in terms of his comparables and the way his periphals are trending I think Teix’s peak as a hitter maybe just at the begining, in that he might have 3-5 year of really really good years ahead of him. (sort of like Jason Giambi 99-01).

As a defensemen he’s always been rather inconsistent from year to year though, he’s landed all over the map in his career, from slightly below average to truely gold glove caliber results.

The past three years prior to 2008 Tex’s UZR has been -2.5, -1.8, and -3.5. If anything one could argue that Tex’s 2008 defensive performance is something of an outlier as he hasn’t had a defensive year even remotely close to this one since 2004 and even then he was +8.8 (career high until 2008). In fact by UZR he hasn’t been above league average since 2004.

Tex wasn’t very good defensively in 2007 by either +/- or PMR. League average to below league average in both.

Tex was league average by PMR in 2006 as well. He wasn’t one of the top 10 fielding first baseman by +/- in 2006 either.

In none of these years where he was below league average by these metrics did he suffer an injury which would make his defensive performance worse than his innate ability.

I agree that subjectively Tex seems to me to be a good defensive player.

However, quantitatively I find no evidence that would allow me to confidently say that I could project Tex to be +5-10 defensively moving forward into the future. If anything 2008 appears to be a relative outlier compared to his recent performance by multiple metrics.

Now you can bring up all of the usual grab bag of issues around defensive metrics and the significant variation we see in them (sample size, etc).

However, it seems to be insufficient to simply use this argument and at the same time to credit Tex as a +10 defensive player moving forward into the future. It seems even more problematic to then value him as an +1 WAR player for the marginal advantage he provides as a defender when trying to quantify his economic value.

There’s been no similar quantitative variation in Pujol’s defensive performance so I find it difficult to conclude that Teixeira would for instance be in the same defensive class as Pujols. Pujols UZR has gone up and down – but he’s never been below league average. And he’s consistently at the top of +/- for first baseman.

This is a critical issue in valuing Teixeira as he is the unusual 1b in that a substantial proportion of his total value lies in his glove and not only in his bat. He’s obviously not Prince Fielder or Miguel Cabrera. So to gauge his value – especially to gauge his value quantitatively – you must be able to gauge his defensive worth.

But if we can’t reasonably conclude from quantitative measures what his defensive performance is due to variation (for whatever reason that variation exists) then it seems the more prudent analytic method to use would be some form of sensitivity analysis looking at the range of potential values Tex might have depending on whether his defense is more like 2005-2007 or more like 2008. In turn one would have to generate a range of potential dollar figures Teixeira would be worth.

Given the dollars Teixeira is seeking variation in his defensive performance could dramatically cost his team millions of dollars. At the very least his recent defense should at least add some kind of risk and possibly a risk premium into estimation of his potential worth in dollars.

I’ve been somewhat surprised that so many quantitative analysts are willing to give Teixeira so much latitude in their appraisal of his defense and his net value. Variation in his defensive metrics is explained away as “noise” more often than I’ve seen for nearly any other player who has similar statistical variation.

Over and over on multiple sites that do terrific work – whether it’s here or at baseball prospectus – readers are told that Teixeira should be valued as a premium defensive performer (i.e. his level of innate skill is so high that he makes first base defense an enormous plus). But these arguments don’t fully account for the empirical variation in his defensive performance across multiple defensive quantification systems. “Noise” should be an argument of last resort. The more conservative analytic strategy would be to factor in this variation into the projection and contract valuation.

The subjective perception of Teixeira’s defensive skill seems to me to be driving the interpretation of his empirical, quantitative defensive value. That’s perfectly valid – that’s what scouts do. But the we should admit we are more in the realm of qualitative data rather than quantitative ( or perhaps in the realm of qualitative data informed by quantitative…).

Teixeira would be a big upgrade over Lowell, but how do you move him now that the sox have agreed in principle to a deal with Tex. 8 for $170. Details are still sketchy but consider the deal done.

http://sonsofsimmons.com/mlb/sources-red-sox-agree-in-principle-on-deal-with-mark-teixeira/2008/