Welcome to the first edition of Stats All, Folks of 2019.
For those who don’t remember, or have blocked it out, this is my ongoing series in which I discuss the use of sabermetrics and analytics in baseball. In previous segments, I’ve discussed Win Probability Added, spin rate, and Catch Probability, among many others.
Normally in this article, I focus on a particular sabermetric statistic and explain why it’s useful. Today, I’m going to focus on an extremely old-fashioned statistic and explain why it’s not useful.
I refer, of course, to batting average, a stat that is nearly meaningless in modern baseball.
Wait! Put down the torches and pitchforks. Hear me out.
I appreciate that batting average has been part of baseball’s statistical lexicon for as long as the game has existed. A version of batting average dates back as far as the 1860s, according to John Thorn, Major League Baseball’s official historian. It’s one of the three rungs, along with home runs and RBIs, of the offensive triple crown. For decades, many teams — and fans — used batting average as the go-to stat for measuring the productivity of a hitter.
But times have changed. And with plenty of more advanced statistics and data now at our fingertips, batting average simply doesn’t carry its weight anymore. It gives only a limited amount of information about how well a hitter is actually performing.
Don’t take my word for it, though. Just ask Orioles assistant general manager and analytics guru Sig Mejdal, who discussed the drawbacks of batting average in a March interview with BaltimoreBaseball.com.
“No baseball fan thinks a home run is as valuable as a single, but the batting average does,” Mejdal said. “No fan thinks that a walk is of no value, but the batting average does.”
Mejdal succinctly pinpointed the fatal flaws of the stat. For one, batting average doesn’t make any distinction between different types of hits. A batter who goes 1-for-4 with a booming home run is treated as the exact equal of a batter who goes 1-for-4 with a weak infield hit. The former was clearly a bigger contributor to the offense, but batting average would mark them both as .250 hitters, no more, no less.
Batting average also doesn’t measure the most important skill for a hitter: the ability to avoid making outs. By leaving walks out of the equation, it’s excluding essential information about how often a hitter gets on base. Let’s say Batter A, in 10 plate appearances, has two hits in eight at-bats, with two walks. Batter B, in 10 plate appearances, has three hits in 10 at-bats, and no walks. If you looked strictly at batting average, Batter B would have the edge (.300 to .250). But Batter A actually got on base more times than Batter B did, and created fewer outs.
Let’s look at an Orioles-based example. If you looked strictly at batting average, you’d think the club’s most productive offensive player this year was the .313-hitting Hanser Alberto. With all due respect to Alberto, who has had a breakout season, most followers of the Orioles would agree that the club’s best hitter has been Trey Mancini. Alberto may have 34 points of batting average on Mancini, but that’s far outweighed by Mancini’s superior OBP (.338 to .331) and especially power (24 homers and .531 SLG compared to Alberto’s six and .409).
Alberto currently ranks fifth in the AL in batting average, ahead of such stars as Mike Trout, J.D. Martinez and Mookie Betts. Does anybody think Alberto is a more productive hitter than those guys because of his batting average? They shouldn’t. It’s no surprise that Mejdal and the Orioles’ administration put very little stock into batting average, and the same is true for every analytics-minded front office in baseball.
So, what are some offensive stats that are more useful than batting average? Well, it’s a long list. Folks who are more sabermetrically inclined tend to favor stats such as wOBA (weighted on-base average) and wRC+ (weighted runs created plus), which I’ve written about in previous installments of Stats All, Folks. In brief, these stats aim to measure a hitter’s true value based on how often he gets on base and the ways in which he does so.
Teams also have their own proprietary methods for measuring offensive performance, and the Orioles are one of them.
“Once you have a model that tries to make sense of the ‘counting stats,’ wOBA, FIP (Fielding Independent Pitching), whatever it is, is really some combination of hits, strikeouts, walks, etc.,” Mejdal said. “So it’s a different person’s attempt at really putting correct value or weight on these components. We have our own, and it’s not too different.”
Granted, those kinds of stats might not be easily digestible for many fans, and you won’t often see them on most baseball broadcasts. So those who are uneasy wading too far into the sabermetric depths can turn to a much more commonplace stat: OPS.
OPS, as most baseball fans know, is on-base percentage plus slugging percentage. It incorporates both a hitter’s ability to get on base and to hit for power, giving a more accurate measure of a hitter’s production than batting average does.
“We know the weaknesses in (batting average), and OPS is one, for instance, that makes up for it,” Mejdal said. “From an offensive standpoint, OPS is wonderful. It correlates very well to the actual run value associated with the single, double, triple.”
OPS has its flaws, too. It weighs on-base percentage and slugging equally, when in fact OBP is almost twice as important as SLG in scoring runs. Still, it’s a much more illuminating stat to use than batting average, and it’s easily accessible. Many baseball broadcasts now include OPS alongside traditional stats like batting average, home runs and RBIs.
Checking back in on our Orioles example, OPS does a better job of delineating which hitters have been the most valuable. Among qualified Orioles, Mancini leads the club with an .869 OPS. Alberto’s .740 mark ranks fifth, trailing Mancini, Renato Nunez (.838), Pedro Severino (.820) and Jonathan Villar (.747). This reflects that while Alberto has been getting his singles, his overall production is dampened by his low totals of walks (eight) and extra-base hits (18).
With all that said, don’t expect batting average to go away anytime soon. At this point, the stat is entrenched into the fabric of the sport. I still cite batting average in my writing, and I think the same is true for the majority of the baseball media. Warts and all, it’s a tradition. There’s no need to airbrush the batting average column out of your old baseball cards.
That doesn’t make it a meaningful stat, though. Using batting average to describe a hitter’s talent is like introducing yourself to someone by naming your favorite pizza topping. Technically you’ve given them a sliver of information about yourself, but not enough for the person to really understand who you are.
It may not be time to retire the batting average statistic, but it’s time to accept its severe limitations. Saying someone is “a .240 hitter” is telling only a small fraction of the story, and doesn’t come close to explaining how productive the batter actually is.
At the very least, try OPS instead. It’s Sig Mejdal-approved.
BALTIMORE—For the second time in three games, closer Craig Kimbrel blew a save chance for…
Orioles left-hander John Means allowed just one hit in seven scoreless innings in his sixth…
Matt Wieters was one of the best catchers in Orioles' history. The fifth overall pick…
BALTIMORE—What happened? Cole Irvin pitched seven scoreless innings, Gunnar Henderson, Adley Rutschman and Ryan Mountcastle…
A day after he was optioned to Triple-A Norfolk, Jackson Holliday was hitless in four…
The plaque on the back wall of the press box at Oriole Park said it…
View Comments
This is spot on. I'll know the tide has turned on BA when the screen shows OBP, SLG, and Extra Base hits rather than the Avg, HRs, and RBI they flash up now. Whenever I watch a broadcast, I always look forward to the second AB when they do avg and OBP rather that RBI. RBI are a non-stat to some extent since it always depends on the batters in front of you. A slow runner may not score from second on a single while a fast runner might score from first on a double (and I have seen instances when they scored from first on a single without any errors).
If OBP is worth twice as much as SLG, why isn't OPS figured as OBP being worth 2/3 of the total and SLG. 1/3? Just sayin'
I coached both baseball and softball at the high school level and incorporated OBP, SLG and OPS when making my batting order. I never found a single player or parent who actually understood what I was doing,even though it is isnt' rocket science.
Agreed, Borg, there are other "traditional" stats that don't carry as much meaning nowadays, including RBIs and pitcher wins. That might be an article for another day. I'm with you -- I'd like to see all baseball broadcasts show more meaningful stats (OBP/SLG/OPS, etc.), though some already do.
Totally cool article, curious if we should have a Platinum Player instead of Silver Slugger now? Lol...go O’s...
This is much ado about nothing. Of course batting average is not the end all stat for hitting. Hence the use of so many other offensive statistics. But to minimize it to such a huge extent is pretty ridiculous.
Tru dat CB.
The problem is that many broadcasts/analysts/fans still treat batting average as if it IS the end-all stat for hitting. You still hear plenty of examples of broadcasters saying things like, "He's batting .250 in such-and-such situations" without providing any other stats or further context. That's where I think we're missing out on important information.
I wonder if ballparks weren't bandbox sized, baseballs weren't juiced to the point of being confused with golf balls, if player training and 'nutrition' weren't what it is ... and most significantly of all ... if the strike zone were simply called as it's written into the rule book ... if this argument would even exist.
No doubt that the long ball rules today's game, and I guess that dictates that Slugging % rules as well, but it's a shame. Seeming every player in the game can go yard at least 20 times a year now. Frankly, Sports Center highlights seeming show nothing but. BORING. The point is that the HR is being devalued to the point that the dunk in basketball has become. When's the last time anybody paid any attention to the NBA dunk contest? My prediction is that the MLB Home Run competition will go the same way. Hell, I haven't watched it in years.
So yes, I guess that the changes int the game have antiquated the old outdated statistic of batting average. But that's a pity. Mr. Manfred needs to realize that Home Runs don't necessarily make the game more exciting. Frankly, I think it's more fun to watch players RUN around the bases than TROT.
Give me Gwynn, Carew, Brett and Ichiro ... not Mark McGwire or Sammy Sosa. And for gods sake ... start calling the HIGH STRIKE!!
Great post there BB Robinson!
One more thought regarding batting average and how we should "accept its severe limitations" as the writer suggests ....
Perhaps if the Angelos family had paid a little more attention to that silly statistic ... the Crush Davis contract may never have become the albatross that it is today.
I'm not so sure that argument holds water-while Davis' salary is overboard, the idea of signing him to a long-term contract did have some merit based on his production for the Os. In 2013 (3 years before the signing) he hit .286-and that is SO hard to believe now!!!- with an OBP of .370 and 53 HRs. The next year his average dropped way off to .196, familiar territory now, but his OBP was still a respectable enough .300 considering how bad his average was. He still hit 26 HRs. Here's where I think the analysis went south-in 2015, his last year before free agency, he went back up to .262 with an OBP of .366 and 47 HR.
I guess they thought the middle year was the outlier rather than a taste of things to come. The first year of the contract (2016) he dipped back to a .221 average, but his OBP stayed decent enough at .332 and he still hit 38 HRs. Of course, we all know what happened after that, but I think his numbers as of 2015 warranted an extension from the Os, though not at the $161 million mark or the seven year marks. Either figure being much lower( 4 years at $65 million for instance) would have made that contract more palatable. Nobody else was even bidding on him at the time.
I think your analysis of the Crush Davis contract extension is right on point ... an extension for Davis was certainly a reasonable move at the time, in 2015 - the problem was that Peter apparently got bamboozled by Scott Boras into bidding against himself, and agreeing to a ridiculously long and expensive contract.
Well if the .196 year didn't scream "DON'T GET A 2ND MORTGAGE" on this guy ... Well ... baseball isn't rocket science. And hits/at bats still tells an awful lot about the player.
With Boras, there is always a mystery team. In this case, I know the Cardinals had interest. Boras went to ownership and convinced PA that the contract reflected a hometown discount because CD wanted to stay with O's. I think the Davis contract will be the "poster child" for how things have to change in MLB. It is being felt even now, as the last two off seasons have seen veterans not getting the offers that they believe are below the market. The FA market and the recruiting process will get redefined.
I think it was Bud Norris after he was traded to the Os who denounced the Houston Astros hierarchy for treating players as just stat items instead of caring about people. One thing stat heads don't seem to value is chemistry because they can't quantify it though it is essential. My issue with all the alphabet formulas especially park adjusted this and probability that and weighted something else is that a ten year old can't just do it with pencil and paper or calculator anymore. For instance I thought Roberto Alomar was a great fielder though someone who never saw him play says the stats reveal he was one of the worst. I can go along with you on OBP, which the Os haven't done well for years, but the drive to turn what we watch and enjoy is being turned into a video game and we don't need season tickets to watch that.
I had never looked up Alomar’s defensive metrics until now. His defensive WAR was 3.3, not great but not horrible. His offensive WAR was 70.8, which is great and his OPS .814, which was good.
FWIW, I just did a quick review of the 2018 American League statistics for team BA, OBP, and OPS, and how they correlated to team win/loss records.
All three stats, BA, OBP, and OPS, actually correlated very closely to the team's winning percentage. Team OPS correlated slightly more closely than team BA, and team BA correlated slightly more closely than team OBP.
I think a better analysis would be to compare those 3 stats with total runs scored. Pitching and defense have too much to do with wins and losses to be ignored.
The simplicity of batting average is the whole point. The casual fan, upon whose eyeballs and dollars baseball lives and dies, does not want to do a math problem to figure out if Robbie Alomar was a good second baseman. I myself do see the value of analytics and appreciate it's deeper insight into the field of play. That being said, MLB does itself no favors by allowing it to dominate the baseball conversation. It makes the product inaccessible, ultimately damaging it's broad appeal.
Look at the soaring success of the National Football League in this town. Ed Reed is about to be inducted into Canton, hailed as arguably the greatest safety in the history of the game. When is the last time you read an article with a bunch of spreadsheet attachments laying out complex arguments for or against him? You dont. People know a good football player when they see one.
Interesting point, Bancells. WAR may be like the quarterback rating.
I agree, Bancells, that a big reason batting average has stuck around so long is because it's so simple. But the fact that it's so simple limits the amount of information it can really convey.
OK, tell me your favorite pizza toppings are pineapple and tofu, and I know all I need or want to know about you.
Seriously, we're using Sig as a critic of statistical formats? That's like asking Frank Perdue about the merits of chicken. It's what he sells. What's nice about a batting average is that everybody knows what it means, warts and all. I confess I've made no effort to comprehend the new, computer-generated numbers, but I have the sense that they are based on a lot of subjective criteria and different statheads like different criteria. With a batting average, about the only subjective data is the official scorer's ruling of a hit or error. These new measures seem to include things like plays that should have been made (beyond official errors) or batted balls that should have been hits according to someone's or some machine's belief about exit velocity and launch angle. and whatever. I'm sure that you can concoct a math formula that will show that Alberto is not only worse than Mancini but also way behind Davis and Martin. No doubt people really believe in these new measures, but what are the odds that the correlations are merely a matter of chance?
Agreed, Boog RR.
The dimensions of ballparks used to be enormous. A home run, unless down the line to the foul pole, was truly a wallop.
In 1955, Memorial Stadium's power alleys were 447 feet and straightaway center was 450 feet. In 1958, they were reduced to 380 and 410 feet.
While I definitely agree with the idea that OBP and OPS tell a much better story about a players production than AVG does, I see one main downfall in this analytical approach: With the mantra that focuses on getting on base, or getting extra base hits comes an actual devaluation of just getting a base hit every now and then. I think Chris Davis is actually a good example of this. When his average started plummeting people pointed to his OBP and slugging to show that he was actually more valuable than his AVG alone would show. The same thing with Bryce Harper after his MVP season: his AVG fell from .330 to .243 but his apologists pointed to the fact he still walked more than 100 times. While this analytical approach is almost entirely correct it keeps telling ballplayers that choking up and "hitting one where they ain't" isn't of value anymore. Unless where they ain't is the second deck. Batting average may be a flawed stat, but it used to be something a ballplayer had to think about. Like if he was scuffling at .220 he might try something different. Now he can just try and get walked or hit a dinger and say I'm hitting .220 but my OPS is over .800 so I don't have to try harder.
So, is hitting .300 the new .400? Go O’s...