Jon Sciambi, who does play-by-play of Major League Baseball games for ESPN, has been kind enough to grace the Baseball Prospectus site with a guest column.
Let me tell you about an argument I had with Chipper Jones. Last year, I came across an interesting nugget on Fangraphs while doing pre-game prep: Besides Albert Pujols, Chipper sees the fewest first-pitch strikes in the majors.
Chipper is open-minded when discussing hitting, even when he disagrees, so I decided to present him this information prior to the game. He was really surprised. He didn't believe the facts, even though the numbers were inarguable. Or, more to the point, he believed in what he knew (and himself) more than my stupid, never-played-the-game facts.
Chipper was so surprised that he went around the clubhouse asking teammates, one by one, if they were surprised. None of them were. Everyone saw it but him, the guy with ostensibly the best view. Chipper has great eyes, obviously, and great belief in those eyes, but those eyes can also occasionally lie to even one of the best hitters in the game.
I went on to ask why he'd swing at so many first pitches when the numbers suggest it's not a great play. Chipper explained that the first pitch is often the only time he'll get a "heater" the entire at-bat. "OK," I say, "but clearly, mathematically, factually, you're not getting a ton of strikes." We go round and round for a bit without concession on either side and eventually I go upstairs to broadcast the game.
Fast forward to the top of the first. San Diego's Tim Stauffer on the mound. Chipper digs inï¿½ and takes a 91-mph fastball right down the middle. He steps out of the box, finds our broadcast booth with those great eyes and, well, here's what follows:
Chipper Jones 1, Stats 0. But Chipper, sample size!
Will Carroll's "Be Stupider" post got me thinking: How can the broadcast world do a better job of delivering sabermetrics to the masses? And, more importantly, how do we make it "illuminating" instead of "coma-inducing?" Additionally, how do I get Chipper to understand I'm not a moron?
Like Will, I'm "mathematically challenged" and find lots of BP, Fangraphs, Tango, etc., tough to digest. I am somewhat "tentative" to push SABR-type stuff on air because of the aforementioned coma potential. That's not an excuse, just reality. But it's also on me to do better. During my years doing Braves games, I had the good fortune of working with a crew that was open to new ideas. All our player stat lines included OBP (not earth-shattering) and the bulk of our graphics were presented in rate-stat form. But the thing that made it work was my broadcast partner, Joe Simpson. If I dug up a non-mainstream stat somewhere online, I'd say, "You know, Joe, I was looking at 'I live in my mom's basement.com,' and I found this note on Javy Vazquezï¿½" He would giggle and poke fun, but he would always listen. You know what else? He would change his mind sometimes. Joe is very old school. But he's fun enough to spitball in one of the classrooms of that old school. What's the line from Mr. Baseball? "I bet you didn't start playing baseball as a kid because you wanted to work?" That's right, Magnum P.I., I just quoted Mr. Baseball.
So how do we make a better broadcast? It's gotta be fun. Entertaining. To us. Yes. But, by extension, to the audience.
What else? Well, it's as much about eliminating certain stats from the lexicon as it is about adding them. We need to slowly erase the "noise." Stop mentioning and graphically supporting with stats like RBI and wins because that stuff doesn't make people smarter.
Let's not forget "it's the search for objective knowledge about baseball." The goal is not unveiling newfangled stats; it's about getting people to understand basic ideas and concepts. To achieve that, we can't just slap stats up on the screen and explain them. Understanding has to come in the form of analysis. We have to use the stat and explain it. Sometimes it needs to be the PBP guy playing analyst and getting the color guy to react:
If Ryan Howard is up, I can talk about RBI and why dependent stats don't evaluate individual performance well; RBI aren't what reflects Howard's greatness, his SLG does. I can mention that Howard's massive RBI totals may be due to the fact that no player has hit with more total men on base than Howard since 1492 (I believe this is a fact but didn't feel like looking it up). Point is, there are dead people who could knock in 80 runs hitting fourth in that Phillies lineup. (OK, I probably wouldn't say that on-air.)
The metrics are getting so advanced that we're in danger of getting further away from the masses instead of closer. We, as broadcasters, have to find better and entertaining ways of explaining the math in bite-sized terms. Simplified, we need to explain that one of the problems with batting average, as opposed to slugging percentage, is that batting average values a single and a home run equally. We can't assume that's understood just because we understand it. And the only way it gets embedded is to keep beating the audience with it so that it becomes ingrained the way ERA eventually did, even though that once passed for advanced math. That, and we should all wear blue blazers with an emblem that reads, "OBP is life."
I also believe, as it relates to the masses, the PBP guys can't move the analysis needle much. The masses will always find former players more credible, period, and the BP base needs to be more open to that-if the goal is indeed to inform the masses and not be "right."
The average fan wants the analysis from McCarver, not Buck. Joe Morgan may not be popular in these parts, but last year I worked a Tigers/Angels game with him in Anaheim. At one point, I mentioned UZR, UZR/150, and other defensive metrics. As we were going to commercial, Joe said that he'd like to talk about defensive metrics when we got back. What followed was a pretty straightforward, simple discussion about defensive metrics without disdain or dismissal on his part.
Look, there's still a ways to go. We need a former player to look into the camera and explain properly why this guy:
is a better offensive player than this guy:
We need to get to where the masses understand there is no choice. This isn't subjective. I evaluate offense with OBP and SLG while you like RBI and runs scored is not the same as "I like strawberry, and you like vanilla." It's "strawberry is better than vanilla." More accurate and, therefore, more delicious. To be clear, I don't speak for ESPN here, just me, but I think we have a responsibility to inform correctly. If a majority of teams are using advanced metrics to inform decisions, then we should do some of the same in analyzing those decisions.
In 2008, Dick Cheney was informed that two-thirds of Americans thought the war in Iraq was not worth fighting. Cheney's response? "So?" We need a little bit more of that 'tude. This isn't to say I agree with his politics. Or the war. Why am I even going here? Eject!
I'd liken this task to how The New York Times operates. The Times places the stories it deems most important above the fold and to the right. The majority of other papers go with what they think we think are the most important. The goal is to be the Times. If we eliminate the noise of RBI, runs, etc., keep it basic and utilize the slash stats, I believe that, slowly, the desert masses will drink the sand. The BP base must understand: VORP, EqA, WAR, and Robert Parish are not walking through that door. Not for a while. But it can only help if the broadcasters are a team, too-in uniformity (together, I mean, not wearing those blazers) while patiently holding that door open.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.Subscribe now
I've always thought that for batters runs created per game (RC/G) would convey a lot of information to viewers using a scale that can be easily understood, a scale only slightly different from the familiar ERA.
Maybe you can impart some of your knowledge on Chip before you go.
This all seems like a good way to start. "Now, this doesn't show up in the stats -- oh, wait, it actually does."
I think your RBI example is a great one and it makes me think of Bill James. There's a reason so many of us got in to Sabermetric thinking through James. James wouldn't lead with the stat. He'd lead with a question -- why does Ryan Howard drive in so many runs? The answer to that question would be a combination of opportunity and SLG%. Suddenly you're teaching without preaching.
And I think you're right about the role of the ex-players too. This has been made very clear recently by Harold Reynolds' deer-in-headlights look during those stat conversations on MLB.com. Can these guys be educated off-air? Can somebody give them a tutor? How can we expect them to contribute if they don't believe it to begin with?
I've had a many a conversation derailed by the "well I watched 5 games and didn't see what you are suggesting is actually happening, therefore I refuse to believe you..." argument. The fundamental acceptance of the rules for quantitative analysis is often the stumbling block.
A) Can be understood at a literal glance... and
B) Offer an inarguable truth that broadens the casual viewer's understanding.
Beyond that, a fine column, including the Rick Pitino quote at the end. More of this, please.
VORP and WARP will take care of that themselves as they are proprietary to BP. We'll see Fangraphs WAR far more often than VORP or WARP simply because it's free for all to use and access.
1. Growing up with Wade Boggs and Dwight Evans as lead-off hitters, I can hardly believe that it's taken so long for the country to accept OBP. I love it when broadcasts show the triple slash stats. Perhaps the first at-bat features the traditional stats when the player goes to the box, then the second at bat features the slash stats?
2. I'd like to be able to see the strike zone better. There's a number of broadcasts where it seems like the CF camera's at a 30 degree angle to the hitter, which makes it impossible to follow balls and strikes. If there's one presentational thing that I'd like to see more of, it's camera angles that show the strike zone from a more accurate angle. With that, you could then do graphics to show whether a pitcher "has" each of his pitches that day.
3. I think BP's base-running statistics, in their raw form, could be useful for broadcasts. By that I mean, Player X is fast but gets thrown out on the bases 3rd most in the league or other observations of that ilk. SBs are only one proxy of a player's speed in game.
4. Ballpark effects broken down by handedness would be nice, especially when they're in the news. Telling people that X Park has X effect on Left handed hitters' BA and HR percentages would be nice.
OPS is far more relatable once you understand that 750 is average and 1000 is good. Broadcasters should pound the spectrum into people's heads until it's second nature for them.
To me the most relatable fact about OPS is that Barry Bonds and Babe Ruth are the only ones to break 1300 for a season. People remember Bonds and watched him play, so they can relate to that. And Babe Ruth is a traditional figure, so when people see a 'new-fangled' stat attached to his name, they're forced to come to terms with the fact that OPS is just a way of talking about baseball.
Introducing sabermetrics during a game is not going to be easy because there are many preconceived ideas about the game. Plus, not everyone watching a game watches every game they can possibly squeeze into their day. Any new statistic will need to be quickly explainable and repeated in many games throughout a season for the general audience to understand it and accept it.
Adding data from the runs scored and winning percentage matrices will be useful. Letting the other analyst explain why a team might choose to do something that disagree with these would be good (i.e. an intentional walk to a slugger with significantly worse hitter on deck).
One thing you could show to get people to better understand run scoring (on a macro level) is the Top X run scoring teams of all time. Then show things that these teams have in common. I do not know which teams comprise this list, but I suspect that high OBP and high SLG% teams are most of them. Over the course of a season, show other things these teams have in common. Show what lower scoring team have in common. Another thing to consider is what winning teams have in common. Run scoring is not always a trait of winning teams, but they do manage to outscore their opponents more often than not. Explain why and how such a team could be even better if they employed a different offensive strategy.
Finally, if you know of anyone working in a MLB front office who has successfully introduced sabermetrics to the decision making process, you should talk with them to find out how they were able to change the mindset in the office. Then figure out a way to convince a larger audience in bits and pieces.
post(and including)-8/4/09: 37/166 = 0.223
And yes I am willfully ignoring the fact that Chipper went on a tear, hitting 16/27 immediately after you made your observation.
Listening to you and Joe during the season was one of life's pleasures. I met you briefly last year before a game near the back parking lot and was thrilled to find that you were just as great in person.
Just a few days before your announcement I was at the Metallica concert and I was talking to a friend of mine who works for the Braves. I told him I was nervous you would eventually leave for ESPN and how upset I would be. A few days later, I was sending him an anguished email.
Seriously, best of luck at the WWL - not that you need it. I can't wait to hear your voice again.
For introducing 'advanced' stats into tv broadcasts, I think you've got to do a couple things:
1) Baby steps.
OBP and SLG are the stepping stones to get people to this side of the pond. Time honored aphorisms can be useful, such as "You can't steal first base," "Leading off the inning with a runner (or walk) leads to more runs in the inning." "Hard to score runs if you don't have runners." Then you follow up with a compelling example of why this is so.
Naturally, this information leads to run-probability and win-probability. For instance, "Why is it bad to make the third out at third base?" or "Pitchers can sacrifice bunt because it's the most likely chance they have to advance a runner and get some small value from the otherwise harmless lumber in their hands, but Jeter should not ever ever ever ever ever sacrifice bunt, even in a playoff game."
I love the idea mentioned above about demonstrating the impact of defensive range with a graphic. Solid dots for balls fielded, and hollow dots for balls misplayed. Basketball has been doing this for a long time with shot location charts, so casual fans should be able to understand it pretty easily.
"the more ambitious the stat, the more complex and arbitrary it almost always becomes. What it gains in sophistication and the intuitive wisdom of its creator, it loses in simplicity and objectivity. How can you love a stat, or use it in arguments, if you canâ€™t really explain it?"
And the smart fan wants McCarver to retire from broadcasting and stop hurting us.
I recognise that a broadcaster's job is to be entertaining, but I don't see why those of us who are already enlightened to care.
So what if the majority of fans don't understand the game? Won't that just create more inefficiencies for us to exploit in Vegas or fantasy leagues? How do we benefit from teaching the masses how baseball works, and if we don't, why should we expend any effort at all doing it? Why should be accept dumbed-down analysis of the sort Will described in his "Be Stupid(er)" post if its goal serves no purpose that we value?
No. I want detailed baseball analysis. I want advanced metrics. I want to LEARN THINGS. I'm not here to teach others.
And if BP is going to spend a bunch of time and resources trying to appeal to a broader audience, that's resources that, from my point of view, are wasted.
Given that there's so much excellent free baseball analysis on the internet, hiding behind a pay wall should require better analysis. If instead they're aiming for more accessible analysis, then perhaps I'm not their target audience and I should instead donate the StatCorner.
It's kind of like convincing people the world is round. Do you want to be right, and no one believes you, which I would assume leads to a certain brand of madness, or do you want to figure out how to convince people to give your world-is-round "theory" a shot? You don't have to say the world is oblong, just say it's round and be smart about it. But bring along new believers. BP is, after all, a business and a brand, pay site or no. And businesses need to grow.
For example, if our audience were multiplied, we might not have to worry about our ability to retain a Joe Sheehan, which could make a significant portion of our audience happier. We might have a larger team working on PECOTA to get those projections to our readers faster, and hopefully with fewer known (and unknown) issues. We're able to do a lot of great stuff given our current resources, but there's a whole lot more we could if they were increased.
I like the idea that "We're telling you, the viewer, about this info because that's what the front office is using." For the better-run teams, pulling back the curtain a little could be a way to enhance the brand and build the community.
1. Stop equating batting average with hitting. Do not say anything along the lines of "He led the league in hitting last year" - if he led in batting average, just say he led in batting average. Don't infer anything beyond that.
2. Ignore anything based on small sample sizes. "Normally it's good to bring in a lefty to face a lefty, but this year Joe Blow is 'hitting' .333 against lefties" 2 for 6? Doesn't mean squat. Give the career totals, or last three years.
But I also wince when I hear PBP guys clearly trying to inject new, more sabr-friendly stats into the discussion. It can be really awkward. The best stats are nuanced (because baseball, and life, is). And frankly, during a game, I have no need to hear a 5-minute discussion on the caveats of a low BABIP affecting a guy's average that year (while noting park effects, etc).
I'd rather listen to a PBP guy and color analyst who really get along just call the game, talk about baseball, tell some stories, etc. A good broadcast is listening to a good conversation that doesn't get in the way of the game. Explaining the concept of replacement player would get in the way of the game, in my opinion.
Ken Funck pointed out a simple stat improvement for broadcast, and it requies no real new explanations in one of his final BP Idol entries last summer. A graphic for the leadoff hitter showing his OBP vs. the league average leadoff hitter. Bang.
One recent example from another sport was Fran Fraschilla bringing up "Kenpom.com keeps track of fouls drawn, and he is near the nation's leaders." Again, a simple stat that just isn't kept by the mainstream.
So the fact that Chipper is second lowest percentage in terms of first pitch strikes means he has the second highest percentage of having a 1-0 count after the first pitch. Doesn't this imply that Chipper is in fact very selective compared to other hitters on the first pitch.?
I also think any evaluation of whether Chipper is chasing too many first pitches has to look at the outcomes when he swings at the first pitch compared to the outcomes for other players and also the outcomes for Chipper when there is a different count.
Joe Morgan thinks like a former baseball player, not necessarily a baseball analyst. He's either not aware of the stats or he doesn't trust them. I wonder if Dustin Pedroia is aware of his UZR rating? Does the average current ballplayer follow the more esoteric numbers on a regular basis, or just around contract time? And speaking of contracts, I'm betting that Joe Morgan's contract with ESPN does not require that he incorporate "VORP" into his broadcast or that he provide x number of stats per game. He's there to enhance the broadcast as a former player, regardless of his title.
Given the large number of players who claim that they don't follow their own stats on a daily basis, never mind the more sophisticated ones, why would we expect Morgan, the former ballplayer, to be any different?
Maybe I can start an informal poll at Coors Field this year...I'll just ask visiting players as they warm up. "Mr. Ramirez? Manny? Do you know your WAR? Or VORP?" I kid but I am serious.
But if he wants to include in his analysis things that can or have been proven to be incorrect through quantitative analysis, he's no longer enhancing the broadcast. And, unfortunately for Joe, there is an increasing amount of territory, particularly in the world of valuation, where sabermetrics is more insightful than Joe.
It's when the ex-ballplayers dismiss out of hand the sabermetric approach and lessons simply because it does not align with their existing perspective, when they cannot recognize the limits of their knowledge, that the problems arise.
I agree that graphics with new-fangled acronyms are a non-starter. Given that so much of broadcast chatter deals with in-game strategy, that's where I see the big opening. Discuss the downside of the CS, the sac bunt and the "productive out." Discuss the fact that LOB's are actually a by product of run scoring and not an impediment to run scoring. Maybe debunk the hot hand theory with some evidence. There's tons of low hanging fruit to incorporate without getting too geeky and preachy.
Points (or what we call runs in baseball) is what it is all about. Score more than the other team- your team WINS!
How does the player(s) contribute to that ultimate team goal (ie more runs = wins) is the context OBP, SLG should be explained.
And the concept of collecting bases via bb (OBP), xbh, hr (SLG)is easy enough for any one to understand.
The walk/base hit gets you on base (1/4 to scoring), double gets you halfway to a run, twice as far as a bb, HR = 1 run, etc...
Player stats have to be explained in how they relate to the TEAM goals (ie scoring/winning). Explaining any stats without tying it into the team concept loses meaning, imo. Like Cal Ripken tells coaches at clinics about teaching baseball fundamentals to kids (I think the same applies here)- you always should explain "why"? Don't always assume that people will always instinctively "get it", even if it seems obvious.
I also like the idea of fielding graphics (lol at the Jeter/Andrus comparison in an earlier post).