2013 MLB Regular Season running thread

We should also ban chemotherapy. Useless.

 

Your humble correspondent does very little thinking on baseball metrics. Frankly, I don't much give a shit about dWAR or whatever the eff statistic is being bandied about. And quite honestly, I thought OOP would come back and point to dWAR's relative lack of effectiveness/efficiency as compared to the status quo re: defensive evaluation. I was kind of hoping to hear how defensive ability was quantified before dWAR. Instead, in an absolutely stunning development, I got "BRETT LAWRIE!!!!!!!"

I care not a whit about defending one metric or the other. What I care about is pointing out how absolutely juvenile intellectually it is to proffer a quirky result as proof -- not suggestive evidence, mind you, but proof -- that the measure that led to that result is useless. To put this into perspective: Hotelling's T-squared test statistic will lead to an inappropriate rejection of the null hypothesis alpha*100% of the time. If we set alpha equal to 0.05, that means that five out of every 100 "significant" results we get when we're using T-squared is an error. OOP thinks that Harold Hotelling should go back to the drawing board. I think that OOP shouldn't have dropped his statistics class.

 

Jeff Francouer signs with the Giants, per Ken Rosenthal on Twitter:

https://twitter.com/ken_rosenthal/status/354585344643043328

 

Is home runs a flawed metric? Brady Anderson hit 51 in 1996, was he really one of the top power hitters in the 90s?

 

I love how oop acts like he has any earthly idea what dq is talking about.

 

That's good because WAR is,while not perfect by any stretch of the imagination, extremely accurate.



http://www.hardballtimes.com/main/article/what-is-war-good-for/

There's some reasoned support for my previous statement regarding WAR.

Is that really what it proves?
'Statheads' like the newer metrics like UZR and Total Zone better than they like the old metrics like fielding percentage and range factor. 'Statheads' also realize that UZR and Total Zone aren't perfect, and make no pretenses that they are perfect.

http://www.fangraphs.com/library/defense/uzr/

That should tell you something about 'statheads'.

There's a distinct reason for doing this.

Hopefully you realize that there is a difference between hitting .300/.400/.500 with 40 home runs in Coors Field in 1999 and hitting .300/.400/.500 with 40 home runs in the Polo Grounds in 1915 .

This should also tell you something about 'statheads'. Instead of hunkering down and insisting that the advanced metrics were correct and that Brett Lawrie was the best player in baseball the 'statheads' looked into what was going on and fixed it.

What the problem was, incidentally, was that the defensive metrics were giving Lawrie credit for plays made when he was shifted. To the metric it appeared that Lawrie was making plays 100's of feet away from his SS position. That was fixed.

So, yeah. New metrics are flawed. Old metrics were flawed as well, which is why people came up with new metrics. People will continue to come up with new metrics to replaced the older, flawed, ones.

 

That's a good article, especially the attempt to use WAR to predict rather than describe. Some may sniff their nose at an R-square of 0.35, but in the social sciences that's a very, very strong result. To give you a sense for how strong a result that is, suppose you were trying to predict a student's First Year (College) GPA (FYGPA) using that student's SAT score and his/her high school GPA. The most recent report (2008) I could get my hands on regarding these variables' association with FYGPA puts their R-square at around 22%.

 

Retired Trenton Thunder bat dog dies days after retirement.

http://bit.ly/186nt9J

 

Does that make him the Bear Bryant of bat dogs?

 

Translation: You are getting your ass kicked on the overall argument, so you latch on to one small part of it and combine it with personal attacks on me meant to distract from your failure.

Edit: Just to stop you from filling the thread with more barely relevant crap, I'll stick to the Lawrie thing demonstrating the flaw in the metric rather than proving it. Happy?

 

I love you you dive in and contribute nothing but a lame personal attack.

 

And your last bit points to my problem with much of this. These metrics are offered before working the bugs out. That doesn't make them useless, but far more flawed than ones that actually measure what they claim to measure.