MLBAM unveils "First-ever complete and reliable measurement of every play on the field”

derekson · Mar 1, 2014

This is absolutely unbelievably amazing. The route efficiency for outfielders will revolutionize our ability to compare outfielders, it seems like a really direct way to evaluate routes instead of the subjective impressions we've relied on in the past. The reaction time stat is really cool too. It'll be funny to see how terrible Jeter scores on that.

ivanvamp · Mar 1, 2014

Ok, quick question about OF route efficiency. I would assume that a 100% efficient route would be a straight line from the point of departure (where a fielder is standing when the ball is hit) to where the ball is caught. However, there would be times when the *best* route would actually not be a straight line.

For example, say there's a runner at third and one out, and the batter hits a fly to medium-depth left-center field. The CF would not want to go straight to where the ball is going to land, because it doesn't put him in the best position to make a throw home. Rather, the best route is to circle behind the ball somewhat, so he can come in on it and make the catch with his momentum coming directly to home.

Do we know if the system would account for such a situation as this?

derekson · Mar 1, 2014

Typically the outfielder only has time to circle behind and line up momentum for a throw on relatively routine plays though, while evaluating routes is really only important on tougher plays. On a fliner in the gap or a sinking shallow line drive the OFer is just trying to get to the ball, not worrying about setting up a throw.

Laser Show · Mar 1, 2014

I can't even articulate how cool the reaction time stat is. Heyward's first step was in 0.2 seconds, to Johnson's 0.5? Wow. Beyond amazing.

This is going to revolutionize how we understand defense.

ivanvamp · Mar 1, 2014

derekson said:
Typically the outfielder only has time to circle behind and line up momentum for a throw on relatively routine plays though, while evaluating routes is really only important on tougher plays. On a fliner in the gap or a sinking shallow line drive the OFer is just trying to get to the ball, not worrying about setting up a throw.

Sure, I understand that. But there are enough of those plays that I describe that I would hope the system would be able to account for it. Because the difference in a ballgame could easily be a sacrifice fly that was "misplayed" by an OF who took too straight a route to the ball and thus couldn't make a strong enough throw to the plate. I want the system to credit OFs for making the *best* route, not simply the "most efficient" route.

finnVT · Mar 1, 2014

ivanvamp said:
Sure, I understand that. But there are enough of those plays that I describe that I would hope the system would be able to account for it. Because the difference in a ballgame could easily be a sacrifice fly that was "misplayed" by an OF who took too straight a route to the ball and thus couldn't make a strong enough throw to the plate. I want the system to credit OFs for making the *best* route, not simply the "most efficient" route.

They don't show it, but I would imagine they can continue tracking the ball speed beyond just the hit. So you wouldn't use *just* route efficiency when evaluating a fielded ball, you'd use a combination of route efficiency, throw velocity, throw accuracy. It is quite possible that by sacrificing some route efficiency you'd be able to increase throw velocity and/or accuracy, so that overall you'd score higher.

It's essentially the same point as the baserunner efficiency they show. 100% efficiency essentially means stopping at each base and making a left hand turn. But, the tradeoff for getting that higher efficiency is a lower overall running speed. So "ideal" is not necessarily 100% efficiency, but maximizing efficiency + running speed. Same with fielding: ideal isn't 100% efficiency (on hits, at least), it's maximizing efficiency + throw velocity + throw accuracy.

What you want ("best" routes vs most efficient route) is going to be some overall defensive metric that combines multiple of these tracking stats, rather than the raw stats themselves. Which some clever person will certainly come up with.

Laser Show · Mar 1, 2014

ivanvamp said:
Sure, I understand that. But there are enough of those plays that I describe that I would hope the system would be able to account for it. Because the difference in a ballgame could easily be a sacrifice fly that was "misplayed" by an OF who took too straight a route to the ball and thus couldn't make a strong enough throw to the plate. I want the system to credit OFs for making the *best* route, not simply the "most efficient" route.

I would hope, and have to think, this is the case. Think of the dynamics on a lined shot to the gap that gets down. Circling around behind the hit some, allowing you to play it in front of you, is likely the better idea than running straight at it and trying to cut it off, risking letting it get past you to the wall.

(Depending on the situation).

Sprowl · Mar 1, 2014

Laser Show said:
I can't even articulate how cool the reaction time stat is. Heyward's first step was in 0.2 seconds, to Johnson's 0.5? Wow. Beyond amazing.

This is going to revolutionize how we understand defense.

I'm very pleased to see MLBAM open up the data stream. Crowd-sourcing will do more with this resource than the clubs will by themselves.

I'm looking forward to seeing batted ball speed and angle (no more fliners!) and Jackie Bradley's route efficiency. Road trips to Minnesota and Milwaukee just became must-see MLBtv.

terrisus · Mar 1, 2014

So we can do all that, but we still can't have automated balls and strikes.

Reverend · Mar 1, 2014

ivanvamp said:
Sure, I understand that. But there are enough of those plays that I describe that I would hope the system would be able to account for it. Because the difference in a ballgame could easily be a sacrifice fly that was "misplayed" by an OF who took too straight a route to the ball and thus couldn't make a strong enough throw to the plate. I want the system to credit OFs for making the *best* route, not simply the "most efficient" route.

As a tool for data evaluation, clubs will simply disaggregate the types of hits. They already do this in differentiating between hard hit balls and softly hit balls when they plan things like shifts and evaluations of batters. They do this to cut out the noise and only look at what has predictive value, e.g. bloop hits have no predictive value as to what the hitter will do next time he's up whereas hard hit balls do.

So with this, for example, they could look at how an OF does just on routes he takes on line drives. Stuff like that. And then they could look at things like whether or not the guy takes the right route for balls that are going to drop to play the bounce or if he always plays the landing first and then adjusts. It remains to be seen, obviously, but Farrell indicated they slice the data up in all sorts of ways beyond just "all contact" which isn't as useful.

mabrowndog · Mar 1, 2014

Sprowl said:
I'm very pleased to see MLBAM open up the data stream. Crowd-sourcing will do more with this resource than the clubs will by themselves.

I'm looking forward to seeing batted ball speed and angle (no more fliners!) and Jackie Bradley's route efficiency. Road trips to Minnesota and Milwaukee just became must-see MLBtv.

Has MLBAM actually said they're making the data stream available? I saw the tweet above noting we'll have "access to this technology", with that last word open to wide interpretation, but nothing in the thread mentions the data itself.

Hopefully it is indeed the case, because you're absolutely right. I was completely disheartened when they decided to keep HitFX and FieldFX in-house as their private toys without even hinting as to their plans. All I could imagine was how much free research MLB could gain by letting a planet of stat geeks take it to the limit under ever-present peer review, and how incredible it would be for fans and the analysis community alike.

soxhop411 said:
It might be a little hard. Apparently they said at the event that each game is 7TB! So I hope that if the public is allowed to use it they will either compress it or let us download the sections we want. Or these videos can replace what we currently get under the highlights for each game. Say JBJ makes an amazing catch. The video under the highlights section would include all of the information that is shown in the pictures

Just want to say this one thread and the content you've provided herein makes up for every unlinked URL and empty message you've ever posted on this site. Seriously, well done.

teddywingman · Mar 1, 2014

This is awesome news. It will be really fun to see how this new data compares to what we already think about certain players.

terrisus · Mar 1, 2014

teddywingman said:
This is awesome news. It will be really fun to see how this new data compares to what we already think about certain players.

It comes just in time too.
One more season and Jeter would have escaped.

timlinin8th · Mar 1, 2014

Reverend said:
So with this, for example, they could look at how an OF does just on routes he takes on line drives. Stuff like that. And then they could look at things like whether or not the guy takes the right route for balls that are going to drop to play the bounce or if he always plays the landing first and then adjusts. It remains to be seen, obviously, but Farrell indicated they slice the data up in all sorts of ways beyond just "all contact" which isn't as useful.

Also, have to think that the data can be parsed to weed out more routine flyballs where a fielder might take a fairly roundabout route to the ball, or a sac fly where they are lining up a throw, vs. the real meat-n-potatoes where a ball drops in for a hit that can be shown where a higher route efficiency might have prevented the hit (essentially, proving whether or not "Iglesias would have had it!" for the crowd that makes such player comparisons)

CoffeeNerdness · Mar 1, 2014

The base running stats will be fascinating as well. Not only on close plays like that picture shows, but getting to see the raw athleticism of a player like Andrew McCutchen in a whole new way will be amazing.

edit - it did seem like Fox used some of this in the playoffs- that amazing Iglesias run and catch comes to mind.

Time to Mo Vaughn · Mar 1, 2014

terrisus said:
It comes just in time too.
One more season and Jeter would have escaped.

It's in one AL park not even in his division. With that SSS, he will escape.

ivanvamp · Mar 3, 2014

teddywingman said:
This is awesome news. It will be really fun to see how this new data compares to what we already think about certain players.

It won't be very long (a few years, at most) before part of our normal baseball vocabulary will include terms like "route efficiency" and "first step time", things like that.

The Sox are looking to pick up player A or player B in free agency.

Player A:
.301/.377/.455/.832
93% route efficiency, 0.08 first step time

Player B:
.278/.334/.435/.769
98% route efficiency, 0.02 first step time

Should make it real interesting.

brandonchristensen · Mar 4, 2014

What a coincidence this comes out in Jeter's last year.

Sprowl · Mar 4, 2014

mabrowndog said:
Has MLBAM actually said they're making the data stream available? I saw the tweet above noting we'll have "access to this technology", with that last word open to wide interpretation, but nothing in the thread mentions the data itself.

Hopefully it is indeed the case, because you're absolutely right. I was completely disheartened when they decided to keep HitFX and FieldFX in-house as their private toys without even hinting as to their plans. All I could imagine was how much free research MLB could gain by letting a planet of stat geeks take it to the limit under ever-present peer review, and how incredible it would be for fans and the analysis community alike.

Just want to say this one thread and the content you've provided herein makes up for every unlinked URL and empty message you've ever posted on this site. Seriously, well done.

Me too. I suspect that the data providers got some hefty proprietary fees from the clubs in 2011-13, but reached the point of diminishing returns more quickly than they expected, and now see the advantages of letting the data loose. PitchFX worked because the world got hacking on it (and Dan Brooks got there first and best). I don't think the best stat-crunchers MLB teams can hire will do nearly as well as Everygeek will do within six months. The internet is a mighty sifter.

Also kudos to soxhop. This is the best baseball news since Game 7, 2013.

Rasputin · Mar 4, 2014

I want to know what team is going to put this in their minor league system and how much of an advantage it will give them.

lambolt · Mar 4, 2014

ivanvamp said:
It won't be very long (a few years, at most) before part of our normal baseball vocabulary will include terms like "route efficiency" and "first step time", things like that.

The Sox are looking to pick up player A or player B in free agency.

Player A:
.301/.377/.455/.832
93% route efficiency, 0.08 first step time

Player B:
.278/.334/.435/.769
98% route efficiency, 0.02 first step time

Should make it real interesting.

you don't think it more likely that this will be boiled down to some overall average run saving metric based on position? That seems more likely, i.e. Player A and B will have some "X runs saved on average" or something

as a snarky aside, what the hell will TV stations do now that there's yet another way to have on screen information that will probably highlight just how many fucking morons are there for "analysis"

lambolt · Mar 4, 2014

Follow up question, given how phenomenally detailed this shit is, how close are we to a discussion of balls / strikes by camera? Too much of a problem with umpires unions etc? I know that player size affects strike zones but surely that could be calibrated by the league on a player by player basis. The umpires could still be there to rule on bat contact and loads of other stuff so it doesn't have to mean them losing their jobs.

ForceAtHome · Apr 18, 2014

soxhop411 said:
Has anyone seen this put into use yet? (ie shown on MLBN or on a broadcast?)

Not nearly to the extent I may have hoped, but the Nats broadcasters on MASN sometimes mention batted ball speed in-game.

derekson · Apr 18, 2014

I don't think they've got it past testing yet. Plus it's only going to be in 3 parks this season: Target Field, Citi Field, and Miller Park.

The batted ball speed comes from a different source that's been around for a few years.

HriniakPosterChild · Apr 19, 2014

Can it determine if the first fucking batted ball of a goddamn game lands fair or foul?

Fred in Lynn · Apr 19, 2014

HriniakPosterChild said:
Can it determine if the first fucking batted ball of a goddamn game lands fair or foul?

I'll bet it can, but you still need human beings who are willing to use the information.

terrisus · Apr 19, 2014

Fred in Lynn said:
I'll bet it can, but you still need human beings who are willing to use the information.

Or robots.

seantoo · Apr 21, 2014

ivanvamp said:
Sure, I understand that. But there are enough of those plays that I describe that I would hope the system would be able to account for it. Because the difference in a ballgame could easily be a sacrifice fly that was "misplayed" by an OF who took too straight a route to the ball and thus couldn't make a strong enough throw to the plate. I want the system to credit OFs for making the *best* route, not simply the "most efficient" route.

I understand your point however you are referring to what is a very small fraction of a very small fraction, to the point that it's likely not going to matter.

cromulence · Apr 23, 2014

ForceAtHome said:
Not nearly to the extent I may have hoped, but the Nats broadcasters on MASN sometimes mention batted ball speed in-game.

This has clearly been a MASN focus this year as I've also seen it a lot for O's games. They tell you that balls hit at certain MPH are X% likely to be hits, and so on. It's kind of cool but not at all what I'm really excited for with regards to this new data.

OttoC · Apr 24, 2014

I thought I'd add some information about attempts at quantifying fielding for younger members. Circa 1985, Bill James's Project Scoresheet, which moved to John DeWan's STATS and then was employed by scorers for Project Scoresheet used the following field chart to record the sector:

where batted balls were fielded on outs,
where ground balls went through the infield on hits,
where batted balls are fielded or stop rolling in the infield,
where line drives, fly balls, or pop flies first drop,
where home runs leave the park.

One thing to notice is that the same chart was used for all parks and while outfield sectors had approximately the same depth, they varied greatly in volume so that a throw from the back right-hand corner of a ball hit to right field would be much different in distance than one thrown from one thrown from the front left-hand corner of the same sector. There also was no record of where the ball rolled in the outfield and there is no record of how well players were positioned.

There basically wasn't much sabermetricians could do with the data and you needed to rely upon your eye to for judgment. And since even this type of data was not available a few years prior, there was no way of comparing players from previous eras except by having seen them play.

MLBAM unveils "First-ever complete and reliable measurement of every play on the field”

news aggravator

Member

captain obvious

Member

Member

captain obvious

superspreadsheeter

Member

mikey lowell of the sandbox

formerly: imgran

news aggravator

for king and country

Ask me about total zone...or paint

news aggravator

Looks like Zach Galifianakis

formerly: imgran

Member

Member

RIP Dernell

captain obvious

Loves Aaron Judge

mikey lowell of the sandbox

Will outlive SeanBerry

http://b.globe.com/13BHr47

http://b.globe.com/13BHr47

news aggravator

Member

Member

Member

Member

formerly: imgran

toots his own horn award winner

Member

Member