Alex Sanchez for POMVP!

My comments on THT, including correlations.

It seems counterintuitive.  If you are trying to hit the ball hard, you aren’t trying to hit a groundball to the right side.

I just don’t think, in an offensive age, giving up outs is a great idea.  Maybe in 1968, when a run was worth a lot more, but not in 2004.

Let me guess. Without looking. The guys with the most POs are primarily groundball hitters that are either lefty or switch-hit; while the hitters with the fewest POs are guys with high K rates, especially right-handed ones. 

How’d I do?

Alex Sanchez? Kris Benson!

Wins to Total Productive Outs: -.316
Wins to Total Productive Out Opportunities: .168
Losses to Productive Outs Allowed: .205
Losses to Productive Out Opportunities Against: .340
Productive Out Percentage to Winning Percentage: -.485
Productive Out Percentage Against to Winning Percentage: .002

Or to math challenged people like me..
Batting Top 5 Teams 140-182
Batting Bottom 5 teams 172-148
Pitching Top 5 teams 146-173
Pitching Bottom 5 teams 148-175
And to think 20 years ago people went postal on GWRBI being the same for 17-0 game as 1-0 game.

Let me put it this way:

The more productive outs you make, the more likely you are to lose.  The more often you make productive outs per opportunity, the more likely you are to lose.

My guess, shoomee, is that POP ends up occupying the same place as GWRBI in the history of baseball.

Warning—this might have already been suggested before in a previous productive-out thread.

Shouldn’t the focus be on the rate of advancing runners, be it by a hit, a walk, an out, etc?  So an advanced-runners statistic.  Guy on 2nd and 3rd, and Morgan Ensberg hits a HR (yeah right, not this year), he had an opporunity of 3 advance-runner points (3 bases) and successfully plated ‘em, so he’s 3/3.  And so forth.  Productive outs become just a small percentage of the metric, as it should.  But this metric will most likely have a large correlation to slugging percentage.  Well, slugging percentage doesn’t include walks, so perhaps OPS or something similar.  You get the idea. 

So ROE data are not easily available (see other thread).  Is advancing-runners data available anywhere other than within pbp data?

Baseball Tonight says Koch to Marlins for “minor leaguers”. As if the Marlins needed another 95 MPH+ fastball.  At least this guy doesnt get anybody out with his heat.

Sorry for the hijack but POP deserves it.

What is it about the Elias Sports Bureau and garbage stats?  They seem to go together like $5 hot dogs and $7 beers.

I was trying to think of any stats that were introduced and faded away. I guess GWRBI and quality starts, the latter has survived somewhat better despit “You get a QS if you pitch 6 innings and give up 3 earned runs? That’s a 4.50 ERA/ How can you call a 4.50 ERA a Quality Start”.
I always wonder why no one didn’t think to address this as 6 IP 2ER, 7+ IP 3 ER.
It’s been years since I read it but doesn’t John Thorn/Pete Palmer’s “Hidden Game of Baseball” talk about things liek “How Many Bases You Advnaced Runners Vs Potential” say runner on 1b you get 3 if hit HR but only 1 for a single. No computers in 1918 made this difficult to keep track of.

I’m just an out machine
And I won’t bunt for nobody but you...

I’m pleasantly surprised that Seattle ranks low in the PO metric, though I have a bad feeling Bob Melvin is reading this and saying, “Dammit, we need to bunt runners to second more!”

Plugging the PO and W/L percentages into Excel, the correlation between the two is -0.48.  It would be interesting to compare these to the Pythagorean W/L to see if POs have a role in teams under/overperforming.

I’m pleasantly surprised that Seattle ranks low in the PO metric

Well, it’s hard for an out to be productive when there’s no one on base.

Just out of curiosity, does anyone have an easy way to determine players’ PO/Un-PO ratios?

Plugging the PO and W/L percentages into Excel, the correlation between the two is -0.48.

Or you could have scrolled up and read the same result…

It would be interesting to compare these to the Pythagorean W/L to see if POs have a role in teams under/overperforming.

The correlation to playing over your Pythagorean Percentage to Productive Out Percentage is -.339.

How does this work?  Comments to News end up on Primate Studies also?  Did everyone else already know that?

Why compare PO percentage to wins instead of runs scored or runs allowed?  Are those correlations just for this one year?

I think we should all be very grateful for Wily Mo Pena and his 000

uh… pecerntage that is. I think he’s the only person on their with a zero

To get downright ‘Repozian’: This is the worst idea since Tricky Dick proclaimed “I am not a crook.”

;-) ...
-----------
trevise :-) ...Ã¥

Why compare PO percentage to wins instead of runs scored or runs allowed?

A) I compared the outs themselves to wins.  I compared POP to winning percentage.

B)Correlation to runs scored:
Productive Outs: -.048
POP: -.227
Productive Out Opportunities: .239

To runs allowed:
Productive Outs: .455 (which looks great compared to the other correlations)
POP: .232
Opportunities: .522

Are those correlations just for this one year?

Yep.  It’s the only avaiable data.

Posting all these correlations is, of course, beating a dead horse, but they’re just so ridiculously bad that it’s comical.

Buster Olney never let facts get in the way of his opinions.

Hey, it never hurts to track it.  Maybe in a few years someone will find a productive use for this new data.

I plan on doing another pointless study to see if this measures anything of even remote value.

Hey, it never hurts to track it.

Was that purely a joke, DKDC?  Because I disagree with it.

There are enough “specialty” numbers and stats that try to define parts of our great game.  This is less predictive than most and seems to track something which simply doesn’t matter much to runs scoring.  Why not track Alex Rodriguez’s home runs on Thursday day games?

Mr. Olney’s system is too simplistic and (again) non-predictive for the stat-savvy, and probably too complex for the casual fan.  So what’s it for?

*still steaming, just a little*

Olney may have been better served by not trying to quantify and study this phenomenon, but to continue writing unsupported opinions about it.

It’s trying to prove flawed “conventional wisdom” to those (like us) who have strayed from the baseball flock.  But it seems to be backfiring, uniting those it was designed to conquer.

Mr. Olney’s system is too simplistic and (again) non-predictive for the stat-savvy, and probably too complex for the casual fan. So what’s it for?

If Larry’s numbers are right, then doesn’t this system prove that productive outs are a bad thing?  That sounds like a pretty valuable conclusion to me, wheather Olney wanted to reach it on not.

APNY,
I don’t think this proves productive outs are bad at all - given the choice between a productive out and a non-productive out I’d take the productive one.

I think what the study actually shows is that the types of hitters that get productive outs (those that hit grounders to the right side) are NOT the same types of hitters that tend to AVOID outs altogether, which is really the outcome teams ought to be striving for.

What Greg Tamer suggested in an earlier post would be more interesting to study…

Btw, are there any pitchers on the batting list BESIDES the guy at the top?

f Larry’s numbers are right, then doesn’t this system prove that productive outs are a bad thing? That sounds like a pretty valuable conclusion to me, wheather Olney wanted to reach it on not.

Yeah, but Olney is “pinker and prouder than previous” and believes he’s the beez kneez. To convince him to concede his failure, someone will have to feed him plenty of cheeze to go with his whine…

-----------
trevise :-) ...

I think what the study actually shows is that the types of hitters that get productive outs…

Well that was the point.  Productibe outs are still outs.  And who makes more outs?  Crappy hitters. 

given the choice between a productive out and a non-productive out I’d take the productive one.

And this is part of Olney’s problem.  If you have a man on 2nd and no outs and ARod at the plate, it’s easy to say “I’d like him to get a hit, but if he makes an out I hope he advances the runner”. 

This goes against his hitting style.  He should not be sacrificing himself.  He should maintain his approach.  This approach leads to lots of “nonproductive” out (K’s, line outs, fly outs), but the exchange is worth it because of all the positives that result from that approach.

The reality is, most of these bad hitters are probably making productive outs by accident, and no good hitter should be trying to advance a runner one base when they can drive them in themself.

Does the definition even make sense to anybody?  Take this provision:  when a grounder or fly ball scores a run with one out. Why can’t no-out run-scoring outs get some love?  Does Olney conced that it’s better get a hit with a man on third and no out, but not when there’s a man on third and one out?  What sense does this make?

Greg Tamer’s right in post 10.  If you’re gonna keep a junk stat like this, at least make it Runners Advanced or something.  Would it show us something new?  Probably not ... but maybe, and it sure has a better chance than this POP nonsense alone.

I think we should all be very grateful for Wily Mo Pena and his 000

As is well-known, PECOTA predicted this for Pena.

What sense does this make?

<sarcasm>
It makes perfect sense!  Remember, getting that first out in an inning is HUGE!  So if there are no outs, you don’t want to make that first out.  But if there is already one out, well, then, shitfuckit (gotcha again, Cyber Nanny!), the inning is screwed anyway, might as well give up the second out to score that one run. 
</sarcasm>

If you’re reading this, Olney (or an Olney-follower), ignore the sarcasm brackets.

Using the Run Expectancy Matrix for 1999-2002 on Tango’s site:

With a runner on third and no outs, you would expect to score 1.482 runs.  If you make an out and get the guy home, the run expectancy drops to .297 ... but you’ve scored a run, so there are 1.297 runs in the inning.  You’ve cost your team .185 runs.

With a runner on third and one out, you would expect to score .983 runs.  If you make an out and get the guy home, the run expectancy drops to .117 ... but you’ve scored a run, so there are 1.117 runs in the inning.  You’ve gained your team .134 runs.

So, it does turn out that the out in the POP formula is more productive than the one I suggested, by .319 runs.  So maybe I’m a moron and Olney’s not.

What if there are runners on second and third?  With no outs, you’d expect 2.052 runs ... having a guy on third with one out is .983 plus the run ... a cost of .069 runs.

With one out, you’d expect 1.467 ... having a guy on third with two outs is .387 plus the run ... a cost of .080.  So the first out is more productive (or less unproductive) in that situation.

(Of course, this is for the average batter ... would there be a way to construct such a matrix for each individual hitter?  I guess this is what Tango did with Barry for that IBB chart ..)

Okay, I don’t feel so dumb now ...

I really thought that the original Olney article would be the end of it.

I was wrong.

Has this incredibly dumb stat been used on ESPN’s tv broadcasts yet?

Maybe it has on BBTN which is an occasion show in my house. I turned in midshow and Gammons was saying something about “you should approach hitting intelligently, have an idea what to do”. Must be Gammons speak for this has a little value, but you think ESPN/Elias would invent/use this stat and then promote it as the best thing since canned beer.

And this is part of Olney’s problem. If you have a man on 2nd and no outs and ARod at the plate, it’s easy to say “I’d like him to get a hit, but if he makes an out I hope he advances the runner”.

The reality is, most of these bad hitters are probably making productive outs by accident, and no good hitter should be trying to advance a runner one base when they can drive them in themself.

Let me first say that I agree there are a lot of problems with this “stat”.  But I think the point is that it is most useful for crappy hitters?  Why?  Well, a crappy hitter will, by definition, make a lot of outs.  If you’re going to make a lot of outs, best make them “productive”.  Now, if you’re A-rod, you’re not a crappy hitter and your job is get on base or hit the ball out of the park. 

So, I wouldn’t say a player is good because they make a lot of productive outs.  I’d say their getting the most of their ability.  A repeatable ability to make productive outs would be a good skill to have in a weak hitting defensive whiz.  So, I think a properly designed metric to measure POs would help you evaluate just how bad a bad hitter is. 

(Say you have to decide between a decent hitter who is lousy defensively and a great defender who - by traditional and SABR metrics - is a lousy hitter.  It might swing the decision in the defender’s favor if he has demonstrable skill at making productive outs.)

Is this a skill?  Wouldn’t you have to look at several years worth of numbers and see if the guys at the “top” this year are consistently there?

If Larry’s numbers are right, then doesn’t this system prove that productive outs are a bad thing?

No, you can’t prove anything with correlations. All you can say is that there is no linear relationship between POP and run scoring, or whatever. Doesn’t mean there isn’t a relationship.

Take this provision: when a grounder or fly ball scores a run with one out. Why can’t no-out run-scoring outs get some love?

They fall under the first category.

Take the analogy of poker.  A productive out is analogous to a pair of deuces.  Sure, it’s better than nothing but you aren’t going to win too many hands with a pair of deuces.

Olney is just trying to show that he can be as stat-savvy as the next guy with a pocket computer.  Not actually being stat-savvy, he comes off looking like a fool.

I like having raw data available. I don’t like to see the raw data get parsed in such a way to make the numbers irrelevant.

Why is a *productive* out not counted as a positive for regular hitters but is for pitchers? It’s because even the constructors of this shoddy statistic recognize outs suck and they are simply trying to shine turds. 

Outs are bad. It makes sense for hitters to do everything within their power to avoid making them. It just doesn’t make sense for hitters to put themselves at a disadvantage by unnaturally trying to hit to the right side.

I know you guys know this already. I hope you’ll excuse me for stating the obvious, but I needed a respite from banging my head against the wall.

Testing the changed comment preview page.

The funny thing is that Kris Benson leads batters in productive outs.  Why?  Well, in one game, Kris Benson laid down 4 successful sacrifice bunts.  That’s a record (and something of an accomplishment).  Of course, none of the player that were advanced by his productive outs scored! Small sample size and all, that tells you something about the “productive out”!

And, of course, it’s not that productive outs aren’t useful, but when you rank the most effective offensive possibilites, the productive out is way down the line.  Why so much importance put on the 7th most effective possibility?

Home run
Triple
Double
Single
Walk/HPB
Productive out
Strike out
Ball put into play resulting in one out
Ball put into play resulting in two outs
Ball put into play resulting in three outs

Was that purely a joke, DKDC? Because I disagree with it.

No, not a joke.  POP is almost certainly useless, but it can never be a bad thing to accumulate more data and have it readily available. 

Will anyone ever put the underlying data to good use?  Probably not, but why complain about it when you can easily ignore it like you do dozens of other useless stats?

Maybe it has on BBTN

Harold Reynolds has praised productive outs several times recently. He made a big deal out of them about 2 weeks ago.

Why doesn’t Larry (or someone more statistically adept than I) send Olney those correlations and see if they can put a stop to this?

Ziggy, If Olney understood correlations, he would have calculated them himself and thought “#### can this idea!”

Home run
Triple
Double
Single
Walk/HPB
Productive out
Strike out
Ball put into play resulting in one out
Ball put into play resulting in two outs
Ball put into play resulting in three outs

I’d rather Reach On Error before making a productive out.  I’d also rather reach via a strikeout and the catcher drops the ball.  Also, strikeouts versus productive outs need to be adjusted for GIDP rates.  Productive outs and GIDP, combined, might be worse than strikeouts.  I’m sure this has been researched before!  Just don’t know where it is at the moment.

We need a name for the award to give to the player who leads this category. And should the winner be the leader in percentage or total outs?

Something along the lines of the Lady Byng.

If somebody grounds into a DP that also scores a runner from third, does that count as two Productive Outs?

I’d rather Reach On Error before making a productive out. I’d also rather reach via a strikeout and the catcher drops the ball.

True.  I was considering “clean” plays (I mean, a single with a three-base error is just as good as a home run).

Also, strikeouts versus productive outs need to be adjusted for GIDP rates. Productive outs and GIDP, combined, might be worse than strikeouts.

True.  Man on 1st and 3rd, nobody out.  GIDP gets you 1 run and 2 outs (and nobody on base).  Strikeout gets you no runs and 1 out (and 2 people still on base).

The more you consider things, the more that Olney’s stat crumbles.  But then again, in his eyes, the 2001 Yankee World Series loss was the beginning of the end to the franchise…

We need a name for the award to give to the player who leads this category.
---
Something along the lines of the Lady Byng.

What about the Buster Olney Trophy?

We need a name for the award to give to the player who leads this category. And should the winner be the leader in percentage or total outs?

POP artist?  P.O. Rox?  Spiffy POP?  PO Folks?  (that was a great restaurant) POetry in motion?  How PO can you go?  The award should be a POPsicle stick with Atta Boy! written on it.

Perhaps the winner should be given a plaque with a special boxscore etched into it that wouldn’t have his name on it since POPs “don’t show up in the boxscore”

I’m not advocating productive outs in any way, but I think we need more than a correlation coefficient to demonstrate that POP is just about worthless. As I stated above, all correlations tell you is whether or not there is a direct, linear relationship between two variables. Obviously there isn’t a direct relationship between POP and Runs scored, but the correlation coefficient assumes that there is a linear relationship between POP and Runs Scored (I’m using RS because that seems to be the most meaningful correlate of POP - if it exists).

It turns out there might not be a linear relationship between POP and RS. Creating a line of best fit through the data produces the following R-squared values:
linear- .049
quadratic- .056
logarithmic- .047
power- .052
exponential- .55
compound- .55

So a linear model of POP accounts for less variance in RS than a quadratic, power, exponential, or compound model. Of course, none of these R-squared values are significant, so I can’t say it’s not linear. But at least we should think of better ways to go about debunking POP than by using correlations.

I’ll try to get the ball rolling…

Adam Everett has made 21 productive outs in 41 opportunities, which is good for 10th in POP at .512.
Adam Everett was projected by Marcel to hit .253/.316/.373, so I’ll say that’s a good estimate of his true talent level. He’d be expected to single in 15.9% of his plate appearances, and so on…
p(1B) = 15.9
p(2B) = 4.2
p(3B) = 0.5
p(HR) = 1.8
p(BB) = 7.6
p(HBP) = 1.8
p(NPO) = 35.0
p(PO) = 33.4

So let’s look at the run expectancy matrix for all four batter PO situations…
runner on 1st, 0 outs = .953
runner on 2nd, 0 outs = 1.189
runner on 3rd, 0 outs = 1.482
runner on 3rd, 1 out = .983

what’s more productive for Everett in each of these situations, a PO or an attempt to reach base?

(I will assume that all singles advance the runner one base and all doubles advance the runner two bases. I’m hoping this conservative approach is offset by not factoring in double plays - don’t need to make things too complicated)

runner on 1st, 0 outs
probability of event; run expectancy before outcome->run expectancy after outcome = difference in run expectancy
p(1B) = 15.9; 0.953->1.573 = +0.62
p(2B) = 4.2; 0.953->2.052 = +1.099
p(3B) = 0.5; 0.953->1+1.482 = +1.529
p(HR) = 1.8; 0.953->2+0.555 = +1.602
p(BB) = 7.6; 0.953->1.573 = +0.62
p(HBP) = 1.8; 0.953->1.573 = +0.62
p(NPO) = 35.0; 0.953->0.573 = -0.38
p(PO) = 33.4; 0.953->0.725 = -.228

value of attempting to get on base = .25
value of a productive out = -.228
difference = -.478

runner on 2nd, 0 outs
p(1B) = 15.9; 1.189->1.904 = +.715
p(2B) = 4.2; 1.189->1+1.189 = +1.0
p(3B) = 0.5; 1.189->1+1.482 = +1.293
p(HR) = 1.8; 1.189->2+0.555 = +1.366
p(BB) = 7.6; 1.189->1.573 = +.384
p(HBP) = 1.8; 1.189->1.573 = +.384
p(NPO) = 35.0; 1.189->0.725 = -.464
p(PO) = 33.4; 1.189->0.983 = -.206

value of attempting to get on base = .06
value of a productive out = -.206
difference = -.266

runner on 3rd, 0 outs
p(1B) = 15.9; 1.482->1+0.953 = +.471
p(2B) = 4.2; 1.482->1+1.189 = +.707
p(3B) = 0.5; 1.482->1+1.482 = +1.0
p(HR) = 1.8; 1.482->2+0.555 = +1.073
p(BB) = 7.6; 1.482->1.904 = +.422
p(HBP) = 1.8; 1.482->1.904 = +.422
p(NPO) = 35.0; 1.482->0.983 = -.499
p(PO) = 33.4; 1.482->1+0.297 = -.185

value of attempting to get on base = -.006
value of a productive out = -.185
difference = -.179

runner on 3rd, 1 out
p(1B) = 15.9; 0.983->1+0.573 = +.59
p(2B) = 4.2; 0.983->1+0.725 = +.742
p(3B) = 0.5; 0.983->1+0.983 = +1
p(HR) = 1.8; 0.983->2+0.297 = +1.314
p(BB) = 7.6; 0.983->1.243 = +.26
p(HBP) = 1.8; 0.983->1.243 = +.26
p(NPO) = 35.0; 0.983->0.387 = -.596
p(PO) = 33.4; 0.983->1+0.117 = .134

value of attempting to get on base = -.031
value of a productive out = .134
difference = .165

Conclusion:
attempting to reach base in productive out situations is worth .758 runs over making a productive out, even at the risk of making a nonproductive out. and this is for a guy who is most likely a true .253/.316/.373 hitter and who is in the top 10 in productive outs! you can imagine that for better hitters, the run expectancy goes WAY down when they attempt to make a PO instead of trying to reach base.

That’s a great post.  Interesting that the only out that turns out being productive is the run-scoring out with one out. 

This all seems like the sacrifice bunt.  There are times and players for which it may make sense, but most of the time it’s better to try to hit.

My favorite bit:

Juan Pierre, the supposed God of POP just a month ago, has a .400 POP.

Mark Bellhorn, the supposed King of anti-POP, has a .379 POP.

Pierre has also hit into 7 DPs to Bellhorn’s 4.

I’m just saying.

Admittedly, I’m shocked Bellhorn’s POP is that high.  Given nearly half his overall outs are Ks, it’s kind of amazing.

attempting to reach base in productive out situations is worth .758 runs over making a productive out, even at the risk of making a nonproductive out. and this is for a guy who is most likely a true .253/.316/.373 hitter and who is in the top 10 in productive outs! you can imagine that for better hitters, the run expectancy goes WAY down when they attempt to make a PO instead of trying to reach base.

Hmmmmm . . . but isn’t there something else you can say, which is not that anyone should “try” to make a productive out, but that it is information worth having to know the following, based on the inarguable fact that all players sometimes make outs (though sometimes Bonds seems like an exception) even though they’re not trying to do so:

Since all players make outs sometimes, a player who can do something productive with those “failures” is more valuable than an otherwise comparable player who makes the same number of outs, with the same OBP and SLG, but does nothing productive with the outs.

Now, obviously, that’s a much more limited claim than lionizing “productive outs” as a good thing, for which a guy should strive.  But it is worth knowing that one guy—in the course of trying to actually succeed—fails less completely than another guy.  If that makes any sense . . . .

But it is worth knowing that one guy—in the course of trying to actually succeed—fails less completely than another guy.

Your applying a skill level.

If your trying to succeed, is the difference between a weak ground ball to second and any other kind of non-strikeout out anything other than luck?

APNY --

I’d say it depends totally on what question we’re trying to answer.  If we’re trying to decide:

Was Player A more valuable than Player B?

Then we don’t care whether it was luck.  Even value attributable to luck is value.  That’s why why a pitcher who happens to have a lucky year when it comes to BABIP is still valuable, even if we don’t think he’ll remain so next year.

But if we’re trying to decide:

Is Player A likely to be better than Player B next season?

Then, the fact that it’s just lucky (if it is) is extraordinarily significant, because (of course) Player A’s high POP isn’t likely to recur.

So I think it’s worth having the in-season data to evaluate the contribution a player may be making with his outs, for purposes of more fully assessing the value his performance.  It’s not worth having for predictive purposes.

Now, obviously, that’s a much more limited claim than lionizing “productive outs” as a good thing, for which a guy should strive. But it is worth knowing that one guy—in the course of trying to actually succeed—fails less completely than another guy. If that makes any sense . . . .

You’re right, but I don’t know if it is true…

There are definitely situations where batters try to hit the ball on the ground to the right side in order to move the runner over. Most sacrifice bunts are not attempts to reach base. Many productive outs are produced in situations where the batter’s primary goal is not reaching base. But unless there is a man on 3rd with 1 out, the best strategy is to try to get on base, not to try to make a productive out.

Why doesn’t Larry (or someone more statistically adept than I) send Olney those correlations and see if they can put a stop to this?

Alex Belth forwarded my first article at THT to Olney.  No response.

Walt,

This is why we all must love the Bellhorn.  I really believe he changes his approach in different situations.

It’s my belief that his high-K total is based at least partly on an understanding of his own limitations.  He can’t make good contact on even borderline strikes, so he lets them go in most situations in the hopes he’ll get a ball call.

But he changes his approach.  A guy on SoSH has been tracking Bellhorn with runners on 3rd, less than 2 out.  He has 17 PAs, one K, a .500 BA and my favorite, 3 SFs (tied for 1st on the Red Sox).

I think the problem with concepts such as “productive outs” is that it’s hard to measure the true context. For all we know, A-Rod may be trying to hit to the right side of the infield every time he’s in a runner on second, no out situation, and maybe the result is that he hits .380 with a .750 slugging percentage while attempting this. The attempted productive out is much like the attempted sacrifice, in that it can do much more than “work.”

of course we don’t know exactly how often this is attempted, so it’s impossible to determine whether or not hitters should actually be trying to do this.

Since there’s almost assuredly a high inverse correlation between productive outs and strikeouts, and strikeouts are already tracked, it seems rather pointless to track this stat.

Hey, the Tigers are 4th in MLB in POP, and second in the AL in Road OPS.

Coincedence?

I..think...so!.

But unless there is a man on 3rd with 1 out, the best strategy is to try to get on base, not to try to make a productive out.

I agree wholeheartedly with baudib that it is context that matters quite a bit because even in the italicized situation, depending on the score of the game, the defense might take that ground ball and go home with it or go to first. The fact that a run scores on the out is only as valuable as the context in which it occurs, which might be something like the extent to which that out and run lead to a greater chance of winning under the circumstances. Productive outs when you are down by five runs don’t seem all that valuable from a contribution to winning angle to me.

I’d have less of a problem with POP if it somehow incorporated hits into it. In this way, players who get hits in POP situations aren’t penalized.

That’s why why a pitcher who happens to have a lucky year when it comes to BABIP is still valuable, even if we don’t think he’ll remain so next year.

Some (most?) of that value should be given to the defense.

I’d have less of a problem with POP if it somehow incorporated hits into it. In this way, players who get hits in POP situations aren’t penalized.

They aren’t.  Opportunities are outs, not plate appearances.