Productive Outs and Measuring Outs That Are Better than Others
by dwishinsky on Feb 26, 2011 4:13 PM PST
Email
PrintI was thinking that I hate it when there is a guy on second with no outs or one out and the guy at the plate strikes out or hits it to the shortstop and that guy is still on second. While I am of the belief that no outs are productive (save for situations where you need one run and hit a sacrifice fly for example - though I am sure there are others) I am more satisfied in the aforementioned scenario when a guy hits it to the right side, and though we have an out at least that runner is now on third as opposed to still stuck at second. I wouldn't call it a productive out, but I would call it a better out.
Much criticism has befallen the Productive Out and Productive Out Percentage (POP) that Buster Olney and ESPN espoused as an "anti-Moneyball" piece of evidence that teams that moved runners along were successful and those that didn't weren't. The Productive Out worked as follows:
"A Productive Out, as defined and developed by ESPN The Magazine and the Elias Sports Bureau: when a fly ball, grounder or bunt advances a runner with nobody out; when a pitcher bunts to advance a runner with one out (maximizing the effectiveness of the pitcher's at-bat), or when a grounder or fly ball scores a run with one out."
The idea is intriguing because at least I am of the belief that not all outs are created equal. Yet, why are those the only situations where an out is productive? Some of these ideas seem stupid. Let's say you had a number four hitter up and they struck out with a man on second and out one. The play would be an 0-for-1 as there were zero productive outs in one potential productive out situation. Yet let's say this same team pinch hit their pitcher who then laid down a sacrifice bunt, suddenly they would up their POP. I hardly think any team that did this consistently would win more games as the POP says they should. Furthermore, isn't it more productive in the same situation, for a guy to hit it to the right side and advance the runner as opposed to striking out and not moving him at all? Also why is a fly ball that scores a runner productive with one out but not with none? If the game is 3-3 in the bottom of the ninth and no one out and the hitter hits it to deep right field thereby allowing the runner at 3rd to score the winning run, I think everyone on earth would call that productive, yet POP would be .000 for that play.
So I am seeking to figure out a metric that captures this. But how do you go about that? I am seeking assistance to find holes in my idea so that I can either remedy it, or scrap a timely process before finding a serious hole later in the game. ESPN seems to use their Productive Outs to say that the team with a higher percentage will succeed more often. I don't think that is necessarily true, but I think the important takeaway from what I will call my "Better Out Percentage" is on the individual level. Can we capture the guy (I am thinking Kevin Kouzmanoff) who strikes out or hits into double plays versus those guys who seem to hit to the right side and move guys along even when making an out (I am thinking Ichiro Suzuki). For all I know the "Better Out Percentage" may prove me wrong and show that Suzuki is terrible and Kouz keeps runners inching towards home?
My initial concept was for all less than two outs situations (since you can't have a productive two-out out) take the number of runners, which I have named "Runners To Be Advanced" and then tally the number of "Advanced Runners". So your formula is a simple AR/RTBA = BOP. For example, from last year's opening day: in the bottom of the 6th inning Ryan Sweeney came up with Rajai Davis on third and Daric Barton on first. Sweeney singled and Davis scored and Barton advanced to second. So there would be two "Runners To Be Advanced" and both advanced so two "Advanced Runners". Sweeney would have a BOP of 1.000 for this play.
But the problem was as I began to tally, not so simple. What about other plays from the game, like the one that set up the above scenario. Barton was up with Davis on first. Rajai stole second. OK, easy stuff, nothing to do with Barton it is ignored. Then a Felix Hernandez wild pitch advanced Davis to third, again easy stuff, nothing to do with Barton it is ignored. But now Barton has in theory one "Runner To Be Advanced" yet walks. Barton didn't advance Davis, but we also don't want to penalize him for walking. I err on the side of ignoring the whole situation so that it is not incorporated into the equation.
What about this situation from the same game? Ryan Sweeney is up with Daric Barton on first and nobody out in the bottom of the eighth. So using our criteria, Sweeney comes up with one "Runner To Be Advanced" (Barton). So Sweeney grounds the pitch to Jack Wilson, who bobbles it, tries to get Barton out with a toss to Chone Figgins at second and doesn't. It is ruled an error. Technically, Sweeney has advanced Barton, Barton was on first and now is on second. But should he get credit for that? He did it solely because Wilson at short bobbled the grounder. Again, I err on the side of ignoring the whole situation so that it isn not incorporated into the equation.
The next hitter in the bottom of the eighth presents another quandry. Kouzmanoff comes up with Barton on second and Sweeney on first. Kouzmanoff hits into a double play that advances Barton but eliminates Sweeney. If we use the aforementioned formula, Kouzmanoff gets a .500 on the play, he advanced Barton but didn't advance Sweeney. Yet it doesn't seem right to give him credit for hitting into a double play that I see as a universal negative. Though I am not a fan of the stat, a run scoring on a GIDP does not yield the batter an RBI. I think likewise the "Better Out Percentage" should not recognize Barton as being advanced. Yet, Kouzmanoff didn't fail to advance two runners, he only failed to advance one. So do you count it as a "Runners To Be Advanced" 2, "Runners Advanced 1" or a .500. Or is it a 0/2 play despite Barton advancing or an 0/1 play to take into account that one batter advanced but we don't want to give credit for that. I'd go with the latter but it doesn't seem to correct take into account the complete situation.
I am wondering what people's thoughts are. I am happy to try and do the leg work on the stats beginning with last season. But what should the full rules for the "statistic" be. Do you think it even measures anything and is even worthwhile? If anyone wants to help me with the "Better Out Percentage" I am happy to have help! I'm all ears, please pick apart and criticize away as I think that the Productive Out does try to capture something that current stats fail to catch, which is that all outs are not equal, but ultimately miserably fails to do so. Can we do better?
48 comments
|
1 recs |
Do you like this story?
Comments
very interesting
regarding the error scenario I think sin qua non is the answer. “But for” the error the pitcher would have been able to eliminate the base runner from advancing. The hitter should be penalized.
I agree with everything else accept maybe if there’s runners on 1st and 2nd and the hiiter singles….maybe the batter should just get credit for advancing the baserunner. The hiiter can either advance the runner or not…how many runners he advances should be irrelevant. The idea of a batter walking with the bases juiced getting credit for advancing 3 runners seems to give him more credit than he is due. He did his job in that specific situation and advanced the runner/s.
Ray Fosse thinks Pete Rose is a prick......
Good points
Yeah I guess it doesnt make a diff depending upon how many people are on.
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
I'm not sure I understand the problem with the ESPN definition
If the game is 3-3 in the bottom of the ninth and no one out and the hitter hits it to deep right field thereby allowing the runner at 3rd to score the winning run, I think everyone on earth would call that productive, yet POP would be .000 for that play.
This is incorrect, in the definition you provided an out that advances a runner with no outs is designated as a productive out. In this case the runner moves from third to home on the fly ball so it would be considered a productive out.
Let’s say you had a number four hitter up and they struck out with a man on second and out one. The play would be an 0-for-1 as there were zero productive outs in one potential productive out situation. Yet let’s say this same team pinch hit their pitcher who then laid down a sacrifice bunt, suddenly they would up their POP.
I don’t really buy this criticism because it involves a situation that would never happen in real life. The metric makes specific allowance for pitcher at bats because they carry a different value. Yes, it is possible to game the system but no one would ever try to do it so that isn’t a flaw in the system.
You are right
I made a mistake regarding the first scenario not counting. With respect to the second situation, I agree it would never happen, and no one would game the system just to do it. But the question regarding the validity of the stat is, the argument that pitchers doing it is somehow good, versus another hitter doing it somehow being bad. I think theyre getting at the fact that pitchers aren’t good hitters but what if you had Carlos Zambrano or Mike Hampton trying to get a hit and they didn’t. Would that be a bad thing necessarily that you tried to have them get a hit? If a crazy scenario ends up as a positive, it sort of throws out the validity of the stat in more normal scenarios dont you think?
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
I guess I don't really agree
If a crazy scenario ends up as a positive, it sort of throws out the validity of the stat in more normal scenarios dont you think?
I think any stat can be flummoxed if you come up with a crazy enough scenario. In my opinion the stat still retains its validity unless you can show that the crazy scenarios happen enough to skew the results of the stat. I don’t think having good hitting pitchers bunt occurs enough to say you should throw out the validity of the stat.
The one question I would have with the stat is that it seems like it would over-rate national league teams. By being forced to send the pitcher to the plate National league teams would probably score higher on the productive out scale. But that would likely happen with or without the special exception for pitchers who bunt with one out.
But a less crazy scenario ends up as a positive too
http://www.fangraphs.com/blogs/index.php/the-wrong-way-bartons-bunts/
That article analyses Daric Barton bunting, many of these times are with no out. Yet reduces the win expectancy. Which isn’t a positive. Which makes me rethink my analysis. Because I wonder if sacrifice bunts should count also, because what I am trying to capture is guys who hit to the right side, or put the ball in the outfield when that is necessary… hmmm… now the thing is all of these are outs, so very few (game winning sac flies) would have a positive WPA…
However… the point of what I was trying to say is all of these would be considered very good by the measure of Productive Outs, and it seems like these are not productive, but I suppose it is more opportunity cost.
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
That sounds very sophisticated
There is definitely room for improvement in the ESPN definition, I guess I just felt like the examples you gave didn’t really point to the problem. I probably should have put more effort into digesting the entirety of your post rather than nitpicking the details.
No not at all nitpick away!
Ultimately I am happy to come to the conclusion that what I am seeking to measure is worthless. I am just throwing out the question really, so your feedback giving me things to think about is valuable.
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
Why does a high productive out stat
Necessarily need to increase a teams likelyhood of winning. Any stat, or at least many stats, if pursued single-mindedly would yield negative results.
A team which always tried to steal second, for example, would probably end up leading in this category. But they’d also loose an awful lot of baserunners, and an awful lot of games.
In other words an out can be productive while not necessarily the best thing to pursue. But this can be taken as a given.
by ilikeike on Feb 28, 2011 12:02 AM PST via mobile up reply actions
You've nailed it
in the sense that I am not looking to measure something that is necessarily something you wish to pursue. You don’t want to make outs for the sake of making outs. However, Im hypothesizing that some guys make outs that end up keeping us in a better position to win that do others. Baseball from a hitting standpoint is a game of failure. So I’m seeking to see if some peoples failures cost us more or less.
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
My definition of a "productive out"
is when a runner is in a position to score without the benefit of a hit (at 2B with no outs or at 3B with 0 or 1 out), and the batter advances him, scoring him or keeping him in position to score without the benefit of a hit (while making an out). Thus, the DP with two on and nobody out is a .000.
They do have the most basic ways already of addressing this: “Runner at 3B less than two outs, what % of the time does the batter get the runner home?” It’s about as complex a stat as pitcher wins, but it’s something.
I like Cindi. A. She never pretends to know more than she does. B. She has unbridled enthusiasm for her "Hotties," and isn't afraid to show it. -IM4Oakgal
I think a far more elegant solution would be to harness WPA.
That way, instead of a simplistic yes/no 1/0 kind of thing fraught with inaccuracies and odd side effects, you have a model that is far more nuanced and responsive to out states, runner(s) on base, and so on.
I’d love if somebody could think their way through this formula and point out anything I did wrong, but what about:
(-WPA)/pLI – (-WPA)/LI
(-WPA) is the WPA only for the plays that lowered the team’s chances of winning. This gets rid of all hits, walks, and run-scoring sac flies, and leaves only the non-run-scoring outs. The first term in the formula counts all of the non-run-scoring outs, and divides it by the average leverage index for those negative plays (to normalize the distribution of opportunities, like if some guy had more opportunities with men on second than men on first). The second term takes that figure and subtracts the context-neutral value of those negative plays, which would be a constant multiplied by the number of opportunities (since, in a context-neutral point of view, all outs have the same value).
Essentially, it’s Fangraphs’ Clutch score, but it only looks at the “clutchness” of outs. Would that work?
This has the added advantage of allowing Bob Geren to make important statements like,
“We just really like the clutchness of his outs.”
I like Cindi. A. She never pretends to know more than she does. B. She has unbridled enthusiasm for her "Hotties," and isn't afraid to show it. -IM4Oakgal
Love the comment
But your comment perfectly illustrates the greater question of this thought process which is would this stat have any meaning? I think there’d be value in that sort of hitter who takes a bad situation and makes it more constructive… but yes, we don’t want to start advocating for clutchness of one’s outs.
But I think there is something there in the sense that you always hear people talk (and I’ve heard it most it seems with Ichiro and Hideki Matsui) that they always put the ball where it needs to be in a given situation, i.e. hit it to the right side with a guy on second, etc. I’ve always thought it was part of that larger commentator cliche of Japanese baseball being so “based on the fundamentals” etc but I wondered if theres any weight to it. Do some people make consistently “better outs” than do others?
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
To answer your last sentence, absolutely yes:
Some batters make consistently "better outs" than do others.
However, it’s because some batters have hitting profiles better suited to this particular outcome.
Placido Polanco, a good hitter who rarely Ks, will have tons of PAs where he either reaches safely or makes an out likely to advance a runner.
Wilson Betemit, an “ok” hitter who rarely Ks, has at least some pop in his bat, and usually bats LH, will make a lot of his outs grounding to the right side or flying deep enough to the OF to advance a runner.
Jack Cust and Adam Dunn, on the other hand, will have tons of PAs where they don’t advance anyone because either they walk or K.
Cust and Dunn are much better hitters than Betemit overall, but not necessarily at this one very specific skill. Just as Rajai Davis is the worst hitter of all the ones I’m mentioning, but is the one I’d want at 1B because he has the best skill at stealing and running fast.
I like Cindi. A. She never pretends to know more than she does. B. She has unbridled enthusiasm for her "Hotties," and isn't afraid to show it. -IM4Oakgal
Agreed.
I’m just trying to think what this “statistic” would illuminate for me. What answer it would hold. It seems like itd find your perfect #3 hitter haha.
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
I imagine it would show you that guys who make more contact,
have some power, and hit the ball more often to the right side, more reliably tend to make more productive outs than guys who don’t.
I like Cindi. A. She never pretends to know more than she does. B. She has unbridled enthusiasm for her "Hotties," and isn't afraid to show it. -IM4Oakgal
It seems itd illuminate the type of hitter you and I expect as oppose to any specific skill set
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
Unless I've mis-guessed the profile of a guy with a high "POP"
I’m doing that on assumption/logic, not on any data.
I like Cindi. A. She never pretends to know more than she does. B. She has unbridled enthusiasm for her "Hotties," and isn't afraid to show it. -IM4Oakgal
Yeah as am I
But I’d be very surprised if we didn’t find that to be true.
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
Well if it's normalized for opportunities
And it’s a ratio of their outs to their productive outs than couldn’t it actuLly tell you a lot?
by ilikeike on Feb 28, 2011 12:08 AM PST via mobile up reply actions
hilarious use of the word "elegant"
theres no elegance in baseball!
Buy some class, act like you've been there before.
Sort of...
I had thought about using -WPA, but my concern is that there are some situations (as you mentioned namely the run scoring sac fly) that you really truly would want to capture. I like the concept with maybe it being (-WPA + the +WPA of SF’s) which gets messy in a hurry, but I think would sort of work right?
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
Well, this thing would be hell to calculate anyway,
because you’d have to not only find the LI of every negative WPA play, but also average the LI’s of the negative WPA plays for the pLI variable.
So maybe not (-WPA), but just, say, oWPA, the WPA of all of the outs?
Absolutely
I was thinking as a do it as the season goes project. But, to clairfy, as I wrote to Nico above: "would this stat have any meaning? I think there’d be value in that sort of hitter who takes a bad situation and makes it more constructive… but yes, we don’t want to start advocating for clutchness of one’s outs.
But I think there is something there in the sense that you always hear people talk (and I’ve heard it most it seems with Ichiro and Hideki Matsui) that they always put the ball where it needs to be in a given situation, i.e. hit it to the right side with a guy on second, etc. I’ve always thought it was part of that larger commentator cliche of Japanese baseball being so "based on the fundamentals" etc but I wondered if theres any weight to it. Do some people make consistently "better outs" than do others?"
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
To me, the skill is not in the ability to intentionally bounce to 2B.
The skill is in waiting for, and swinging at, a pitch you are likely to make a productive out on, if you make an out at all, so that you either “succeed” (by getting a hit) or “succeed” (by at least making a productive out).
For example, with a runner at 2B and nobody out, Ryan Sweeney should be looking for pitches on the inner half to swing at — even though with a runner at 3B and two out, he should take advantage of a fastball away to put his unusually good “poke it to LF for a single” swing on it.
I like Cindi. A. She never pretends to know more than she does. B. She has unbridled enthusiasm for her "Hotties," and isn't afraid to show it. -IM4Oakgal
Yeah that was what I wrestled with
What does it say about those guys who do get the job done?
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
Pshhh, as if Ryan Sweeney can ever pull a ball even if it is an inside pitch.
"We were shit, pathetic," Guillen growled early in spring training. "We hit too many home runs."
by lenscrafters on Feb 26, 2011 9:48 PM PST up reply actions
Waiting for and swinging at a pitch that you want to bounce to the right side?
That seems weird. Why wouldn’t you just try to get a base hit instead? A line drive is more likely to be productive than a bouncer.
Official Athletics Nation Rotating Tagline Editor - Pam liked my old sig better.
My thoughtful watermelon is easily mistook for an early American catapult.
My guess is that formula would shed some light on that.
The “clutchness of the outs” phrase may not have been the best way to put it. But what it will do is see if people make “better outs” than others.
Wouldn't a very, very, very likely answer be:
“Yes, they do, but that doesn’t mean they’re better hitters — just relatively good to have up in those particular situations”?
I like Cindi. A. She never pretends to know more than she does. B. She has unbridled enthusiasm for her "Hotties," and isn't afraid to show it. -IM4Oakgal
It definitely won't say anything about who's a better hitter.
I’d guess that Cust and Dunn and co. would fare far better than you might expect, because they never GIDP. And a GIDP probably wipes out the positive effects of three or four sac flies.
And in either case, I’d also guess that any positive contributions given by any productive outs are absolutely dwarfed by, you know, actual hitting.
I wonder if over the course of the season though productive outs add up to something meaningful
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
Hmm.
This is something that I’d love to see. And it should be readily testable too, given a pretty sizable chunk of time and effort.
I'd do it, but I'm already in the middle of another big project,
which is going to hospitals to track whether there’s really such a thing as a “productive cough.”
I like Cindi. A. She never pretends to know more than she does. B. She has unbridled enthusiasm for her "Hotties," and isn't afraid to show it. -IM4Oakgal
One reason Cust/Dunn might fare better
is that they don’t get penalized for walking. It’s the guys who K a lot but don’t BB who kill you, especially if they tend to hit into DPs. I’m looking at you, Kouz.
I like Cindi. A. She never pretends to know more than she does. B. She has unbridled enthusiasm for her "Hotties," and isn't afraid to show it. -IM4Oakgal
Dwishinsky- the productive outs that seem
to be the most important in tracking are: runner on 3rd less than 2 outs..runner on 2nd with no outs or runner on 1st with no outs…but what about the baserunner? take the runner on 2nd with no out scenario- if the hitter hits a ball to the SS, BUT the runner follows the rule and is able to get ahead of the grounder so the grounder passes behind the runner or to his left thus allowing him to make it to 3rd…do you think the hitter should still receive credit for the runner’s heads up play? Same for the runner who tags on a shallow fly..are we just concerned with the result?
Ray Fosse thinks Pete Rose is a prick......
Yeah it is difficult to isolate one from the other which is why I think I'm going to go with a WPA based format
Though I think there still may be a bit of that at issue.
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
This is a great post...
A couple things that need to be thought out….
1) does there need to be a distinction between voluntary productive outs and involuntary? Bunting, for example, is a voluntary out.
2) who gets credit for the productive out? Sure, the hitter laid down the bunt or hit a grounder to second moving the runner along, but what if the manager told him to? Certainly, the manager plays a big role in this.
3) should the study be a cost benefit analysis? When Barton bunts Davis to third, he, or the manager, should get credit for the so-called productive out, but at what cost? Certainly, there’s a cost to voluntarily making an out, because someone like Barton has a 40% of getting on base anyway. So is the cost of the voluntary out more than the benefit?
Agreed.
Given that I think think I am going to try the WPA based formula proposed by danmerqury. We shall see what shakes out, as this season I will be keenly following all at bats with runners on and less than two outs! (hopefully there are many of those!)
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
This seems like the point of trying to pin down this stat
If he has a 40% chance of getting on base, but, say, a 90% chance of making a productive out, than might it make sense? I hav no idea…
by ilikeike on Feb 28, 2011 12:14 AM PST via mobile up reply actions
I think I'm beginning to get a handle on this.
The hardest part will be determining values based on runner location and number of outs created. Where the best result will be a bases loaded, no out sac fly where everyone advances, while the worst result would be a bases loaded, no out triple play. A two on, no out triple play is not quite as bad as the bases loaded triple play because more potential runs were destroyed in that case. Advancements incurred by errors should not positively count toward the batter. Basically, the outs should be judged on how the base situation was before and after the batter created the out(s). If the runners end the at bat in the same location as they were before the at bat, that should be a neutral out. If the runner situation ends up worse, that’s negative. And if the runner situation improves, that’s positive. When we can determine 1.) how to record these occurrences, 2.) what values to assign for all outcomes, and 3.) how to convert to a coefficient that can be applied to something like wRC+ or WAR, then we have really got something.
Try run expectancy
The table is at http://www.tangotiger.net/RE9902.html
Then all you need to do is look at the run expectancy (RE) after the productive out and the runner(s) has moved and subtract the run expectancy for if the runner(s) had stayed and there had been 1 more out.
Example: bases loaded, no outs at start of AB
Force out at first, run scores, runners move up : 1 (run) + 1.467 (RE 2nd_3rd 1 out)
Subtract 1.65 (RE bases loaded 1 out)
Hits into double play, run scores, runner at third : 1 (run) + .387 (RE 3rd 2 outs)
This is not a productive out as the RE is less than if the batter struck out.
Thanks
I am going to have a look and see how that works out too. I think Im goign to create a massive spreadsheet with all the culled data from 2011 of A’s hittign with men on and less than two outs. i will be the authoritative source! haha
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
Comments For This Post Are Closed
Secondary Sidebar
User Tools
Related FanPosts on Athletics Nation
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
Baseball Nation Recent Stories
In Association With
Big League Stew - MLB - Yahoo! Sports
- The Juice: Long Adam Dunn homer punctuates Chris Sale’s 15 Ks, Mark Trumbo goes deep after Jered Weaver goes down for Angels
- Leyland begs media to make umpires ‘accountable’ for bad calls (VIDEO)
- Get a grip: Reds’ Todd Frazier loses bat, still connects for home run (Video)
- Eric Hosmer redefines the Baltimore chop with infield double (Video)
- College player makes impossible over the fence catch to save tournament game (Video)
