THE BOOK--Playing The Percentages In Baseball

In 2006 I used Hit Tracker to analyze every home run hit in MLB, and the data is generated was very interesting and useful.  It also whetted the appetites of a lot of people to have detailed trajectory information on all batted balls, not just home runs.  I am committed to doing everything I can to make this a reality, but realistically, it is a huge jump to go from analyzing 5,386 regular season home runs to analyzing about 150,000 batted balls.  I’ve done a lot in the offseason to streamline the process, but still, I can’t do it myself.

Right now, I have two volunteers who are going to help me out, and between the three of us I am hoping to be able to “cover” the following:

- all home runs
- all non-homer fly balls landing “near” the fence
- all batted balls of any type taking place during Red Sox, Yankees and Cardinals games.

The first item provides another year of home run data, which will hopefully help persuade MLB teams to adopt HT as their home run measurement system.

The second item will provide what I think of as the “lucky/unlucky” metric, i.e. for all hitters, what is their net gain or loss of home runs due to atmospheric conditions, and also, whose homers are barely clearing the wall, and who is coming up just short. 

The third item is the “Holy Grail” of batted ball data: an exact trajectory for every batted ball, which should be great for hitting, pitching and fielding analysis.  The three teams are chosen because a) I’m a Red Sox fan, b) there are a lot of Yankees fans and I want to contribute to the ongoing saga of the Sox-Yanks rivalry, and c) the Cards are the premiere NL franchise and the reigning champs.  But, if I get some more capable volunteers, I may be able to expand beyond this.

So, here’s how a volunteer could help.  First of all, to become a volunteer, a person has to be willing and capable to accurately make “observations” of batted balls.  This involves having access to MLB video (via MLB.TV and/or MLB EI, Directv says hello), owning a stopwatch and some spare time.  Needless to say, it also helps if you are a baseball and sabermetric fanatic

A note about stopwatch accuracy: When most people think of stopwatch accuracy, they picture a 100 yard dash, with a starter gun and a bunch of guys lunging across a finish line tape.  That is pretty far from the situation with Hit Tracker in many ways. 

In a race, the starting gun is deliberately set off at an unexpected instant in time, to foil runners who want to anticipate it.  In baseball, you can see the pitcher wind up, and the ball leaving the pitcher’s hand and approaching the plate.  It is easy after a bit of practice to “nail” the bat/ball impact.

In a race, a runner is judged to finish when their torso crosses the line, but obviously this is a bit of a judgement call, and it can be tough to tell, at full speed, when a runner is executing a last-second lean.  In baseball, the ball can be seen as it descends, and often you can see either a fielder or a fan reach for the ball, essentially previewing the impact.  And since the ball is a sphere, there’s no need to guess when its flight is finished, it’s a clean event.

Finally, in a race you get one chance to time it correctly, and if you mess it up, tough luck.  In HT, everything is done off video, which means if you mess it up, you can rewind and do it again, with an even better chance to get it right, since you would be familiar with the angle and the approach of the ball to the impact point.  With a little pratice, a person can tell when they’ve “nailed” the timing and when they’ve botched it, at which point you do it over.  I’ve had no trouble timing many home runs exactly the same (to 2 digits) repeatedly, and when there is a difference, it is typically on the order of 2-3 hundredths of a second, which I’ve analyzed and satisfied myself to be insignificant.  It is not only possible to extremely accurate, but I think just about moderately stubborn person can make HT observations that are dead on.

To the volunteers I have provided an observation program (in Excel), and then checked them out on some homers to make sure they were doing an acceptable job of timing and spotting.  Closer to the start of the season I am going to provide each volunteer with some additional programs to pull weather and box score data, and to put the observation data in the right format for upload to the site.  Once the season starts we’ll figure out who’s going to cover which games, and hopefully we’ll all be able to take time off when we need it .

I would be delighted to have anyone who’s interested contact me about volunteering!  You don’t have to do any set amount of games; if you’re able to make solid observations, I’ll be happy for any contribution you can provide.

Many thanks to Tangotiger for letting me get the word out, and of course if anyone has suggestions, questions or comments about Hit Tracker, please send them to

Dave was kind enough to alert his passionate readers about Greg’s project:
http://ussmariner.com/2007/02/07/sexson-projection-sent-out/

All you bloggers out there: please spread the word, by linking here, or cut/pasting Greg’s entire post.

The Holy Grail, the convergence of performance analysis and scouting, is upon us.

Here is the link if someone wants to be a spotter. 

http://www.hittrackeronline.com/spotters.php