Yesterday on RAB Joe Pawlikowski wrote a post asking if and were of equal offensive value because prior to Tuesday’s game they both had an OPS of .779. The post is excellent, and worth reading in its own right, but to summarize here the answer is that Swisher is the better offensive player because of his power. It is just a coincidence that Gardner’s and Swisher’s lines had aligned to that point in the season.
To dig into the numbers, Joe also pointed out that Nick and Brett had produced similar wOBA’s to that point in the season. Swisher had hit to a .345 wOBA entering Tuesday’s game, while Gardner had hit to a .344 wOBA. This too, however, was more of a coincidence as well. At a certain point Swisher’s power should allow him to produce a higher wOBA and OPS than Gardner, but both are valuable players who produce offense in different ways.
To help illustrate this point, Pawlikowski referenced an article on fangraphs that Matt Klaassen wrote. In the article Klaassen asked whether or not AVG/OBP/SLG are still valuable now that wOBA and similar stats have become so popular. Klaassen correctly argues that the slash line remains valuable, because it serves a different purpose than wOBA. wOBA is valuable because it tells an analyst which player produces offense at a better rate, while the slash stats are valuable because they explain how a given players produces that offense, for example through getting on base or through hitting for power. Neither stat (or stat line) alone tells the full story.
Both Klaassen and Pawlikowski were spot on in their analyses. My purpose here is to add an element that reinforces why it is smart to provide both a player’s slash line and his wOBA when describing performance. wOBA on Fangraphs is calculated as follows, :
((.72*NIBB)+(.75*HBP)+(.9*1B)+(.92*RBOE)+(1.24*2B)+(1.56*3B)+(1.95*HR))/PA
Where the less well known acronyms NIBB and RBOE stand for non-intentional walk and reached base on error, respectively.
The equation holds the secret to why the slash line still has considerable value when describing a player, compared to wOBA alone. Simply put, the bulk of wOBA’s mathematical components are also components in each of the slash stats. There is almost nothing in wOBA, for example, that is not in one of the individual slash stats. As a result, when the slash stats are presented together, but not blended, mathematically you have a series of numbers that explain almost precisely what wOBA does, by design.
The difference between the two, of course, is that wOBA is weighted. It assigns a logical value to each of the events that occurs during an at-bat. The slash line doesn’t do this, but it doesn’t need to. OBP describes one aspect of a player’s game, while SLG describes another. (Of the three, AVG is the least important.) Left separately, these two statistics tell us a lot about how a player produces offensively. While they may not reduce to a single number, as elegantly as wOBA does, they actually give someone evaluating an entire lineup more to work with because they work to explain where a high (or low) wOBA comes from. These two methods of evaluation will virtually always point in the same direction because they are different ways of presenting almost the exact same numbers.
The statistic that actually obscures value is OPS. That’s because OPS purports to do what wOBA does, but applies no intelligence to how it weighs a hit. Doubles count twice as much as singles, while SLG is equal to OBP. This is wrong. For example, a player with a .300 OBP and a .500 SLG is a productive hitter, but is not as valuable as a player with a .400 OBP and a .400 SLG, even though they both have an .800 OPS. The player with the higher OBP is more valuable. OPS may be fun because it rounds to nice, easy to understand numbers, but it is incorrect to add OBP and SLG to create a stat to draw comparisons across players. The best way to do that is to use both wOBA and the slash line.
If you enjoyed this article, please consider sharing it!
I’m having trouble reconciling these two statements:
1-The post is excellent, and worth reading in its own right, but to summarize here the answer is that Swisher is the better offensive player because of his power.
2-For example, a player with a .300 OBP and a .500 SLG is a productive hitter, but is not as valuable as a player with a .400 OBP and a .400 SLG, even though they both have an .800 OPS.
I agree with the second and completely disagree with the first. Also, yes scoring is down, but it’s not down to the point that makes consecutive offense less valuable. Gardner getting on base is just as important as it would have been in 2003, (and speed on the basepaths has a lot of value in run-scarce environments, anyway).
[Reply]
Mike Jaggers-Radolf Reply:
June 29th, 2011 at 1:55 pm
Hi Oldpep, thanks for the comment.
I see why you’d find those two comments confusing. Let me provide some additional context for clarity.
The first comment you highlight should probably read to say that given that both Swisher and Gardner get on base at roughly the same clip this year, Swisher is actually the better offensive player. He and Gardner currently have the same OPS and OBP, but at season’s end Swisher will in all probability have a higher SLG than Gardner, if his SLG is not already higher after last night’s offensive explosion.
Even with the added value of steals (which are not included in Fangraphs wOBA, at least according to their calculator) Swisher’s increased power makes him, in all probability, the better offensive player. It is also why it is so uncommon for Swisher and Gardner to have similar OPS’s or wOBA’s. If Swisher had gotten off to a more normal start the RAB post probably never happens.
I’m happy to say that I took on the risk of being glib in summarizing the first article, and could have done a better job.
With regard to point 2, the one you agree with, that is as close to a mathematical fact as you can find. Before writing that as an example I actually checked the numbers using the Hardball Times metric GPA (which weights between just OBP and SLG). It held true, so I went ahead with using the example.
It is important to note, however, that these two hypothetical players .300/.500 and .400/.400 are NOT Nick Swisher and Brett Gardner, respectively. As I mentioned earlier, Swisher and Gardner for the past two seasons have had must closer OBPs than the hypothetical players in this example. As a result, power and steals will be the ultimate arbiter between which is better in a vacuum.
Hope that helps, and hope you enjoyed the post otherwise.
[Reply]
TedK Reply:
June 29th, 2011 at 3:17 pm
Mike, could you explain a little more why you say OBP is more valuable than SLG? I thought it was the other way around, and your reply to oldpep still doesn’t seem to give the details to justify your statement.
Without getting into different roles, it seems like a team of nine players with a .400 OBP and a .400 SLG is going to score fewer runs than a team of nine players with a .300 OBP and a .500 SLG. Although both are .800 OPS, the team with the lower SLG will strand a lot more runners. That’s not a scientific analysis, so maybe I’m missing something. Is it that SLG is more “valuable,” but that the scales are different and a .020 increase in OBP represents a much bigger jump than a .020 increase in SLG?
[Reply]
Kevin M. Reply:
June 29th, 2011 at 3:35 pm
You’re completely wrong here. A team full of 9 players with .400 OBP would probably set a single season record for runs scored….even with that .400 slugging percentage.
Would they strand more runner? Of course they would because they’d put a boatload more runners on base than a team with a .300 OBP.
It’s also very bad math to compare OBP and SLG because they have different denominators. OBP maxes out at 1.000 and slugging goes up to 4.000. It’s why OPS is such an ineffective stat.
[Reply]
Mike Jaggers-Radolf Reply:
June 29th, 2011 at 3:58 pm
I’ll try to address both TedK and Kevin M. here.
Ted, OBP is far, far more important that SLG. One of the main criticisms of OPS+ as a statistic is that it does not favor OBP ENOUGH in its final calculation.
I’ve seen people argue that OBP is about twice as important as SLG. The entire premise of the book Moneyball, in fact, was the fact that the Oakland A’s were one of the first teams in baseball to recognize that OBP was the most important stat when it came to scoring runs.
With regard to Gardner and Swisher, let’s look at their 2010 totals. Both players had excellent years. Gardner posted a .383 OBP and a .358 wOBA while Swisher posted a .359 OBP and a .377 wOBA — all good numbers all around. Swisher has the higher wOBA, however, despite the lower OBP, precisely because he hit for so much more power than Gardner. Gardner’s better on base percentage couldn’t make up for his lower than average SLG in total. In general, if Swisher doesn’t start a season miserably, as he did in 2011, he’ll be a more valuable offensive player in a vacuum than Gardner is because he gets on base comparably well but hits for more power.
Kevin M., spot on.
[Reply]
I would not try to determine which player (Gardner & Swisher) is the better offensive player because while they have some overlap (getting on base), they serve different functions in the game. Gardner’s value is getting on base and in position to score runs. Swisher’s value is more for driving in runs. That being said, I did enjoy the article.
[Reply]
Mike Jaggers-Radolf Reply:
June 29th, 2011 at 2:51 pm
Thanks Phil C. I appreciate that. I was not sure if our readers would enjoy this post or not.
Regarding Gardner and Swisher comparatively, your comment is precisely why a good analyst would present both AVG/OBP/SLG and wOBA. A team composed entirely of Gardner or Swisher clones would have problems. Instead, a team would be composed of a mix of players with strengths that fill different needs up and down the lineup, and at each position.
The complete stat line, with all four metrics, would demonstrate which player can do what (get on base, hit for power) through the slash line, while wOBA would tell us which player in a vacuum was comparatively better, which would be valuable for assessing two players with similar skill sets.
[Reply]
My problem with stats is they only tell of past performance and aren’t 100% indicative of future results. That sounds like a commercial, haha.
Jeter is a perfect example. Who the hell would’ve imagined he’d have posted .270 BA and .340 OBP last year and worse this year after his stellar 2009?
Last night’s game is another example. Grienke bombs against a team who at times has made #4 and #5 starters look like aces, #2s, and #3s.
Baseball is like life: random. Stats tell the general tale and that’s really it.
An interesting experiment for this blog or any baseball blog would be go a week without mentioning stats in any article.
[Reply]
Mike Jaggers-Radolf Reply:
June 29th, 2011 at 3:14 pm
Your criticism goes beyond just baseball statistics. You are criticizing the entire discipline of statistics in all its applications.
For example, one of the main uses of statistics is the linear regression, which is meant to help an analyst make future predictions based on past data, or make unobserved predictions based on observed data.
Linear regression models, however, are never 100% accurate. I’ve developed them as part of my profession, and have encountered instances where a model that was only 30% accurate was deemed a success. You show me a statistician who is 100% accurate and I’ll show you either a fraud or a very wealthy man.
I’m not sure our site would be quite the same if we declined to mention statistics for a day, never mind a week.
[Reply]
Kevin M. Reply:
June 29th, 2011 at 3:31 pm
Actually, quite a few people predicted Jeter’s demise based on the historical comps based for 36 and 37 y.o. shortstop’s. So yes, statistics can be effectively used to predict future events.
[Reply]
oldpep Reply:
June 29th, 2011 at 3:40 pm
amen-ditto Jorge
[Reply]
Phil C Reply:
June 29th, 2011 at 3:35 pm
Well Duh, then you’ll like my favorite statistic. Whenever Romero Rena comes to bat, he has a 50% chance of hitting a home run. Because either he will or he won’t.
[Reply]
Mike Jaggers-Radolf Reply:
June 29th, 2011 at 4:00 pm
That assumes that Pena is just as likely to hit a home run as he is to do anything else in a given plate appearance.
[Reply]
Kevin M. Reply:
June 29th, 2011 at 4:49 pm
Phil….I’ll bet you $10,000 Ramiro Pena does not hit a home run in his next at bat?
Deal?
[Reply]
[...] One more look at wOBA and AVG/OBP/SLG: Another follow-up on why BA, OBP, and SLG complement wOBA, this time from Mike Jaggers-Radolf of The Yankee Analysts. [...]
I just got the biggest boner of my life. Thanks Mike.
[Reply]