions. 3. Racial or ethnic slurs. 4. Use of edgy language (including masked profanity). 5. Direct personal attacks, flames, fights, trolling, baiting, name-calling, general nuisance, excessive player criticism or anything along those lines. 6. Posting spam. 7. Each person may have only one user account. It is fine to be critical here - that's what this board is for. But let's not beat a subject or a player to death, please.

Thank you, and most importantly, enjoy yourselves!


RedsZone.com is a privately owned website and is not affiliated with the Cincinnati Reds or Major League Baseball


Contact us: Boss | GIK | BCubb2003 | dabvu2498 | Gallen5862 | LexRedsFan | Plus Plus | RedlegJake | redsfan1995 | The Operator | Tommyjohn25