A recent column about TV networks feeding viewers junk statistics led to a bunch of letters on the subject, which in turn led to a conversation I had with a reader who works in television. That conversation provided a big insight into how things work in sports television, and with the reader's permission, I'd like to share it with you.
The original column criticized ESPN for providing viewers the context-free stat that the Anaheim Angels scored 40 percent of their runs with two outs. The graphic was flashed after a two-out rally against the Chicago White Sox, and was clearly meant to show that the Angels were good at scoring runs with two outs. On seeing the graphic, commentator Buck Martinez launched into a little riff about how the Angels are scrappy and never give up on an inning.
On doing a little research, I found that the Angels are only a little better than average at two-out scoring, fifth in the 14-team American League, slightly better than their opponents that day, the White Sox. I also found that the leader in this category was the awful Kansas City Royals, and the division-leading New York Yankees and Minnesota Twins trailed the league. Regardless of whether the Angels were any good at it, if you were looking for a stat that said nothing, two-out scoring might have been it.
Reader Dave Perez commented, "They bank on the notion that nobody will take the time to look it up like you did. Just like politicians' ads."
That prompted my TV-industry reader, who will remain anonymous, along with the reader's employer, to write.
"I've been in these production trucks -- those statistics, while indeed context-free, are thought of by people under pressure at the moment, not as a concerted strategy to mislead the public. The pressure they face is to get a stat on the air at the appropriate time of the game, not to place it in a larger context.
"This phenomenon is less forgivable on scripted shows (like a highlight show), but during a live game, announcers need something to talk about, so they ask for and receive statistics of debatable merit."
I responded that while I don't think reader Perez meant that it's "a concerted strategy to mislead the public," I do think that the networks "bank on the notion that nobody will take the time to look it up," or even think too deeply about it. That is the same as the politicians' ads, and the same as many other kinds of ads too, though the motives differ in each case.
I don't buy the explanation that the guys in the truck are under time pressure and have to get something, anything on the air. If the announcers can't find "something to talk about" without getting bogus stats from the truck, it's time to get some new announcers.
The way to have things to talk about is by doing homework before the broadcast starts, by being prepared. That's also the way to have relevant, context-driven stats at the ready. It's one thing if something surprising happens -- if six guys in a row get hit by a pitch, I can forgive some obviously scrambled-for stats on consecutive hit-batsmen. Even in that case, though, I would expect to hear something along the lines of "We're scrambling here, folks, but this is what we've come up with so far."
But if the Angels' two-out run scoring, for example, is notable at all, was it inconceivable before the game that the Angels might score some runs with two outs? Shouldn't the truck have had some stats handy? And if a stat is worth talking about, it's worth putting in context. Otherwise it's just nonsense. Time pressure is a bogus excuse. A stat not in a larger context might as well be wrong, because it has no meaning.
My TV-industry reader wrote back:
"It's my failure to be clear here. Pretend you are a graphics producer for a baseball broadcast. The network's lead announcer mentions on the air that the Angels are particularly proficient at scoring two-out runs. Your job is to make this announcer look as good as possible. You immediately look up how many runs the Angels have scored in that situation and get it on the air, not to get in the announcer's ear and patiently explain that this stat has no context.
"If you're a rookie producer, say, working a baseball game, and the network's lead announcer -- the top on-air guy -- says something on the air that you don't agree with, is it likely that you'd present your theory to the audience at the risk of offending him? Or would you put up the statistic that appears to mean something and be satisfied that at least 80 percent of your viewing public will be satisfied with it?"
Here was my epiphany."Your job is to make this announcer look as good as possible." There's the problem. And here's a bigger one: "You put up the statistic that appears to mean something and be satisfied that at least 80 percent of your viewing public will be satisfied with it." Holy cow.
My job, as the graphics producer or anything else, should be to to serve the viewer by producing the best game telecast possible. My contribution should thus be to present statistics that are relevant and informative, that help the viewer understand and appreciate what's going on. It's not enough for 80 percent of the viewing public to think -- incorrectly -- that what I put up means something.
Because the problem isn't that our example stat had no context, it's that the lack of context made the statement being made a false one. The statement being made was, "The Angels are very good at scoring runs with two outs." The stat appeared to illustrate that point. But if the stat were given in context -- "The Angels score 40 percent of their runs with two outs, which is fifth best in the league" -- it would actually disprove the statement. The Angels aren't very good at scoring runs with two outs. They're only decent at it, never mind whether the stat means anything in the first place.
So it's not like this context thing we're talking about is some esoteric, only for experts kind of thing. We're talking about the difference between true and false, between accurate and inaccurate. And the TV networks, or at least the one my reader works for, are willing to ignore this fundamental difference, the very crux of the difference between good and bad information, to make the announcer look good.
If the announcer says the Angels are very good at scoring runs with two outs, my job as the graphics producer should be to find him a stat that confirms that, sure. But if, in searching for that stat, I discover that the Angels are actually not especially good at scoring runs with two outs, then my job at that point should be to tell that to the announcer. I would then expect the announcer to say something like, "I said the Angels are good at two-out scoring, but the guys in the truck looked it up and it turns out the Angels are just middle of the pack. They're fifth in the league. They're nothing special."
This could lead to an interesting conversation about how looks can be deceiving, or perceptions can be different from reality. Or the subject could be dropped. In fact, the announcer could even choose not to correct himself, since an off-the-top, spoken comment that's inaccurate is more forgivable than an inaccurate graphic, which the viewer can only assume has been vetted.
Instead, what's going on under the regime my reader describes is that whatever the announcer says becomes this alternate reality. It's true because Joe Buck or Al Michaels or Jim Nantz or Jon Miller or Bob Costas says it's true. Let's get some numbers up there! If the graphics producer's job, anyone's job, is to make an announcer look good at the expense of true, relevant, contextually accurate information, then why should I ever believe anything that network has to say, about any subject?
An analogy to that high-profile announcer, the reader writes, "is you: You have your name on your column on Salon, but I'm sure there are people who work just as diligently as you do who don't have your profile. If these people really disagreed with you on a point, would they have the gumption to do so publicly, making you look uninformed?"
The answer is yes.
Even though I have my name and my caricature on the column, and I have a higher profile (though not higher status within the company) than the people working as editors, and there might be the odd person who'd cancel his subscription if I left, and I don't know that anyone's ever canceled a subscription because an editor left, anywhere: The answer is yes. If the most junior of editors who might edit my column found something in there that they disagreed with, they would and should challenge it.
It might come down to a judgment call, and I'd like to think I'd have a decent chance of winning that argument, but if I say the Angels are a great two-out scoring team, to pluck an example out of the air, and an intern who's been given a shot at editing me as part of her ongoing education says, "Wait a minute, they're only fifth in the league," that word "great" is going to be changed to something like "decent," even if I throw a fit and stomp my little foot.
Because otherwise where is the line drawn? What if the lead announcer says something that's just plain wrong? What if instead of saying the Angels were great at two-out scoring, and the Angels are fifth in the league, what if he said they're great when they're actually the worst team in baseball history at scoring runs with two outs? Does that graphics producer still not get in his ear?
I realize I shouldn't have been surprised or outraged at what my reader told me about how TV sports broadcasting works, and I do want to point out that the reader was only talking about one network, though we agree that none of the networks that broadcast sports are particularly good or bad on this issue relative to the others.
I guess I already knew that it's a major priority in television to make the talent look good. What I hadn't realized was that the truth is not too big a thing to sacrifice in that pursuit. So I've had to amend my list of who is being served by the networks that broadcast sports in what order. It now goes like this:
1. The on-air talent
2. Non-sports fans
3. Sports fans
I want to give the last word to my reader, who was only explaining, not apologizing for, this behavior: "I agree with your original column. It's best for the home viewer to be skeptical. [If a network puts up a graphic that features a statistic], take it to the bank. As for what that statistic means or doesn't mean, it's best for voices like yours to call them into question the next day.
"I think we have common ground here. We agree: A) a context-free statistic has little if any meaning, and B) there is no effort to mislead the sports viewing public for the most part."
Previous column: Readers write: Clemens, Ali, Shaq
- - - - - - - - - - - -