Comments on: On Productive Outs

By: michael standish

michael standish — Mon, 19 Oct 2009 19:43:02 +0000

Re Productive Outs:

I may be overlooking something here, but I don’t see how “…logic dictates that productive outs are preferred to non-productive outs.”

Conferring a positive value on an out strikes me as trying to sneak baggage past the customs agent.

By: Marc Schneider

Marc Schneider — Wed, 14 Oct 2009 14:18:30 +0000

I can’t contribute to the statistical analysis discussion. Obviously, no one goes up looking to make an out. But, clearly, if you are up with a runner on third and less than two outs (assuming the infield is back), it’s better to hit a ground ball to second than to strike out. (The Yankees won the 1962 World Series on a GIDP with the bases loaded.) And, presumably, this is somewhat within the hitter’s control–you could cut down your swing, for example, to increase your chances of striking out.

The problem is, whether this is really a “productive” out depends on the situation. If it’s a tie game in the 9th inning, a sacrifice fly is productive. If it’s the first inning and the bases are loaded, the sac fly actually is helping to kill the rally. “Productive” outs often lead to cheap and meaningless RBIs (another reason why RBIs are often meaningless stats), such as when a team is behind by several runs in the late innings and the hitter drives a runner in from third with a ground ball. It helps the hitter’s RBI numbers but doesn’t really help the team. In a case like this, it really isn’t a productive out (even though Joe Simpson would probably say it is) because it will actually depress the number of runs scored compared to a non-out.

By: JC

JC — Tue, 13 Oct 2009 16:55:14 +0000

Re: #2

It means the estimated effect is not outside the the normal variation if the effect is zero. I’m fine with stating the effect is zero. Considering that the coefficient estimates are negative, claiming no effect is actually charitable.

By: Aaron W.

Aaron W. — Tue, 13 Oct 2009 16:45:44 +0000

Perhaps the bigger point is that we shouldn’t be congratulating people on making a lot of productive outs. Whether the correlation is insignificant or just incredibly small is relatively unimportant.

By: Devon

Devon — Tue, 13 Oct 2009 15:58:04 +0000

Sounds like POP is kinda like the modern GWRBI.

I noticed a few nights ago on TBS that they showed what teams had the most productive outs during the season and I instantly wondered “yeah, but what % of their total plate appearances is that?” …I think that might hold more significance to their run scoring +/- than just lookin’ at productive outs. I could be wrong though.

By: Tom

Tom — Tue, 13 Oct 2009 13:36:29 +0000

Just because the coefficient is not statistically significant does not mean it is zero, just that it is small (relative to what could be detected, power, etc.). I think you should not say it’s zero, you should just say that it doesn’t help much with prediction, given what you already have to use in your prediction. In fact, maybe that is what you DID say!

By: Tweets that mention Sabernomics | On Productive Outs -- Topsy.com

Tweets that mention Sabernomics | On Productive Outs -- Topsy.com — Tue, 13 Oct 2009 13:12:35 +0000

[...] This post was mentioned on Twitter by J.C. Bradbury and Patrick Sullivan. Patrick Sullivan said: Good reminder as playoff audiences grow and are subjected to such falsehoods RT @jc_bradbury: On Productive Outs http://tinyurl.com/yzqg47w [...]