Gold Glove Selection will add a Sabrmetric component

[SteveA]

http://www.cbssports.com/mlb/blog/eye-on-baseball/21834257/gold-glove-awards-adding-sabermetric-component

[DrungoHazewood]

It's interesting. At first glance it appears they're (at long last) responding to the criticism of the Gold Gloves being no more than a poorly-designed popularity contest. But it's bizarre that it appears they've convinced SABR to invent their own proprietary black-box fielding system instead. SABR is as close to a scholarly arm of baseball as you can get, but unless I'm wrong it looks like Rawlings has gotten them to come up with a trademarked system that is probably not going to be transparent and use it for marketing Rawlings. If that's all true, I'm a little disappointed in SABR.

Let's hope they do what's right and at least release the details of the system to the public for review and comment.

[Frobby]

It's interesting. At first glance it appears they're (at long last) responding to the criticism of the Gold Gloves being no more than a poorly-designed popularity contest. But it's bizarre that it appears they've convinced SABR to invent their own proprietary black-box fielding system instead. SABR is as close to a scholarly arm of baseball as you can get, but unless I'm wrong it looks like Rawlings has gotten them to come up with a trademarked system that is probably not going to be transparent and use it for marketing Rawlings. If that's all true, I'm a little disappointed in SABR.

Let's hope they do what's right and at least release the details of the system to the public for review and comment.

I look at it this way: I'm excited that someone qualified is going to take another crack at coming up with a set of defensive statistics, and I'm excited that some effort is going to be made to weigh some metrics and give the managers and coaches some metrics to look at in the Gold Glove process. It may not be perfect, but it sounds better than what they've been doing. I wonder if they are going to use "Field f/x" to come up with their statistics -- is that up and running yet? http://www.sportvision.com/baseball/fieldfx

[waroriole]

It's interesting. At first glance it appears they're (at long last) responding to the criticism of the Gold Gloves being no more than a poorly-designed popularity contest. But it's bizarre that it appears they've convinced SABR to invent their own proprietary black-box fielding system instead. SABR is as close to a scholarly arm of baseball as you can get, but unless I'm wrong it looks like Rawlings has gotten them to come up with a trademarked system that is probably not going to be transparent and use it for marketing Rawlings. If that's all true, I'm a little disappointed in SABR.

Let's hope they do what's right and at least release the details of the system to the public for review and comment.

You mean you're going to be dubious when their new formula says that Derek Jeter, recently recovered from a broken ankle at 38, is actually the best defensive SS?

[DrungoHazewood]

You mean you're going to be dubious when their new formula says that Derek Jeter, recently recovered from a broken ankle at 38, is actually the best defensive SS?

You know (especially with managers and coaches inputs still being the majority of the weight) there will be some head-scratchers. And if the new system isn't transparent it'll be at least as controversial as before. Now there's opportunity for both the objective analysis folks and the traditionalists to say the process sucks. Now you (probably) have the GGs saying "Trust our black box, and trust the managers. Still no need to look behind the curtain, we know more than you."

[DrungoHazewood]

I look at it this way: I'm excited that someone qualified is going to take another crack at coming up with a set of defensive statistics, and I'm excited that some effort is going to be made to weigh some metrics and give the managers and coaches some metrics to look at in the Gold Glove process. It may not be perfect, but it sounds better than what they've been doing.

May be better, may not be. And we might not have any way of knowing. If it's a black box and they're asking us to trust them on it, well... I'd rather have publicly available data I can review. With what we have now we can look for stringer bias, we can look for home/road splits, we can see if there are consistent idiosyncracies in the data. With the SABR method there's a good chance it's going to be "Joe Smith topped our rankings and combined with manager opinions he's the Platinum award winner!" and that's all the data we get.

Although I'd also guess that within days of the first awards we'll have people who reverse engineer the process and more-or-less figure out what the inputs are.

I wonder if they are going to use "Field f/x" to come up with their statistics -- is that up and running yet? http://www.sportvision.com/baseball/fieldfx

Yes, it's up and running, and I think it's highly doubtful it will be a part of this. FieldF/x is still a proprietary thing licensed to MLB and a tightly kept secret, and I doubt a) they'd let a bunch of SABR panelists see the results, and b) they'd let the metric be published even in the filtered form of GG results.

[isestrex]

Excited about Rawlings wanting to make this change... but puzzled as to the execution. It sounds like they made it more complicated than it already is.

Personally I would prefer to have sabermetrics simply choose 5 nominees at each position and let everyone vote on that. That way we wouldn't get any horrible candidates.

[skanar]

The GGs have mostly been not horrible selections recently. I went back and compared the Fielding Bible stats with the GG selections over the last few years a couple of months ago during the Trout/Jones brouhaha and the winner was usually in the top 3 in his league per FB. Still, when they pick badly, they usually pick VERY badly.