Greenmonsters
Wannabe Duck Boat Owner
This is a bit alarming. I learned this in a graduate statistical analysis courses and then also use this principle throughout my long engineering career. I deal with this principle with co-workers and venders and customers alike (fairly standard stats stuff). I would suggest you might want to look into it. It is CRITICAL in making assumptions based on a data set.
Now a person COULD assume that you have a valid data set based on the number of overall pitches. I would disagree. In an extreme example one could analyze the flight off a single pitching machine up to a pitch count of a million. Would this mean you can make exacting conclusions on the flight of a softball?!?!? Absolutely not. I hope you can see this problem. You need to have pitches from a varying selection of pitchers that needs to be greater than or equal to 30 pitchers to have statistical significance to allow confidence in your data assumptions.
A MINIMAL number of pitchers should be 30. (and this data set of pitchers should be hand selected accomplished pitchers that have at least some indication from input from catchers/coaches/etc that "wow,...this pitcher has that late-break characteristic".) I submit that it IS that rare and it is directly attestable to an exacting speed/spin/angle on the pitch. I would be willing to bet large sums that there is a dynamic, like drag crisis, this IS occurring on some pitchers pitches.
Are we talking about the same MIT? The Myopic Institute of Technology's study I read found that Sir Isaac Newton got hit on the head by an apple was because he swung and missed due to late break.