Jon “Boog” Sciambi never found himself the subject of a scathing critique on the “Fire Joe Morgan” blog, and for good reason. Few, if any, broadcasters have been as saber-friendly as the former Marlins and Braves, and current ESPN, play-by-play man. Sciambi was in attendance at the MIT Sloan Sports Analytics Conference earlier this month, and between sessions he discussed the use of advanced stats in the broadcast booth.
Sciambi, on giving the audience what they need: “Some broadcasters simply aren’t interested in advanced stats and there are others who are worried that the audience will have their eyes glaze over if they use them. But if you are interested, by and large you’re going to use them. As I’ve said before, it’s kind of like the New York Times and the New York Post. The Post puts on their front page what they feel the readers think is most relevant. The Times puts above the fold, to the right, what they think is most relevant. They print the news. Kim Kardashian is going on the front page of the Post, because it will help sell papers. What happened in Syria is going to go above the fold in the Times. Similarly, I want to deliver ‘the news’ and tell people what is actually happening. Advanced stats help illuminate that.
“I’m a big believer in perception. The mainstream, by and large, dislikes or devalues Three-True-Outcome guys. And they love the counter. Even when Adam Dunn was productive, a lot of people didn’t like him because it looked like there was nothing happening. He was either trotting around the bases, walking to first, or walking back to the dugout. Conversely, Juan Pierre was all movement. He was hitting ground balls and running fast. It looked like a lot of stuff was happening. Of course, the main thing that was happening was he was making a lot of outs.
“In a broadcast, I’m more than willing to say, ‘Everybody hammers this guy, but he’s actually very good,’ or ‘this guy has struggled and here’s the evidence.’ It’s a must to do that, to be honest. You’re giving out good information. There’s also a way to do it. If I’m talking about Yuniesky Betancourt, or some other low-OBP guy, I don’t need to give their OBP and then say that they’re really bad. I just need to give their OBP. The savvy fan will know. Sometimes I might say, ’His OBP is .250 and the league average is .320.’ I don’t need to put an equal sign at the end of it, followed by ‘stinks.‘ They’ll figure it out.
“This year, if Drew Stubbs walks more and raises his OBP to .345 — which he’s capable of — and hits for a little more power, he’ll be a very productive player in Cincinnati. But if he strikes out just as many times and doesn’t markedly increase his RBIs or home runs, I don’t think most fans will assess him correctly. Some will, but most will just see the strikeouts and not notice that he’s a better player than he was last season because he’s not making as many outs. If a player is adding statistical value beyond counting stats, it makes the broadcast better if I point that out.
“I’m not sure that advanced stats will ever become completely mainstream. One of the problems is that if I’m arguing with a player and saying ‘this guy is good,’ and he’s saying ‘no he’s not,’ I’m never going to win in the eyes of most fans. A guy who never played won’t win an argument against a player who is saying ‘This is what’s important: .300, 30 and 100.’ The fan is going to say, ‘He played, so he knows.’ That doesn’t mean I shouldn’t state my opinion and give stats that are meaningful.”
On using predictive stats in a broadcast: “I’m always cognizant of the stats I’m putting out there and whether they have predictive value. That’s the biggest thing. I try to stay away from runners in scoring position, and if I use it, I’ll try to explain, ‘This is what has happened in the past.’ I want to make it clear that I’m not giving the stat to say, ‘This will predict what will happen in this at bat.’
“There is so much information out there that you’ve got to be careful. I try to pick and choose, but I definitely have an idea. I know the stats that are more predictive in nature and the ones that can mislead. First and foremost, I don’t want to mislead. I think the most effective use for stats in what I do is to try to eliminate the noise.”
On sacrifice bunts and run probability: “I have Tom Tango’s bunt tables and I will use them from time to time. Most of the time we speak about total amount of runs — the potential amount of runs that could be scored — and I think we more often need to include, especially in late/close games, ‘What is the chance that you’ll just score?’ By and large, when we use bunts tables in statistical evaluation, the first place we go to is that spot with base-state and out-state. You’re expected to score 0.42 runs and in this state it’s 0.65 runs. I think both tables should be included when you’re analyzing bunts. What is the total amount of runs you’re expected to score, and what is the percentage chance that you’re expected to score?
“The fans want to hear stats, but at a very basic level. I feel you have to be as efficient as possible when explaining them. You have to make them understand. If you can do it efficiently, they absolutely want to hear them. If you put it as simply as, ‘If you bunt in this spot, you’re expected to score this percent of the time/this amount of runs and if you don’t bunt in this spot, so on and so forth.
“As far as strategy, if you describe the concept of giving up outs — if you’re taking a guy with a .370 OBP, who does a good job of not making outs and you keep bunting him — the fans should be able to understand that it’s probably not a good play. I don’t need to say outright, ‘The manager is making bad decisions.’ You don’t need to beat anybody over the head with it, you just give the information. As a broadcaster, that’s part of my job.”