Predictive Range and Ian Snell

These days we know that defensive evaluation is as inexact a metric there is. Lately there’s been a trend to quantify defensive performance by putting it within a range. Griffin did just that with the UZR projections in his (excellent) post below. (Seriously. Go read it.)  For example: he puts the A’s Ryan Sweeney’s UZR at 7 < X < 12 in right field.

This gives you a couple pieces of important information.
1) Ryan Sweeney is good defensively.
2) His downside is better than average.
3) His upside is really good.
4) We don’t know what his defense gonna be like exactly but it’s gonna be between better-than-average and really good.

This is fantastic. That’s a lot of information presented in 6 characters. It’s like tinyurl. Presenting inexact information in this fashion is inherently self-effacing and more accurate. If there’s two things armchair sabermetricians can be it’s more self-effacing (self-effacing-er?) and more accurate. Now, not all statistics should be presented this way. Some statistics are very specific and should be presented exactly. There’s no way to present a range of a pitcher’s K% from a specific year. It is what it is. But inexact information should be labeled as such.

Projections, by nature, are inexact. There are variables that you cannot take into account in a mathematical formula. You can get close, but a range, like in defense, would make more sense.

Which brings me to Ian Snell.

The Fans, per Fangraphs, predict Snell to be 1.7 WAR in 2010.
Marcel ~1.5
CHONE ~2.0
Bill James ~2.3

So the best EXACT guesses at Snell’s 2010 performance are between 1.5-2.3 WAR, a large range in itself. There is, of course, natural variance based on how much upside, downside, and the % chance of reaching said upside and downside are. A wider variance would mean a greater volatility.

If his inability to find the plate continues like it did in the latter part of last year his downside is down to about 1 WAR. If he finds his 2007 fastball/slider combo, regains his control, and fixes his mechanics he’s likely to be a ~3.0 WAR pitcher, that is to say, slighty below his 3.5 WAR 2007 season.

Now, I love the idea of Ian Snell. I am subjectively bullish on his upside. There’s a lot of helium there. He has demonstrated dominance in the past and was in a bad place both geographically and mentally in Pittsburgh. Stupid Pittsburgh.

On the other hand once he hit Seattle he kinda fell apart. His strike-outs plummeted and the walks soared. A lot of people thought that he was trying too hard. I dunno, he stunk throughout 2009 after most of the pressure was seemingly off so… (aposiopesis pregnant with inference!)

His % chance of remaining who he is currently is higher than his % chance of reverting to 2007 Super-Snell so I’ll weight his downside potential heavier than his upside potential. His absolute range’d be between 1.0 and 3.5 WAR but adding a subjective % chance of reaching each extreme I’d put him at about 1.3 < X < 2.5 WAR for 2010. Now then, isn’t that’s more instructive than an exact guess? Isn’t that what we’re looking for here?