Padres bloggin' since 2007

Old_Padre’s tRA talk

January 16th, 2009 by Melvin

In case anyone missed the comments section from this post, Sac Bunt reader Old_Padre had some concerns about my use of the tRA statistic in forming my opinion on a Jake Peavy trade. Here’s what he said:

How much of a predictor of future performance is a stat if it has swung back and forth by at least 20% over the last four seasons?

I responded that what I like about tRA is the way it breaks down a pitcher’s results into components for which we can more accurately apply credit or blame. Unfortunately I was not able to add much more to my answer because I am not a statistic guru, I merely enjoy sharing what I have learned with others who might not have the time, interest, or whatever other reason to follow sabermetrics news.

I suggested Old_Padre get in contact with Graham MacAree, the main developer behind tRA for a better answer to his questions. Awesomely, Old_Padre did just that, and thanks to both gentlemen I learned a lot in the process. I would like to re-post their e-mail conversation, in full, so hopefully more readers can benefit from Old_Padre’s diligence.

Mr. Macaree,
I was recently introduced to tRA through a blog post elsewhere written about Jake Peavy. I got into a bit of a conversation with they blogger who suggested I pass my question on to you.
His implied belief about tRA is that because it analyzes all components of the batter/pitcher interaction is that it is more useful as a predictor of future performance. The question I asked in the comments is “How much of a predictor of future performance is a stat if it has swung back and forth by at least 20% over the last four seasons?”
Now, as I read your primer, I don’t know that you would make the assertion that it’s a more accurate predictor of future performance just because it paints a more accurate picture of past results. However, if you DO feel that’s the case, could you help me understand why I saw such significant variance (I say that without having enough time or competence to truly do regression analysis or to look at deviation numbers) in the year-to-year tRA numbers for Peavy and the other guys I listed in the comments.
I’ll admit I’m not that bright (I’m a Padres fan, how smart could I be???) and it’s been far more years since I used SPSS than it was months that I used the damn thing, so my statistical prowess is not, ahem, remarkable. However, I would love to try to understand better.

Graham’s response:

tRA obviously isn’t a perfect future performance indicator – but nothing that we have is a perfect performance indicator. Every stat suffers from large swings year-to-year – if you look at the variation in ERA and compare it to tRA’s, you’ll find that the former’s is much higher. In fact, tRA is something like 2.5x as stable year to year as ERA, and also beats FIP (the most common ‘advanced’) pitching stat by a fairly healthy margin. So although it’s prone to large variations, it’s still more predictive than looking at other stats. Baseball is just too variable for any statistic to remain static.
I hope that clears things up.


Thanks, Graham. I appreciate that you replied SO quickly!
I certainly didn’t mean to imply that I thought there was no value, or that because it has some variance, it isn’t predictive at all. I guess my beef is more with the tack taken by the blogger… look, his tRA ballooned last year, quick trade him before anyone notices.


It may be of some interest to you that one of the alternate names for tRA in development was ‘the Jake Peavy is amazing stat’.

Personally, I think that a ballooning tRA is a large red flag, but it’s certainly not the only thing that should be considered.

Definitely fun to hear our favorite Alabaman was on Graham’s mind when working on tRA. Thanks again to Old_Padre for contributing such great dialogue, and to Graham MacAree for his insight.

I would also like to clarify that I don’t base my opinion that the team should make the right deal for Jake based solely on his 2008 tRA. In fact, that reason would probably land at #4 or #5 on my list. I put Jake’s violent, all effort delivery, age combined with injury risk, the large percentage of team payroll his contract covers, and the holes on the roster that need filling before the tRA thing. Sorry if that wasn’t clear in my original post.


Posted in statistics | Comments Off

Comments are closed.

Search Posts

The Sacrifice Bunt on Facebook The Sacrifice Bunt on Twitter


Sacrifice Bunt Shop

Sacrifice Bunt Shop