Lookout Landing

The Typical Idiot Fan - November 13, 2009 reply

He had a couple years of negative UZR and possible age decline.

I could see that influencing a decision.

Who voted 20-25?

baetown415 - November 13, 2009 reply

Why not?

Optimistic version of a great defender. That’s not really that much of a stretch. -5 is just a nonsense vote – the kind of vote that people do so that if he gets injures or sucks they can point to it and say “SEE? I told you!” even though their decision was based on nothing but the will to be proven right in the event of an outlier. That’s a stupid reason to vote for anything.

-5 is far more likely than +20

I think Gutierrez spoiled us. +20 at a premium defensive position is insane. It almost never happens.

davidcameron - November 13, 2009 reply

Well, easily.

But this is a poll of anyone that visits this site, not necessarily the educated fan. Jack Wilson is known for outstanding defense, so an excited fan can easily pick +20 without issue. However, no one – educated or not – has any excuse to expect -5, even if it is more plausible due to injury and age. -5, however possible, is a guess that someone would only make to be a dick. +20 is a pick someone would make because they’re super happy he’s on our team.

Or, to elaborate

If we had a projection table for Pujols, and the options were:

.370/.525/.700

and

.250/.375/.400

The latter is more realistic than the former, because the latter can easily occur if Pujols hurts himself and/or has an extremely low BABIP, while the former would be extraordinary. But the guy that votes for the former is just high on Pujols and believes he can easily exceed even the best expectations while the latter is being an asshole who only makes that guess because he wants to claim he’s right in the event Pujols falls off a cliff.

I actually think the latter is more likely

I could see Pujols having a season where everything comes together and he slugs .700, but he’s just not really fast enough to hit .370 or patient enough to post a .500 OBP unless gets intentionally walked 70 times.

Poochie - November 13, 2009 reply

It wasn't an "or" - it was if those representedthe upper and lower ends of a projection poll like this one.

The Typical Idiot Fan - November 13, 2009 reply

Or because they don't like aging players coming off injuries?

I don’t know who voted for -5 or what their motives are, but it’s not that crazy.

davidcameron - November 13, 2009 reply

Well, I would actually say he is 15-20 per 150 games,

But I think he can only be counted on to play 100-120 games, and thus be worth about 10-15 during the partial season.

ARock - November 13, 2009 reply

My thoughts as well

seattlecougar - November 14, 2009 reply

I went a bit conservative and voted 5-10.

Still would be huge breath of fresh air after watching Yuni the past couple years.

Benne - November 13, 2009 reply

Just for fun(?)

The best qualified SS defenders from 2006-2008 by UZR/150, then their 2009 UZR/150.

Omar Vizquel, +13.3, +30.2 in way too small of a sample
JJ Hardy: +14.3, +8.8
Jimmy Rollins: +9.1, +2.9
Jason Bartlett: +9.4, -6.8
Orlando Cabrera, +7.2, -13.7

Projecting Wilson for anything past +10 is probably a little crazy.

davidcameron - November 13, 2009 reply

Oh stop influencing the vote.

Let them have their fun.

Presumably people have voted before they read the comments

The Typical Idiot Fan - November 14, 2009 reply

I agree with whatever Jeff says.

I'm gonna go with around +7 or so.

Still good, but injuries and age will start to affect his defense.

Goose - November 13, 2009 reply

I feel silly asking,

but where do you find Revised Zone stats?

erik - November 14, 2009 reply

THT has RZR.

how do you put it into runs?

erik - November 14, 2009 reply

There's an article that touches on that

here. It’s not something I’ve ever tried to do, but it seems fairly straightforward.

much obliged.

erik - November 14, 2009 reply

So far the community projects Wilson to be +14 / 150

Graham MacAree - November 14, 2009 reply

That's a wee bit optimistic

Poochie - November 14, 2009 reply

The amusing thing is a +7 Jack Wilson is +20 on our last shortstop

So he's a +20 shortstop!

seattlebruin - November 14, 2009 reply

So I've been wondering about how worthwhile something like this would be for a little while:

Link.

This seems like a good place to ask; given UZR’s unreliability on a year-to-year basis, how useful/predictive is an exercise such as this? My initial reaction is to think that it’s fairly realistic for someone with a large sample (like Jack Wilson, say, who is projected to be +4.3,) while it would be less than optimal for someone with a smaller sample (Franklin Gutierrez is projected as +15.5 in CF, as another example.)

The problem I can see, though, is in the weighting; because you need three full years of data to make any conclusions about a player’s true level of defensive ability, weighting the year-to-year data in such a manner seems less than optimal. If you simply average Wilson’s last three years of UZR/150, you get ~12.5/150. The linked projections, though, are including data outside of that three year range (even if it is carrying less weight) and including a penalty for aging, which seems fair until you consider that more than a few players have shown a lot of improvement in UZR as they age. That seems counterintuitive, but it’s still there.

Thoughts?

Let's see

1) Any projection is going to be better for players with larger sample sizes

2) Recent information is always more valuable than older information, so you always want to weight heavier what’s newest. We could, for example, probably improve our ability to project offense going forward just a little bit by splitting seasons into halves and weighting post-ASB a little stronger than pre-ASB

3) Information from more than 3 years ago is still useful, just much less so

4) Aging penalty is fair. Not all players decline as they age, but most do, hence the trend. These projections are, I imagine, intended to be conservative

Cool, thanks.

As far as #1 is concerned, sure. But I’d wager that the sample size issues are a much, much bigger deal with UZR than they would be with offense or pitching.

Otherwise, that sort of puts to rest any concerns I had about the usefulness of these projections. Neat.

Sample size is a big issue, but if a player has limited time, you can either do your best or throw up your hands and say "whatever"

Most people elect to try to do their best. As long as people know that guys with small sample sizes will have less reliable projections, then it’s okay.