THE BOOK--Playing The Percentages In Baseball

Give him a camera, too.  Have him take a picture of the field before every pitch.  Later, for the 60 or so batted balls per game he can log those positions into the system.

Bonus: on a grass field, (with some work) you’ll be able to mark the initial positions and landing spots of batted balls very accurately because the pictures will have the mowing pattern of the field.  Before the game, take a picture of the field from up high, and superimpose a direction and distance grid on it.  Later, when a guy rips a line drive to the gap, locate the landing point on the grid formed by the mowing pattern, and read off the “precise” landing spot.  Should get accuracy within a few feet.

All this, of course, if people really want precision.  It’s time consuming work, but should keep people busy until the next gen. technology gets here…

The camera idea sounds like it would be pretty easy to do for fielder positioning, but might be a little trickier for every batted ball due to timing issues and resolution issues (I’m ignorant about how high a resolution you can get out of a modern camera).

Do they need stopwatches? Can’t they just get the time of most of the events off a video recording of the game? The booth may occasionally miss a shot of a play, but you can generally tell when a flyball lands or when a grounder gets past the infield.

Well, whether they need a stopwatch or can get it from the video feed, the point is: RECORD the time.  Such a tiny and simple thing to do.

I like the picture idea too.  Again, so easy to do.

BIS, STATS, etc., are in the business of recording data in order to sell it to people who want to do research.  It is amazing how resistant they can be to easily changing for the better.  Eventually they will, but it is like pulling teeth.  People just don’t like to change their routine, even when it is fairly easy and when they know it is for the better.

If nothing else, these data collection companies, including the RS guys, should be telling us if there is a shift, if the infield is playing in, the outfield is playing up (we can actually infer that from the score, outs, and inning), or the corner or corners are in expecting a bunt.  It would make A LOT of the research, especially the defensive stuff, much more accurate.

And you would think with so much competition for it, that the market would force the data providers to do this.

But, then you look at the NHL, which collects the data itself, and it is leaps and bounds in terms of effort they put in, compared to what the baseball data providers do.

The NHL will record the ice time, by the second for all 12 guys on the ice (though, easy to do with the goalies).  They will tell you the number of feet a shot was taken, whether there was a takeaway, giveaway, hit, who drew the penalty, who won the faceoff, and on and on.  Basically, the things that you would try to describe verbally, they are capturing it.  There is little that they are not capturing.  And, hockey does not have the stoppages in play that baseball has.

In baseball, as mgl said, you want to know if the IF is playing in or not, if the 3B is covering the line, if the CF is playing left, right, or straight away, if there’s a shift, etc.  But, for whatever reason, the data collectors don’t do that!  There’s a ton they don’t record, which is in contrast with the NHL, which tries to record as much as it can.

Very weird.

Is there a chance individual teams are doing this type of data collection already?  Or is to too costly/labor intensive?

No way are they doing things systematically like we proposing.  I do know some do some things, but more as a side project (like they’ll make the observation if it was a good fielding play or not), and not integrated with the PBP files.

Well, maybe the Redsox are.  They do have Tom Tippett.

Isn’t there an overhead video camera surveying the whole field for pretty much every game in every stadium?  Maybe not the *whole* field ... but even in the incredibly rare event that a game is not televised somewhere, there’s still the closed circuit cameras in the stadium.

... incidentally, glad to see the green back.

On European Champions League soccer games this year, they have started showing how far players have run during the game.  For example, they’ll put up a stat showing “Fabregas 10,450 meters” (I think it’s in meters, it could be feet, but it is in Europe so probably meters).  Every time I see it, it has astounded me that a) the players run so far (several miles during a game), and b) that they are somehow able to track every player and see how far they have run.

I don’t know what the technology is, or anything about it, but if they can do it for a whole soccer game - and it’s presumably automated in some way - they can do it for positioning in baseball.

That technology is being done by a French company called Sport Universal Process.

http://www.football-business.net/tracking-device.html

They use eight fixed videos cameras to cover the pitch.  Technologically, I don’t think there’s any reason Sportvision couldn’t do this, too, or MLB could use Sport Universal Process, but I wonder if the cost would be a barrier that would prevent this from being implemented in MLB.

Yowza.

Does the French “Ligue 1” generate over 6 billion$ in revenue?  I don’t think cost is a barrier.

Here’s the company:
http://amisco.eu/

And here’s another link:
http://mastercoach.de/index.php?option=com_content&task=view&id=43&Itemid=4〈=en

Are you kidding me?  This is unbelievable!  This is what we’ve been waiting for, and it’s been here at least 2 years?

Tango, I meant cost vs. payback, not cost relative to revenue.  How is spending money to develop and implement this capability going to increase revenue for Sportvision and MLBAM and the broadcasters.  I believe it could, but I can see various parties in that equations considering it a big risk at this point.

In soccer, tracking the movement of the players gives huge insight into strategy of the game.  In baseball, most of the action happens around the pitch, which is probably why we have information on pitch trajectories before anything else.

This technology is by no means limited to champions league and the biggest national soccer leagues.

We have it here in Sweden as well and our soccer league is 3rd rate at best. (Think of it as high A-ball of european soccer.)

Mike/15: One marginal win is worth two or four million dollars.  The payback will be there, as it is for Bjorn’s 3rd rate league.

Tango - I am all for having more data, but just out of curiosity, what question(s) will you be able to solve if you finally get the information of where the players are positioning themselves?

1. You can split up the player’s fielding between his range and his positioning.  It’s alot easier to work on someone’s positioning than his range.

2. Compare a hitter’s spray chart against the defense he sees, and determine if he is changing his approach based on the fielders he sees.  If so, you can try to figure out an optimal setting against each hitter.  Spefically, for those shifts against Ortiz, Hafner, et al.

3. Is a pitcher allowing alot of hits because his fielders simply aren’t positioned optimally for him?  Look at his spray chart, and where his fielders are.

***

Ideally, you’d create a model where the hitter’s spray chart, the pitcher’s spray chart and the fielder’s positioning have an inter-relationship, and you can try to come up with the ideal spot for all fielders for each combination of hitters and pitchers.

Tango/#17, marginal wins don’t help Sportvision or MLBAM or the broadcasters, and those are the parties that would need to spend money to implement the system.

If you are arguing that the clubs should want the data enough to invest money to get it, I would agree, but I see no sign the clubs actually think that way.  Is there any single club, let alone multiple clubs, that is investing millions into analysis right now in order gain a few marginal wins?

Will it probably happen eventually?  Yes, because the technology will finally get cheap enough or MLBAM will find a way to sell the data.  But unless the clubs change their investment model, I don’t see anyone having a monetary motive to implement it quickly.

Tango - I am sure that the teams use the data in some form in their analysis. But at least in Sweden (and I am relativly certain that this is the general case in european soccer) the economic driver for these system(s) is that the data can be sold to the fans in various “premium packages”.

Soccer fans have a pretty high acceptance to pay for content. Most of the big games are already televised on subscription based channels that are often relativly expensive.

Not saying that the same case would be true for US baseball, but I do agree with you that there is probably enough economic incentive to go around.