MLBAM Announces New Data Stream: The Future is Almost Here.
Well, this is something. At the Sloan Sports Analytics Conference in Boston this morning, Bob Bowman — CEO of MLB Adavanced Media — announced that they are creating an entirely new model to capture data at the MLB level. And it sounds pretty great.
Here’s a video example of what MLB hopes to be able to do with this data.
And a few comments from people involved with the building of this product.
For instance, on a brilliant, game-saving diving catch by an outfielder, this new system will let us understand what created that outcome. Was it the quickness of his first step, his acceleration? Was it his initial positioning? What if the pitcher had thrown a different pitch? Everything will be connected for the first time, providing a tool for answers to questions like this and more.
There will be something for everyone, far beyond what has been available in the past. Miller Park in Milwaukee, Target Field in Minnesota and Citi Field in New York will be operational for this tracking in 2014. The plan is to start rolling out the rest this season so that all 30 ballparks are operational by 2015 Opening Day…
“There’s a speed component to the game from an offensive side, too. You can start to see how quickly they get down the baseline as they make contact, and as they hit a ground ball, or as they hit a double into the gap. If they didn’t score, we always say that the game is a matter of inches, well if he gets thrown out or he is safe at home, you can actually go back and measure it from an evaluation tool, and say, did he get a big enough lead, was he running hard enough, did he take the right angle, you now have the ability to measure that, which we’ve never been able to do.”
Five-time All-Star Steve Sax experienced the new data and said it will be a game-changer for players.
“Players today have really been able to take advantage of technology, and they’ve been able to see video before, during and after the game,” he said. “But this really does take it to a whole new level. They’re going to be able to measure distances, they’re going to be able to see it from different angles, and they’re going to be able to take their knowledge of what they’re doing, instantaneously, to a different stratosphere.
“This technology is outstanding to hitters, but it can also be a great help for baserunners. Not only for the players, but for the coaches as well. This tool can be utilized in gauging distances where it could never have been done before.”
This is what people have been imagining for years, and we’ve seen glimpses of what this kind of data could do with PITCHF/x from SportVision. However, FIELDF/x has yet to turn into a fully realized product, and it sounds like MLB has decided to go in a different direction, as this new data stream is being collected in partnership with Trackman, who use a radar-based system instead of the camera-based SportVision model. According to Eric Fisher of the Sports Business Journal, live tweeting the discussion from Sloan, PITCHF/x will remain active in 2014, but beyond that, its status is “TBD”. MLBAM’s announcement makes it seem like this is an entirely new product, not an add-on to existing F/x technology already in place.
Obviously, the big question from a public standpoint is what kind of access is MLB going to allow to outsiders. PITCHF/x gained hold in part because it was released publicly, but HITF/x was proprietary to the teams themselves, and FIELDf/x was certainly going to be as well. By creating one single data stream that incorporates all the aspects of that promised technology into one data feed, it seems that there could be an all-or-nothing outcome with this new system, where MLB either allows outside analysts access to the entire dataset or restricts it almost entirely to broadcasters and the teams themselves. While no concrete answers have been given to these questions, Bowman’s comments do make it sound like they plan for this data to have a high profile, higher certainly than tools like HITF/x have had during its existence. We’ll cross our fingers and hope that this new tracking system is both a tool for the advancement of understanding for team employees and general fans and analysts alike.
I’d expect to hear plenty more about this system in the days and weeks to come. While hopes for a system like this have been hanging around for years, this is perhaps the most concrete promise yet of a future with comprehensive data that could revolutionize how baseball is seen and discussed.
Dave is the Managing Editor of FanGraphs.
As a developer I am drooling over the possibilities of what could be done with the data…. soooo much information!
This could be massive news. The route-running efficiency calculations by themselves could bring huge changes to how defence is assessed.
MLB: make this information open to the public.
“defence”
I c what you did there
Maybe I’m missing something obvious here, but why would MLB keep the new system proprietary to the teams? If it is an improvement, and we already have access to Pitch F/x, I don’t know why they wouldn’t release it publicly.
Anyone have any thoughts?
The current model is that individual MLB teams and broadcast partners had to purchase the rights to the data in order to pay Sportvision to maintain the infrastructure of the system. This sounds like MLBAM has decided to employ Trackman/Hego to build them something entirely new, which certainly isn’t going to come free. They may very well decide that they see this as a revenue opportunity, and focus on selling the data to broadcasts and using it as an incentive to buy MLB.tv and MLB At Bat.
I absolutely hope it gets released publicly, but I don’t think we can be certain that it will.
Ah, I think I’m understanding. I had assumed that the league pays for the technology as a whole, which I think makes a bit of a difference. Thanks for the info, DC!
It’s clear that MLB is so cash poor that it needs to make money wherever it can. Why make public a source of profit?
/sarcasm
Seems like they can still license the data to broadcasters or use it for pretty graphics on MLB.tv or whatever and still make it available to the public by just delaying the data dump until the game is over (or a daily morning update for yesterday’s games, whatever). We don’t care if we get the data a day late. Broadcasters don’t lose much of anything if we have it after the game. Seems like the likeliest holdup is the individual teams themselves being dickbags worried about the data itself, not its monetization.
What in heaven’s name can a broadcaster do with data this sophisticated and complex.
channel:
Play the video embedded above. This will make for a lovely addition to broadcasts.
But how in the heck does a broadcaster stuck in ERA/BA world use this. Heck swing plane analysis is Martian for most of them. Can you see Kruk dealing with this?
Money, money, money, money
Well, sure. Would you give it away if you paid money to get it? Don’t see an issue with a for-profit company making a profit.
Well, giving the data away to the baseball analytics community could in turn help the teams by getting more people coming up with new ideas on how to use it. Also a good way to engage the fans. I’m pretty sure this will make me enjoy baseball more, which means more money for MLB.
Of course, the benefits may be outweighed by annoyances like requiring metal detectors at all parks. That’d probably be just as big a negative as this is a positive, probably much more, and could really hurt ticket revenue.
I would think Data analysis would only significantly help teams if they have a advantage over other MLB teams. Thus any public analysis wouldn’t improve a front offices decisions OVER those of its competition. However keeping this info from the public wouldn’t offer any advantage either as still all 30 teams would again have the same info, it would just be how they used it that would make a difference. So…. the Rays win?
This data is also massive in size. I don’t think MLB wants to handle public access and the bandwidth involved.
Click my username for a picture of outfielder route maps, from Zachary Levine on Twitter.
This is exciting and scary. We could have a lot more data or we could lose out on one of our only sources of pitching data.
This is nonsense. Who cares? Computer mumbo-jumbo. When this thing can account for guts and heart like pitcher wins and RBI do, then come talk to me.
Dang, I had an /HaroldReynolds at the end but it didn’t post it.
It’s fine; it wouldn’t have been funny anyway.
Not if it was a Ben Zobrist rant. Those are hiLARious…
Nice name
I’ve always thought that one of the only things really holding the WAR equation itself back was the general unreliability and arbitrary nature of dWAR and the defensive metrics that feed into it. This could change that in one fell swoop. I am extremely excited about this.
Maybe maybe not– if a lot of the “error” in defensive metrics is tied to sample sizes not data/formula, then this might really be a minor upgrade…
There is still the issue of how we determine value.
Is an out an out? Are we giving more credit for difficulty, or situation? Who gets credit for positioning? It’s an interesting slew of questions.
Most of those questions could likely be answered with data. If we had a means if tracking speed/acceleration we could determine if a fielder is able to consistently execute in certain scenarios our how much of a defensive play is luck. We would likely have the data to really expand our understanding of how fielding works.
This looks like it could be huge if it is accurate enough and made public.
I read everything until the first video and can already tell how FILTHY this is going to be…It sounds sort of like SportVU in basketball. I should obviously continue reading before I make any judgments, but this already sounds amazing
Sounds great!
However, it would be a shame if Pitch F/X is discontinued. We’re starting to get some nice long-term data out of it.
My understand it is going to be replaced by this which will delivery the same, if not more data on pitching.
But it has to be made public. The good news is that MLBAM is probably the smartest and best run part of baseball. But we have to make sure they, not the teams run it.
The sad part is this means someone out their knows the truth behind Juan Lagares’s insane defense. And we don’t.
Hawk Harrelson’s head just exploded
Hawk Harrelson wouldn’t even be able to comprehend the press release, much less have his head blown apart.
good riddance
Fantastic! (If it is released to the public.)
Waste of time and money, this is just an example how companies are trying to monetize the MLB and it’s clubs, and it’s a shame that they would be duped into seeing a trajectory and think somehow they can figure this game out better.
In fact, technology is moving one step closer to ruining this game with instant replay as well, it used to be a human game and now they are trying to make us slaves the ‘the machine.’
Yeah! How dare a company try to monetize baseball! This isn’t some money-making business, this is God honest, down-to-earth baseball. And technology, give me a break! It’s pretty clear that technology cannot be used to amplify people’s enjoyment of the game. What the people want is to watch a age-old tradition without that fancy replay and especially without those sissy helmets, wall padding, and elbow guards. For God sake, we want to see more blood and broken bones!
“And technology, give me a break! ”
Exactly– what is the BS with baseball on TV?? And before that radio! Don’t worry about this change– baseball was ruined by technology 93 years ago…
If somebody wants to peel off a few million bucks from the used car salesmen country club… more power to them.
I think someone said that about radio broadcasts in the 20’s.
Man, I completely missed the part about how human fielders will be replaced by robots! Thanks for pointing that out. Here I thought this would just tell us more about what the humans playing the game actually did so we can appreciate their skills and athleticism more.
People like you are the reason we don’t have teleportation and a cure for cancer. STFU and let us move forward.
I see that this will certainly help improve the accuracy of outfield defense metrics, but how will this help improve evaluation of infield defense?
Accurate ball tracking and positioning of the fielders.
Say Player A is 10 feet away and Player B is 15 feet away. Player B fields the ball. We now know player A should be “punished” and Player A should be “rewarded”.
C’mon, it’s player A’s last season in the Majors, give him a gift instead.
Derek Jeter?
25feet between them? Punish both for poor field positioning.
If it’s not made public, maybe Anonymous can work some magic? /totallykiddingpleasedontcometomyhouseNSA
Why is everyone drooling about this? This isn’t going to make the game better. MLB is just trying to make money any way they can. They don’t care about the game. If they did they would have started drug testing players a long time ago. The Olympics started drug testing athletes in 1968 because they knew athletes would use drugs to enhance their performance so why did MLB take so long?
k
You sound like a smart person who knows what logic is.
Seattle Slew was actually a horse, not a person. So the missing logic at least has an explanation.
Would you like a side of rice and vegetables to go with that red herring?
He may have been inebriated. In which case the herring was pickled.
The most impressive part of this post is that Heyward catch. Holy hell the man is 6’5″ 240 and can still do that. Efficiently as well
apparently!
Personally, I don’t find this terribly exciting.
This will definitely allow a team to optimize defensive positioning. And will probably help identify injuries (or at least help identify players who have injury inducing tendencies). But not much else.
This will only exist in major league parks. We already know how good most major leaguers are. Sure, this will make the numbers a bit more precise, but tells us nothing about minor leaguers, or foreign players, or amateurs. We already know a hell of a lot about major league players, and diminishing returns are certain at some point.
As for defensive metrics, someone already mentioned that sample size is one of the prevailing issues here. This system will help, but the sample size isn’t getting bigger, so the error bars will remain large. I guess the question is how much of the current error is due to sample size (and other statistical issues) and how much of it is measurement error?
Plus, there’s the question of how reliable will this system be? I’m assuming it’ll be pretty good, but until the players are required to wear sensors throughout their bodies, there will still be problems.
Personally, I think this system will be more important for broadcast than analysis. Hopefully I’m wrong
A big part of the reason sample size is an issue with defensive measurements is that the measurements are so imprecise. In fact, they aren’t even measurements, they are just human judgements. The reason the sample size needs to be enormous is because most of the data is simply random noise generated by the people who are judging the plays.
I understand your point, but sample size has nothing to do with measurement precision. Hence the statistical error vs measurement error comment.
“…but sample size has nothing to do with measurement error”
That is just not true. The more precise the measurement, the smaller the sample size that is required for reliability. Insofar as these data would significantly improve the precision of the defensive measurements, then one can reasonably expect the required sample size for reliable conclusions to decrease accordingly.
Now we just have to keep Fox from hiring McCarver for another postseason. He is destined to ruin it for everyone…
I saw this on MLB Network last night. One word…Outstanding.
All I know is I’ve never seen a higher route efficiency than Heyward’s 97%.
I know this is kind of jokey, but it certainly seems like you couldn’t lose much more time/distance than Heyward did and still make that play. Game of inches, indeed.
Aside from the value of this technology for baseball analysis, there is an ongoing research effort by various people (not me) to understand how players track down fly balls. Dr. Mike McBeath from ASU,who gave a talk at the 2012 Saberseminar, is an expert in this area. He and I worked on a paper together a few years back about “paradoxical popups”, for which the usual tracking algorithm used by fielders will sometimes fail for popups. You see this quite often where an infielder settles under the popup, then has be backpedal at the last minute. It’s all discussed in the paper: http://baseball.physics.illinois.edu/AJP-Aug2008.pdf. Geting real-life data from actual MLB games would be important input into this type of research.
Now if only they can figure out the “glowing puck” technology that Fox debuted on it’s NHL coverage circa 2002…
The glowing puck technology was one of Sportvision’s (of Pitch Fx) first products. Closely followed by the projected first down lines in football.
The puck thing was stupid and took away from the game. I think the problem is sought to correct (not being able to see the puck) was largely corrected at this time with mainstream use of High Definition TVs. The first down line on the other hand, has very much improved NFL games on TV.
I thought the first-down lines came first… I think I remember an SI article on it when it was new, that they had to bring a trailer full of high end computers [SGI] on-site to do it.
http://www.popularmechanics.com/outdoors/sports/4301993
So, I did recall the article [Popular Mechanics, not SI], but it even references borrowing the NHL tech.
Could you do a piece on Brian McCann and what you can expect? After looking at his stats I’m thinking he’s someone who might fall off fast. It seems that his best years are already behind him. I’d love to see what you think.
A sad day indeed. Where have you gone Joe DiMaggio?
He should be pretty easy to locate; he hasn’t moved very much in the past 15 years.
/Dave Cameron puts in 2 week notice,/Signs with MLBAM to head their new sabermetric paysite, /We follow.
Weird. Last I heard of Steve Sax, he was serving six consecutive life sentences.
Steve Sax: fun.
His brother Yakety Sax: waaaaay more fun.
If MLBAM *doesn’t* make this information public, then another whole line of inquiry can be opened up on the differences between what front offices are doing v. what we see with existing stats.
“Well, our read on WAR, UZR, DRS and some defensive metrics on Fangraphs say that this player has no business being out in center field… but the ballclub is running them out there anyway. They must see something in the MLBAM data that we can’t see.”
It’s an “appeal to authority” debate flaw, I know, but really for the first time we have some proof that the clubs will have access to significant and likely much more accurate data than say FG or BR does, which means more than ever that FG and BR’s advance statistics are best guesses possible from a very incomplete data set.
I’m taking things like UZR and WAR with a much, much bigger grain of salt than I did before, now.
Very true. UZR and DRS depend too much upon the pitching and hitting. So there’s a lot of luck going into it. We need to somehow have hitting/pitching independent fielding. MLBAM data gives us that.
This is really cool. The information that’s being collected should lead to all sorts of interesting analysis which will make for great *reading*.
But as for watching a game on TV, I could definitely see this being over-used in another misguided attempt at making the game “more interesting” (cough, Fox, cough) when in fact a baseball game is inherently interesting already without added bells and whistles like super extreme close ups of pitchers faces (when we’d rather see what sign the catcher’s putting down), needless crowd shots, and cameras buried in the dirt a few feet in front of home plate.
I think the real benefit of this data will be analysis of large quantities of data – trends, players’ tendencies, etc. – but I am already braced for the meaningless ways Fox and others will put it to use.
Heard you talk about this on the podcast with Carson.
I believe you left out one key covariate that’ll be important in evaluating (and predicting) fielding value: hang time on the ball.
With that info, along with initial player position as you mentioned, the SAFE model will be significantly more accurate and, thus, incredibly useful in the MLBAM new data setting.
FYI, here’s a link to the model’s info: http://www-stat.wharton.upenn.edu/~stjensen/research/safe.html
I think data will make baseball video games more realistic. It will not in general improve players. I have been coaching forever, you can teach skill, but not talent. I teach anyone to catch, throw and hit, but that extra called talent, is what you must possess to make it to the bigs and video can’t teach you that.