June 24, 2003
Ron Antinoja, Part II
Continuing from Part I...
Baseball Prospectus: How can a team use Tendu to, for example, evaluate individual personnel?
Ron Antinoja: Let's say a team wants to find a guy who can handle lefty pitchers. They can go to Stats Inc. or MLB already and get data. But we can break things down even further, provide more information and a more complete view. One player can be the difference between winning a championship and not--it's important to get these decisions right. Teams can also use (Tendu) for contract negotiations. Both the player and his agent as well as the team can use Tendu to go beyond the basic stats and better make their case.
BP: What was the initial reaction when you came out with Tendu?
RA: When I first introduced it, there (were) no data collected, just the software. I took it to all different levels of baseball and I don't think I've shown this to any baseball person yet who wasn't intrigued. I showed it to 100 college coaches once at a conference once; they all said they wanted to use it, but we got zero sales out of it.
The picture changed once we started collecting data, and the focus changed a bit toward major league teams. Even then though, there's a barrier, in that most teams don't budget much for this kind of information. Some of our competitors charge $50,000, $100,000 for this kind of material, so their eyes might glaze over when they're first approached. When I tell them our price--we charge $5,000 for a license for one person, $30,000 for a team, which includes 10 accounts--they sort of say, 'oh, really?'"
It still takes time to get them to come around completely though. A lot of teams will say they don't trust anyone that hasn't been trained by their own people to put the data in. I've come to understand that objection--they have this real shield all around them, and it's hard to get to them.
BP: Well you talked about how your scorers collect the data. But what sets them apart from the rest--in other words why should teams trust your people above others?
RA: I have high standards--I only expect that they be perfect. I compare it to taking an exam in school. With 100 questions, if you get 96 of them right, you get an A. If they get 96% right for us, that's one our worst results. These aren't like those exam questions; it's more like homework, where you have every opportunity to double check and find every mistake you've made. When you're done it needs to be perfect though. Teams depend on our data to be as accurate as possible. If it isn't, we haven't done our job.
BP: What types of people become scorers for Tendu?
They're mostly college level baseball players, some minor leaguers. We also have Sean Spencer, who's been in the minor leagues for a while and played briefly with the Mariners and the Expos. Sometimes we'll find people who are good at recognizing pitches, but don't have the focus to do this well. One of my competitors once said that baseball players can never be meticulous enough to do a job like this. But to me it's just another cross section of the population, where some people will have the focus to do it and some won't. The next step is to find people who can handle statistical analysis on a high level. That can give clients--the teams that buy the data and software--added value.
BP: What does a team get for its $30,000?
RA: They get 10 user accounts, plus we offer some consulting as well. We'll also go in and give a training class to get them started. I will also do a custom extract for someone who wants it. So every day, every week or whatever I can send them custom reports based on what they want.
BP: How do you get people--a lot of whom aren't necessarily experts at digesting this kind of research--to soak it up and understand it?
RA: The basic principles aren't that complicated when you break it down. If you can learn that a pitcher throws 95% fastballs on the second pitch, that's going to help a great deal.
BP: Who are your clients right now?
RA: The Mets have paid $30,000 for 10 user accounts. Three more teams have several user accounts, but have not paid yet. I'll often set things up that way, where a few people from a team will give it a try to see if they like it, then we look into making accounts official later on. Ten more have accounts, but are trying it out and are too sporadic to be considered regular users right now. Part of this issue is a matter of focusing on getting funding for the company rather than enforcing payment strictly right now.
BP: What types of people are your clients?
RA: They're thinking baseball players. Some people have suggested that no players would be interested in this. But I think we can reach 20% to 40% of them eventually. Maybe a lower percentage will use the Pitch Sequence Analyzer. But I anticipate that a lot of them will use something like the spray charts. We feel it's going to spread...someone will be using it and another player will say 'what is that, let me see that.'
BP: You're not naming names, are you?
RA I will say that (Oakland A's pitching coach) Rick Peterson does a lot of work with us. The Mets became customers because of Vern Ruhle, who heard about us through Peterson, and also partly because Art Howe left the A's and went to the Mets. The whole Mets coaching staff has an account--it's seven coaches and three people in the front office total.
BP: How do you scout for potential customers?
RA: The most valuable place for Tendu to be used is in preparation for an upcoming game, typically by the pitching staff and those that work with the pitchers. For example, I know that Peterson does a lot of work with the catchers, getting them to help pitchers prepare for a game and go through it. Also hitting coaches can use it to gain an edge, or a bench coach can spot tendencies and better position fielders. If they're not very computer-oriented, we may also want a more technical person in the organization to help them get the data they need--every team has some techies, either associated with video systems, or sometimes with data.
Ultimately the real secret to anything in baseball is that a team often has to learn about a new concept from someone they trust. Peterson has been helpful in that regard, as has Craig Pippin, who was a minor league pitcher in four or five organizations and knows everybody.
BP: You mentioned catchers--why are they so important in the process?
RA: Traditionally catchers are the ones that call games. And we know that some catchers handle pitchers better, whether it's calling the type of pitch to throw, location, pitchouts, or anything else. If a pitcher has too much to think about, it can disrupt his concentration and the flow of the game. So yes, we'd really like to work directly with catchers and show them what we have to offer.
BP: Are there other ways you think you can leverage Tendu?
RA: We've been in talks with ESPN. I believe that most organizations like an ESPN will ask for packages, just as they did from different stats organizations. But I also believe the best media organizations will have people who understand baseball well, will like to dig and explore. What I want to do is offer the license and the tools to do that digging and exploring. The database is so rich, you could have different teams or media groups developing their own style for using this. And wouldn't it be great if some reporter finds something in the software that no one else can find?
I think this can be extended to fans too. My belief is that fans have been denied much of the most interesting data in the game, really a whole different view of the game. Hopefully one day we could get this out to fans, who could access Tendu through wireless devices. Everyone could come into the ballpark with their own Palm Pilot or cell phone and begin to do this stuff, to follow along in real time with what the team sees.
BP: Are you working on improvements to your existing software?
RA: One big thing is when we find a pitcher or hitter's results in a certain situation, teams want to be able to easily compare that performance to league average, so that's something we'll need to address going forward. Also, right now the (Pitch Sequence Analyzer) tool is designed mainly for a human being to go through and find red flags. But I think eventually we'll be able to find them for you. We want to develop a way to automate tendency detection. We want people to be able to eventually see not only what a tendency is, but if or when it gets changed--when a player identifies a problem and makes an adjustment. People always say how it's a game of adjustments. When a player wins Rookie of the Year then struggles the next season they call it the sophomore jinx, but really it's just that the league has adjusted to the player and he hasn't yet learned to make adjustments back. We'll also want a tool that will cover all aspects of basestealing...pickoff attempts, pitchouts, hit and run attempts. Something with the same predictive flavor that you see with the pitching and hitting tools.
If I have one pet peeve about baseball statistics though, it's fielding percentage--it's a complete joke. I want to record exactly what happens on every ball in play. I want to track poorly fielded balls, where the official scorer doesn't score an error. Or what about when a hitter gets a single with a man on first, but the fielder does something--or doesn't do something--that allows the runner to get to third base, where he might otherwise only have gone to second. That kind of play often doesn't show up in fielding percentage, and I think teams would benefit from having more accurate information on those kinds of plays.
I think there isn't enough reliable information out there for teams to really evaluate fielding the way they could or should. When I talk to people from places like Elias, they always say they don't want stats that are subjective. I say I don't mind at all making subjective judgments taking into account concepts like fielding difficulty, as long as it's done by the right people, in the right way.