Thursday, May 23, 2013
Thursday, May 23, 2013

Phillies

Bases Created, Outs Made, OPS, and the Phillies (and You!)

The latest Phillies and baseball news from Philly.com.

Bases Created, Outs Made, OPS, and the Phillies (and You!)

POSTED: Monday, June 18, 2012, 1:46 PM
(Elise Wrabetz/Staff Photographer)

One of the tricky parts about my job is straddling the line between accuracy and accessibility when using statistics. A lot of people know what OPS means and what it represents. But a lot of people don't, and both groups of people want to read about the Phillies. Sometimes I get emails complaining about my use of RBIs and pitcher wins in a story. Other times, I get emails complaining about my use of OPS and WHIP. Sometimes, I get both types of emails in the same day.

The problem with "counting" stats like RBI and hits and strikeouts is the fact that they do not consider the big picture, like how many runners a hitter had on base in front of him, or how many plate appearances the hitter had. The problem with a lot of percentage stats is that they are not intuitive. The difference between a .260 batting average and .300 batting average is four hits for every 100 at-bats, which is maybe 24 hits over the course of a season. What, exactly, does that mean in the context of an individual plate appearance? Try explaining the same for on base percentage and slugging percentage. Fast forward all the way up to a less accessible stat like Wins Above Replacement and you can understand why a lot of people just don't want to be bothered with it.

But what if there was away to lessen the disconnect by finding a way to quantify the events that really impact the outcome of a game in an intuitive fashion? Linear weights, the backbone of WAR-related stats, just aren't intuitive enough for a lot of people.

Here's my attempt. It comes with a disclaimer: I am not a math whiz. In fact, I hated math all the way up through college. But the following makes sense to me, and from my perspective it provides a relatively accurate and intuitive measurement of offensive performance. Maybe some form of this already exists. Whatever the case, it killed some time.

1) Outs, not hits, define success and failure.

Part of our problem lies with our desire to measure a player's success at the plate. But baseball is a game of failure. A game is not decided by a team's batting success. It is decided by a team's batting failure. A hit does not bring a game closer to its conclusion. An out does. Every out that a hitter makes brings his team closer to defeat. So why don't we measure performance with outs as baseline?

2) Two potential outcomes

Every time a hitter steps to the plate, he can create one of two outcomes: out creation, or base creation. This is the basic concept behind OPS: add the percentage of plate appearances that a hitter reaches base to the number of bases he creates per at-bat (single equals 1 base, double equals 2, etc.) and you get his OPS. The flaw, of course, is the fact that you are adding two fractions with different denominators (Times on Base divided by Plate Appearances, and Total Bases divided by At Bats).

Or. . .

TOB/PA + TB/AB

The basic rules of math say that we are not allowed to do what we do when calculating OPS. Which is why it is a flawed statistic.

3) Outs as the common denominator

To me, the simple solution is to use the one potential outcome that is common for all hitters as the denominator. In other words, the solution is to measure a hitter's production, or the bases that he creates, against the outs that he creates. After all, every out is worth the same amount: 1/27th of a team's allotted time at the plate. When a player doesn't make an out, he both creates offense and extends the amount of time his teammates have to score. When he makes an out, he decreases the amount of time his teammates have to score.

A player's value, then, is the offense he creates, or the bases he accumulates, measured against the outs he makes. The question every general manager should ask himself when he adds a hitter is, "Are the bases that this player creates worth the outs that he makes?"

Or, "How many bases-per-out does this player create?"

To answer the question, you simply divide a player's Total Bases Created by his Total Outs Made.

Think about it in this light: In the first inning of a game, Ryan Howard and Pete Orr both make outs during their at-bats. Both at-bats result in the elimination of 1/27th of the outs the Phillies are allotted for the game. But sheer intuition tells you that Ryan Howard is more valuable to a lineup than Pete Orr. And when you measure the offense he produces against the outs that he makes, the result tells you the same thing.

4) Calculating Bases Created

Total Bases is already a common statistic, one that counts a single as one base, a double as two, a triple as three, and a home run as four. But it doesn't include walks and hit-by-pitches. So just add a player's BBs and HBPs to his TBs and you get the total bases he has created.

Take a player's plate appearances and subtract the number of times he has earned his way onto base (Hits + BBs + HBPs) and you get the number of outs he has made.

Just divide. So, for example, Carlos Ruiz has creates 129 bases while making 125 outs this season. Divide 129 by 125 and you find that he has created 1.032 bases-per-out. Now take Pete Orr, who has created 19 bases while making 32 outs for a .594 bases-per-out ratio. In other words, for every out that Carlos Ruiz has made, he has created 1.032 bases, while Orr has only created .594 bases. Which, layman's terms, means an out by Carlos Ruiz is almost as justifiable, or almost as easy to swallow, or almost half as detrimental, as an out by Orr. Assuming, of course, they maintain their respective ratios.

The number looks similar to OPS, but in my mind it more accurately combines a player's on base percentage (or out-avoiding ability) with his slugging percentage (or base-creating ability). Ruiz, for example, has a .982 OPS against a 1.032 BCOM, which suggests that OPS undervalues Ruiz, while Orr has a .731 OPS against a .594 BCOM, which suggests that OPS overvalues Orr. The difference, in this case, lies in their on base percentages: Orr's is .302 while Ruiz's is .416.

7) A quick test

Obviously, a pitcher's success is based on the same two principles as a hitter's. Except the goal is to make as many outs as possible while allowing as few total bases as possible. To judge the accuracy of our little formula, I took the hitting and pitching performances of each team in the National League from 2011 and subtracted their Bases Allowed/Out Made (pitchers) and their Base Created/Out Made (hitters) and compared the result against their winning percentage. To give us a number that is easier to compare, I multiplied both sides by 27 outs. In other words, the table below shows the total bases an offense created over 27 outs (the average game, in other words), and the total bases a pitching staff allowed over 27 outs. I then compared their NL rank in that category to their NL rank in winning percentage.

 Team BA/27 Outs BC/27 outs Difference Win Perct. NL WP% Rank 1. Phillies 15.309 17.665 +2.356 .630 1 2. Brewers 16.486 18.537 +2.051 .593 2 3. Cardinals 17.306 19.140 +1.834 .556 4 4. DBacks 17.657 18.438 +0.782 .580 3 5. Braves 16.116 16.809 +0.693 .549 5 6. Dodgers 16.447 16.774 +0.328 .509 7 7. Giants 15.759 15.957 +0.198 .531 6 8. Nationals 16.826 16.681 -0.145 .497 8 9. Reds 18.437 18.157 -0.279 .488 9 10. Marlins 17.525 17.231 -0.294 .444 12 11. Mets 18.303 17.911 -0.392 .475 10 12. Rockies 19.179 18.391 -0.788 .451 11 13. Padres 16.705 15.515 -1.190 .438 14 14. Cubs 18.720 17.314 -1.406 .438 15 15. Pirates 18.516 16.113 -2.403 .444 13 16. Astros 19.452 16.088 -3.363 .346 16

So no team's rank in the difference between Bases Allowed and Bases Created per 27 outs (which is the same as their rank in the difference between their BAOM and BCOM), was more than two slots better or worse than their final rank in overall record.

This year? The Phillies' differential ranks 11th. Which should jibe with what you've seen.

8) The Phillies' offense in 2012

Here is the rundown of the BC/OM of each Phillies hitter (min. 60 PA), with their OPS in parentheses:

1. Carlos Ruiz 1.032 (.982 OPS)
2. Jim Thome .857 (.844 OPS)
3. Hunter Pence .799 (.822 OPS)
4. Shane Victorino .665 (.721 OPS)
5. Ty Wigginton .658 (.715 OPS)
6. Juan Pierre .624 (.734 OPS)
7. Brian Schneider .596 (.675 OPS)
8. Jimmy Rollins .575 (.688 OPS)
9. John Mayberry Jr. .560 (.652 OPS)
10. Placido Polanco .545 (.617 OPS)
11. Freddy Galvis .507 (.617 OPS)
• 0 like this / 0 don't   •   Posted 1:52 PM, 06/18/2012
not a stat guy, but interesting. any consideration for sacrifices?
phils_tnj1
• 0 like this / 0 don't   •   Posted 2:13 PM, 06/18/2012
That is a good point. Sacrifices should be taken into account. For each sacrifice you should add 1 to the numerator(Bases Created) and 1 to the denominator(Outs Made). It is addition and not multiplication, (2+1)/(3+1) instead of (2/3)*(1/1), so it would have an effect on the BC/OM percentage.
edjsull925
• 0 like this / 0 don't   •   Posted 2:09 PM, 06/18/2012
I was wondering the same thing. If you sacrifice a guy, do you still get 1 base created (since you advanced a runner)? Might be too hard to research that though.
MattMullin
• 0 like this / 0 don't   •   Posted 8:55 PM, 06/18/2012
If you sacrifice a guy, you get to bring a new player up from the minors.

Perhaps this is why the Phils are avoiding sacrifices - they might lose their jobs, literally, not figuratively.

• 0 like this / 0 don't   •   Posted 2:16 PM, 06/18/2012
Cheese this breakes down even further players OPS. this is the way it should be to show production for all at bats. as the stat says only Thome and ruiz are showing better than their OPS % and if you watch the games that is all backed up. This also gives you some insight on who's hitting with men on and what the outcomes are. The only two that suprised me are Poppins and Pence who only seem to get hits with no one on base. Thosa guys and Vic i thought would have a more negative Bcom vs. OPS.
Trot
• 0 like this / 0 don't   •   Posted 2:17 PM, 06/18/2012
One last thought, what about sacrifices and what about hitting into double plays? One should be neutral or positive and the other is a negative..right?
Trot
• 0 like this / 0 don't   •   Posted 2:19 PM, 06/18/2012
What about stolen bases,caught stealing, and base running blunders(getting picked off, caught stretching a single into a double)?

When you steal a bag you are essentially turning a single into a double, double into a triple, etc.

Caught stealing or is getting picked off is an out just like any other out.

edjsull925
• 0 like this / 0 don't   •   Posted 3:19 PM, 06/18/2012
For the same situational-dependent reasoning, you should account for sacrifices (i.e., a hitter doesn't have control over the fact that the manager asked him to sacrifice to move up a runner). I also don't think it should be a positive to convert a sacrifice (again, because you want to really limit this stat to batting events the player has control over), so you should just take ABs (which already don't include BBs, HBPs, and SFs) and subtract hits to get number of outs.

This would really give you a better sense of a hitter's productivity for all opportunities in which the batter is at the plate just to try to get a hit.
• 0 like this / 0 don't   •   Posted 3:41 PM, 06/18/2012
Also, a related statistic is Total Average, but I like your approach better, which is essentially Total Average without steals and GIDP: http://en.wikipedia.org/wiki/Total_average
• 0 like this / 0 don't   •   Posted 5:15 PM, 06/19/2012
Murph, how exactly did you breakdown your calculations to get 17.665 BC and 15.309 BA? Based on the formula you used, the Phillies offense created 2,797 bases in 2011 (2,202 TB + 539 BB + 56 HBP). The Phillies pitching staff allowed 2,437 bases in 2011 (1,986 TB + 404 BB + 47 HBP).
sferrell710
• 0 like this / 0 don't   •   Posted 2:52 PM, 06/18/2012
I've always had a problem with OPS because it undervalues On Base Percentage, or the ability to create bases, instead of outs. I like this formula because it more accurately depicts the flow of the game. If you've watched every Phillies game for the past two years, it's nearly impossible to argue with this formula. Last year, as evidenced by your chart, they had hitters who created more bases than outs, resulting in a lot of victories. This year, they are hitting better than last year (at least up to this point of last year - average wise) yet they obviously have a much worse record. How can that be? Well pitching has a lot to do with it, but it also has a lot to do with the number of outs guys are creating, specifically outs when they need a hit the most. It's fascinating to see how closely related that chart is to the success of a baseball team.
jacobrepko
• 0 like this / 0 don't   •   Posted 3:05 PM, 06/18/2012
Super column and it really is just as simple as that. Now if only Amaro and Manuel could figure this out and stop bringing in guys who are weak in creating offensive production per out (Polanco, Pierre, Nix, Wiggington). I'd like to see hitting expressed in terms of failure as you suggest. Instead of on-base-percentage = .300, an outmaking percentage = .700 stresses the negativity of making outs. In 2009 Jimmy Rollins was one of the 'best' outmakers in the NL and in 2011 Raul Ibanez also had that distinction (something he would be vying for again in the AL if he had enough at bats). Great job on trying to help fans focus on what is valuable!
jtj10
• 0 like this / 0 don't   •   Posted 3:18 PM, 06/18/2012
Dave: did you figure times reaching base via a fielder's error into the equation?
Chester Goode
• 0 like this / 0 don't   •   Posted 3:22 PM, 06/18/2012
I learned a long time ago to separate "good" hitters who hit the ball hard, to all fields even if they make an out, from "poor" hitters who hit a lot of pop ups and weak ground balls because they are always trying to pull the ball.
farley
• 0 like this / 0 don't   •   Posted 3:25 PM, 06/18/2012
This crappy season has come to this? An analysis of a different way to micro analyze statistics? The season really is over, isn't it?
thingfish
• 0 like this / 0 don't   •   Posted 3:25 PM, 06/18/2012
analyze this: Phils suck -- their g.m. has no clue
warbiscuit
• 0 like this / 0 don't   •   Posted 5:38 PM, 06/18/2012
@war - another incredible to the point post. The only thing you missed, war, was a comment on the manager.

The stats on 'sucking', however, are not incontrovertable. They are in fact, subjective, and so much so that you might not have as many people believe the 'suck' data.

Now, on the GM 'no clue' part: that IS data: no = 0 therefore, there is zero chance in your opinion that the GM has clues. And no chance at clues. So, therefore, the GM is clueless.

Well done.

Next point.....
• 0 like this / 0 don't   •   Posted 3:25 PM, 06/18/2012
I am now in a deep sleep. Thanks.
1danny
• 0 like this / 0 don't   •   Posted 3:38 PM, 06/18/2012
Yeah, after five or six beers, a thoughtful analysis can put you to sleep.
• 0 like this / 0 don't   •   Posted 3:32 PM, 06/18/2012
my brain exploded. Thanks, DM!
palmyra21
• 0 like this / 0 don't   •   Posted 3:43 PM, 06/18/2012
Excellent column. Sacrifices and stolen bases should be included in the total. Another factor that needs to be considered is how a player performs with RISP. That has a lot to do with teams performance in terms of wins and losses. Same points need to be considered when evaluating the pitcher's side of the equation. Again, good job.
hwsmith316
• 0 like this / 0 don't   •   Posted 3:51 PM, 06/18/2012
Here's my key indicator stat which you do not mention - performance in the late innings of a close game. One year, I did a breakdown of Ryan Howard's performance using this stat, which is in the mlb.com website. He's great at hitting 3-run HR's in the first few innings when there's many innings to go and little pressure, but terrible in the late (or extra) innings of a pressure-filled game when he has to deliver a key hit under the microscope. The OPS of a guy who can deliver in clutch time is different than the OPS of a player who hits bigtime in blowout games. Using this stat, Ryan Howard is not Ryan Howard the \$125MM man.
hairball
• 0 like this / 0 don't   •   Posted 3:57 PM, 06/18/2012
Great stuff!!! Off the top of my head, I like some of the suggestions being made for including sacrifices etc, ...and I'm also wondering if the stat might be even more reflective of a players contributions if the format were changed from the "bases created" stat which counts only the bases reached by the batter, to a "Bases contributed" format which would not only give the batter credit for the bases he created by hitting a single (1), but would also give him credit for the bases advanced by whichever runners were on base and advanced by the hit, ...so a single with the bases loaded would count as 4. A sacrifice fly with a man on third would be 1, a grand slam would be 10, a GIDP a -1 etc...
andyd
• 0 like this / 0 don't   •   Posted 3:58 PM, 06/18/2012
Great stuff!!! Off the top of my head, I like some of the suggestions being made for including sacrifices etc, ...and I'm also wondering if the stat might be even more reflective of a players contributions if the format were changed from the "bases created" stat which counts only the bases reached by the batter, to a "Bases contributed" format which would not only give the batter credit for the bases he created by hitting a single (1), but would also give him credit for the bases advanced by whichever runners were on base and advanced by the hit, ...so a single with the bases loaded would count as 4. A sacrifice fly with a man on third would be 1, a grand slam would be 10, a GIDP a -1 etc...
andyd
• 0 like this / 0 don't   •   Posted 4:25 PM, 06/18/2012
The reason to avoid the "bases contributed" is: how do you handle a single with a man on second, or even a man on first. It is often the baserunner's speed, and not anything related to the batter's performance, that determines whether the runner scores from 2nd or advance to 3rd from 1st.

If you like the bases created idea, though, you might like Weighted On Base Average (wOBA), which takes all aspects of hitting and weights them in proportion to their actual run value (i.e., how much each proportionally contributes to a run scored): http://www.fangraphs.com/library/index.php/offense/woba/
• 0 like this / 0 don't   •   Posted 4:36 PM, 06/18/2012
OPS essentially double counts base hits, even if you ignore the incorrect mathematics of adding with differing denominators. So, in the component of slugging %, if you just subtracted base hits from the numerator, it should capture the intent of what Murph reflects here. So, just take OBP + (Total Bases - base hits) / PAs = Adjusted OPS. This would reduce the relative implied value of a singles hitter like Juan Pierre or Pete Orr, and inflates the relative value of hitters like Ruiz or Howard.
FakeName
• 0 like this / 0 don't   •   Posted 5:28 PM, 06/18/2012
Murph, this is a good way to evaluate players instead of just looking at there batting averages. Batting averages arent broken down near enough and it makes some players look better than they really are. For instance, 280 average vs 260. The guy hitting 280 looks better because of the number is bigger, but the guy hitting 260 walks more and scores more runs which in the end, makes him more valuable for the team then the other guy. Its kind of interesting if they will adopt this method because you should get rewarded for helping the team win rather than putting up just a look at me batting average!
huntnmike3666
• 0 like this / 0 don't   •   Posted 5:29 PM, 06/18/2012
Yankees fans don't like stats such as these because they show Derek Jeter and Don Mattingly in a not too favorable light.
bigeastbeast
• 0 like this / 0 don't   •   Posted 5:43 PM, 06/18/2012
Excellent! Well thought-out. I agree with the suggestions to score sacrifices more liberally and deduct for double plays. Thanks for a very interesting read.
CrotchetyOldMan
• 0 like this / 0 don't   •   Posted 6:22 PM, 06/18/2012
This wasn't killing time, this was wasting time. You could have spent your journalistic skills investigating why Monty and Giles play their game of duplicity with the paying fans and press and why they are clearly demonstrating a lack of interest in solving the Phillies descent into ignominy. Its obvious by the fact they are not firing Cholly, his coaches and not bringing up Domonic Brown, that they have an ulterior motive. Where are Woodward and Bernstein when you need them?
daystrum
• 0 like this / 0 don't   •   Posted 8:52 PM, 06/18/2012
Why bother? There are no stats, no data available for that sort of inquiry. And why hire Woodward & Bernstein when you can get the Pink Panther for that investigation, or Get Smart?

MontCo, last I looked, was the owner-in-charge of this team. He has a powerful group of 'investors' behind and aside him. As they continue to print money, why change?

They epitomize conservative managerial ethics. Do nothing until you absolutely have to. Then write off the losses and write off your customers. GM did the same thing.

Personally, I like this column. It takes my mind off of things I can't control. And takes my mind off of a team that currently epitomizes pathetic. Sadly, as a fan, I can't watch.
• 0 like this / 0 don't   •   Posted 9:32 PM, 06/18/2012
Well if the answers are so obvious, why should he do a column on it?
cloudkitt
• 0 like this / 0 don't   •   Posted 6:22 PM, 06/18/2012
This wasn't killing time, this was wasting time. You could have spent your journalistic skills investigating why Monty and Giles play their game of duplicity with the paying fans and press and why they are clearly demonstrating a lack of interest in solving the Phillies descent into ignominy. Its obvious by the fact they are not firing Cholly, his coaches and not bringing up Domonic Brown, that they have an ulterior motive. Where are Woodward and Bernstein when you need them?
daystrum
• 0 like this / 0 don't   •   Posted 6:56 PM, 06/18/2012
@BJG83 hit it: wOBA is really the Sabermetrics stat that attempts to address a lot of this. wOBA is, in my opinion, a better metric than OPS. But Murphymetrics isn't too bad an idea. I would definitely factor in sacrifices somehow, although wOBA does not. But it factors in more than OPS and if you look at wOBA leaders they're the typically the top hitters you would expect. wOBA weights different plate events which (being an old math guy) is the right idea. Taking that approach and Murphy's idea, and some of the comments you could probably come up with an enhanced wOBA that factors in more. For example you could add a weighted value for sacrifices and deducts some weighted value for outs. This exercise is left to the reader but it will be on your final exam.
s
• 0 like this / 0 don't   •   Posted 8:25 PM, 06/18/2012
I eagerly await the "Murphyball" book and movie. Don't forget us on the way up David.
s
• 0 like this / 0 don't   •   Posted 8:47 PM, 06/18/2012
I did some rough calculations and this shows what most sense, that Carlos Ruiz has been the Phillies' most valuable hitter the last three years. It's funny how the hitting guru always bats him at the bottom of the order. Overall, I think it would be a good idea to express all of the counting stats in terms of "per outs made." It drives me nuts when a guy like Rollins has 725 plate appearances (as he did in 2009) and scores only 100 runs and dimwitted fans, Phillies' management, and much of the press think he is a catalyst because he scored the magic number 100 runs, despite being a lead off hitter with the third worst OBP among all regular players in the NL.
jtj10
• 0 like this / 0 don't   •   Posted 10:45 PM, 06/18/2012
Thanks for the article Murph. Good numbers. Now can you do one that is just as comprehensive on the pitching? :)
Cheers.
zubzub
• 0 like this / 0 don't   •   Posted 10:45 PM, 06/18/2012
Thanks for the article Murph. Good numbers. Now can you do one that is just as comprehensive on the pitching? :) Cheers. (HTML deleted)
zubzub
• 0 like this / 0 don't   •   Posted 11:40 PM, 06/18/2012
If you look at the stats for the three areas -- pitching, batting, fielding -- it's really the pitching that's been the clearest problem. Both starting and relief pitching are way down from last year's pace. While you couldn't realistically expect a repeat of last year's pitching you also wouldn't expect the sort of drop we're seeing. Some of the hitting is actually better, although maybe not situational enough (although Murphy also had some stats questioning that). The fielding sure feels a lot worse and the standard fielding numbers are down. But some of the Sabermnetrics fielding numbers are actually better, which is mind boggling.
s
• 0 like this / 0 don't   •   Posted 10:59 PM, 06/18/2012
I get a chuckle out of the fixation with bringing up JaMarcus Russell...oops, I mean Domonic Brown. The Phillies had a speedy left handed hitting outfielder in camp that they practically spit on and then just kicked to the curb. He's now hitting over .350 in Boston. They had another lefty outfielder who was the MVP for LV last year who's now hitting bombs for Oakland. No thanks, we opted for Bowker instead.

Regarding this statistical analysis: pretty heady stuff for a game played by such mental giants as Victorino.
dasher
• 0 like this / 0 don't   •   Posted 12:18 AM, 06/19/2012
stoky
• 0 like this / 0 don't   •   Posted 2:24 AM, 06/19/2012
while I like the idea, it forgets one very important fact. RUNS!! lets say team A has 12 total bases in 9 innings and team B has 21. Based on Murph's stat you'd think team B was better, however, team A hit 3 HRs and scored 3 runs. Team b spread a few doubles, walks, and singles over the game and ddin't score. Who wins? I suggest adding Runs and runs produced into formula somehow. We have all seen this year that being able to creat a base with a man on 3rd is more important to the team then creating one with no one on.
jeff gross
• 0 like this / 0 don't   •   Posted 5:27 AM, 06/19/2012
Like it. It always bugged me on some level that OPS was adding two different percentages. Thanks, David.
Robbo
• 0 like this / 0 don't   •   Posted 9:19 AM, 06/19/2012
I always maintained that Placido Polanco is one of the most overrated offensive players of the last decade or so, without looking at those numbers. His BA looks good all the time, but he's not the great hitter people think he is.
EL Zorro
• 0 like this / 0 don't   •   Posted 8:05 AM, 06/20/2012
A couple of points: 1) As interesting as the BC/OM stat is, when Murph puts it side-by-side with OPS he comes up with two very similar assessments of the Phillies' lineup, except for slight anomalies with Pierre and Rollins. So is OPS really all that far off-base? 2) Some batters are more likely to cause fielders' errors than others. Those who routinely hit the ball hard -- line drives and sharp ground balls -- are going to get on base through errors that, arguably, they forced. This consideration should be factored in to any BC stat, but it would be awfully hard to do.
Dave Clemens

1
Popular Posts