Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Baseball Newsstand > Baseball Primer Newsblog > Discussion
Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Wednesday, May 21, 2014

A look at MLB’s new player motion tracking system - MLB - Ben Reiter - SI.com

When this goes online I may never leave my computer.

Jim Furtado Posted: May 21, 2014 at 01:49 PM | 18 comment(s) Login to Bookmark
  Tags: mlb network, sabermetrics

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

   1. Pat Rapper's Delight Posted: May 21, 2014 at 02:19 PM (#4710837)
In mankind's history of great inventions, this may be the best.
   2. salvomania Posted: May 21, 2014 at 02:24 PM (#4710840)
That first video is pretty incredible. First, it's an amazing play by Tejada and Murhpy, but then, all that data!

As interesting as it is, though, I don't know how it will be used---although as with many new things, people will figure out useful applications.

I suppose if, as in the case highlighted in the story, you know one of your shortstops is twice as slow to react to a batted ball as your other SS, you could maybe focus on that as something to work on...
   3. ellsbury my heart at wounded knee Posted: May 21, 2014 at 03:08 PM (#4710899)
As interesting as it is, though, I don't know how it will be used---although as with many new things, people will figure out useful applications.


It's basically ultra-refined, objective scouting information. It seems like any situation where you might rely strongly on scouting information (short-term player analysis where traditional, outcome-based stats haven't stabilized, or advanced scouting or more detailed defensive analysis looking at range or arm strength or something) this stuff could be very useful. Some naive model where you just throw all the measured variables into a selection model and see which tools or whatever correlate best with good outcomes would be neat. You might even be able to look at performance changes that would be difficult to detect with the naked eye as pre-injury markers.
   4. Sunday silence Posted: May 21, 2014 at 03:21 PM (#4710922)
how many variables to primates think it would take to create a reasonable defensive metric using this data? I can think of: reaction time, closing speed, arm strength/accuracy, IQ (e.g. throwing to the right base, covering second etc.)

what do you think?
   5. Random Transaction Generator Posted: May 21, 2014 at 04:05 PM (#4710990)
how many variables to primates think it would take to create a reasonable defensive metric using this data? I can think of: reaction time, closing speed, arm strength/accuracy, IQ (e.g. throwing to the right base, covering second etc.)


Maybe something about original positioning location vs final fielding location.
That you can differentiate between players who are smart enough to position themselves before the play happens, and those that react to the play well enough to overcome bad/normal original positioning.

If you see player A has to move less often (or move less total distance) versus player B who is always having to hustle to get to the right place (but does so because of his speed/skill), then maybe there is something in the defensive metric that could account for that.
(the method vs the results)
   6. bjhanke Posted: May 21, 2014 at 04:27 PM (#4711018)
Random - You have a good insight, but isn't the following an additional help?

If you have one guy who has great speed range, and another guy who lives by positioning, then should not your proper response as a manager or infield coach be to grab hold of the speed guy and go upside his head until the 2x4 has gotten the attention of the mule, and say, "Look, you need to learn how to position like the other guy. Your speed is a SUPPLEMENT to that, not a REPLACEMENT. Yeah, you can equal his range now, but you can have much MORE range if you can add positioning to your speed. Lou Boudreax PLUS Ozzie Smith equals the greatest defensive shortstop ever. Same principle applies at every position. - Brock Hanke
   7. Jose Can Still Seabiscuit Posted: May 21, 2014 at 04:43 PM (#4711042)
Reaction time seems like the important thing to me. That feels like the thing that indicates a "good" defensive player who gets the most out of his abilities rather than the guy who can close down a ball because of his speed.

For example, two outfielders make diving catches that are "web gems." The guy who had the better reaction time made the better catch because he maximized the ground he covered.
   8. Greg K Posted: May 21, 2014 at 05:03 PM (#4711067)
I wonder if breaking down infield defence into things like first step, release time, arm velocity, could be used to determine whether you can move a guy around or not. For instance at 3B you need this particular mix of ingredients, at 2B this one. In other words, if the Jays had this data would it help them determine whether moving Lawrie to 2B to get Francisco in the lineup is worth it?

If nothing else it's fun as a fan to have info on who has the best first step, which catcher pops up the fastest, who has the strongest arm.
   9. Walt Davis Posted: May 21, 2014 at 06:19 PM (#4711124)
In terms of a metric, the most important info is probably the starting position and the range covered. That should eventually replace the current zone-based ratings. You still have a problem of whether to credit the defender or the team for good positioning but this is the essence behind the concept of expected plays made. At a minimum, in/out-zone would be defined in terms of distance from starting position.

Whether that would really make any difference in Rfield/UZR is another question.

Like pitch-framing, these stats might feed into C defensive metrics. But I'm not sure that pop time vs. throw velocity are going to tell us substantially more than what CS% is already telling us.

Scouting/development-wise there would seem to be a lot of value here but I assume this system is not in minor-league parks much less college/high-school. I assume this sort of technology could be set up at a baseball "combine". There is some ML scouting of course and I suppose it could turn up things like "don't let Hamilton get more than 10 feet" or even that a base stealer might tip his attempts.

There may be several early indicators ... e.g. if the velocity off of a young/aging player's bat is dramatically different in April than it has been in the past, maybe this is an early indicator of breakout/collapse. Of course if it's April, it may be too late to really do anything about it.

It will obviously provide a lot of the answers to why a specific play turned out the way it did but that doesn't always mean that it will have predictive value. For example, the Tejada-Murphy DP was wicked fast but Tejada is considered average at best and Murphy is considered one of the worst defensive 2Bs around. Presumably with a larger sample, these numbers would bear that out (or convince their reps are undeserved) but, at the end of the day, will it tell us more than the number/rate of DPs and other plays made?
   10. Lassus Posted: May 21, 2014 at 06:31 PM (#4711126)
Jhonny Peralta adds at least another two feet if not three to his lead, and then stops, and then starts up again after the measurement of his lead is listed.
   11. puck Posted: May 21, 2014 at 06:32 PM (#4711127)
As interesting as it is, though, I don't know how it will be used---although as with many new things, people will figure out useful applications.


People will figure it out if they have access to the data.
   12. I am the Can Posted: May 21, 2014 at 06:37 PM (#4711132)
One thing you could certainly measure is assists and putouts made by infielders versus their expected values. If you know the precise velocity of a ground ball, as well as the break, acceleration, and speed of the runner, you can calculate the likelihood of an out on similar plays and measure fielders' performance versus average/expected values. That seems pretty darned useful in figuring out who the best fielders are.

And, wow, does this seem awesome!
   13. cardsfanboy Posted: May 21, 2014 at 07:54 PM (#4711162)
Like pitch-framing, these stats might feed into C defensive metrics. But I'm not sure that pop time vs. throw velocity are going to tell us substantially more than what CS% is already telling us.


We'll get a better idea of how much credit for caught stealing to give to the catcher, the pitcher or just a bad jump. I think all of this information is going to be valuable when there is enough data to parse it down and come up with a framework for "average".
   14. Walt Davis Posted: May 22, 2014 at 12:44 AM (#4711296)
One thing you could certainly measure is assists and putouts made by infielders versus their expected values. If you know the precise velocity of a ground ball, as well as the break, acceleration, and speed of the runner, you can calculate the likelihood of an out on similar plays and measure fielders' performance versus average/expected values. That seems pretty darned useful in figuring out who the best fielders are.

Yes, it should allow you to more precisely measure the difficulty of a play made/not made compared to the zone ratings we currently use. But that's no guarantee it will substantially effect the results -- i.e. zone-based systems are probably over-valuing some plays and under-valuing others but coming out about right on average.

Reducing measurement error is a good thing and it would reduce the year-to-year bouncing around in Rfield which would be nice but we probably still end up with a guy who totalled +15 Rfield over 3 years totalling +15 SuperField over 3 years.

We'll get a better idea of how much credit for caught stealing to give to the catcher, the pitcher or just a bad jump.

Sure. But unless these things are stable from attempt to attempt for a given pitcher, catcher, runner ... where does it lead us. And teams seemed (or maybe claimed) to have pretty good ideas already of which pitchers are slow to the plate, which catchers are slow out of the crouch and of course which runners can get to 2B fast. There's apportioning of past value, there's predictive value and then there's predictive value above the predictive value of what you've already got.

I think all of this information is going to be valuable when there is enough data to parse it down and come up with a framework for "average".

I think a bit of this information is going to prove very valuable, some of it will prove to be of marginal value and some of it will be distracting noise that leads nowhere. Such is the nature of data collection.

I think all of this information will have a big impact on broadcasts, media, baseball blogs, etc. I await "every 6 inches of range at SS is worth 9 runs a year which is why Brendan Ryan is the highest paid player in the game" and "Sure, Bronson Arroyo's career 98 ERA- is low for an HoF pitcher but what those numbers don't account for are his intangibles like his .8 seconds to home plate and his 86.4 MPH throws to first base with an accuracy of 99.8% and that his breaking ball rotated at 2,500 rpm, more rpm than Clayton Kershaw, and he magically led the DBacks to the 2014 World Series title." :-)

I just don't want folks to get their hopes up. For now, just marvel at it the way you do a David Attenborough nature documentary.

   15. bjhanke Posted: May 22, 2014 at 08:04 AM (#4711325)
Walt - this is just a bit off topic, but it's something I've been trying to find out for months, because it sounds really useful. What's the difference between 98 ERA- and 98 ERA+? I know it has to be a different method, but what's the difference, and what is it supposed to gain you? I've read a lot of comments saying that ERA- works lot better than ERA+. I Just don't know why. You're going to know that, and you made the horrible mistake of using ERA- in a comment that I read.

Thanks in advance - Brock Hanke
   16. Walt Davis Posted: May 22, 2014 at 06:13 PM (#4711754)
Wrong guy to ask on that one Brock but here's what I think I know ...

ERA+ is measured as (lgERA/playerERA) after park adjustments. ERA- does them the other way around. So high ERA+ is good, low ERA- is good. So that's the first thing to keep straight.

This leads to "scaling" differences too.

lgERA 4.50; ERA1 = 4.00; ERA+ = 113; ERA- = 88
.... ERA2 = 2.00; ERA+ = 225; ERA- = 44

The ratios are the same but ERA- is more interpretable -- the 2.00 ERA guy gives up 44% of the runs relative to average, as opposed to, what, 125% better than average? At one point Sean went with 2 - ERA- so the 88 would become 112 but called it ERA+ anyway ... that lasted about 2 days cuz it confused the heck out of all of us.

Everything else we do is player/lgAVG so part of it is just consistency.

Mainly just keep straight that low ERA- is high ERA+ and that ERA+ may seem to show bigger differences -- i.e. a 50 ERA- is a 200 ERA+.

But given the horrors of OPS+, this always struck me as an odd thing to "fix."
   17. cardsfanboy Posted: May 22, 2014 at 08:33 PM (#4711832)
Walt - this is just a bit off topic, but it's something I've been trying to find out for months, because it sounds really useful. What's the difference between 98 ERA- and 98 ERA+? I know it has to be a different method, but what's the difference, and what is it supposed to gain you? I've read a lot of comments saying that ERA- works lot better than ERA+. I Just don't know why. You're going to know that, and you made the horrible mistake of using ERA- in a comment that I read.


I like era- because it's more intuitive scale. Add in that it works in the same direction of era and it just works better.


Look at Pedros big year vs Gibson big year by these two metrics

Pedro 291 era+, 35 era-..

Gibson 268 era+, 38 era-.
   18. Best Dressed Chicken in Town Posted: May 22, 2014 at 09:48 PM (#4711863)
I cringe thinking about Michael Kay getting his hands on this data. Even more meaningless numbers he doesn't understand tumbling out of his mouth nonstop.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Harveys Wallbangers
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogAL WILD CARD GAME 2014 OMNICHATTER
(1129 - 8:33am, Oct 01)
Last: Jim Kaat on a hot Gene Roof

NewsblogLinkedIn: 10 Sales Lessons From “The Captain”
(2 - 8:28am, Oct 01)
Last: Jeltzandini

NewsblogSpector: Stats incredible! Numbers from the 2014 MLB season will amaze you
(36 - 8:15am, Oct 01)
Last: Never Give an Inge (Dave)

NewsblogNL WILD CARD 2014 OMNICHATTER
(2 - 8:10am, Oct 01)
Last: DKDC

NewsblogOT: NFL/NHL thread
(8174 - 8:01am, Oct 01)
Last: Norcan

NewsblogWSJ: Playoff Hateability Index
(16 - 8:00am, Oct 01)
Last: Swoboda is freedom

NewsblogOT: Politics, September, 2014: ESPN honors Daily Worker sports editor Lester Rodney
(4083 - 7:59am, Oct 01)
Last: Bitter Mouse

NewsblogThe Calm-Before-The-Storm and Postseason Prediction OMNICHATTER, 2014
(111 - 7:14am, Oct 01)
Last: Jolly Old St. Nick Is A Jolly Old St. Crip

NewsblogThe Economist: The new market inefficiencies
(19 - 2:21am, Oct 01)
Last: David Concepcion de la Desviacion Estandar (Dan R)

Hall of MeritMost Meritorious Player: 2014 Discussion
(14 - 2:17am, Oct 01)
Last: bjhanke

Hall of MeritMost Meritorious Player: 1958 Ballot
(13 - 1:55am, Oct 01)
Last: neilsen

NewsblogBrown: Winners And Losers: MLB Attendance In 2014, Nearly 74 Million Through The Gate
(33 - 11:27pm, Sep 30)
Last: Bhaakon

NewsblogMLB’s Biggest Star Is 40 (And He Just Retired). That Could Be A Problem.
(76 - 11:27pm, Sep 30)
Last: Walt Davis

NewsblogESPN: Ron Gardenhire out after 13 Seasons with Twins
(42 - 10:49pm, Sep 30)
Last: The District Attorney

Hall of MeritMost Meritorious Player: 1959 Discussion
(6 - 10:35pm, Sep 30)
Last: MrC

Page rendered in 0.1676 seconds
52 querie(s) executed