Posted by John Dewan on December 16, 2015
November 09, 2015
Two weeks ago, we posted a Stat of the Week highlighting the new Stolen Base Times section of The Bill James Handbook 2016 that is based on data collected at Baseball Info Solutions (BIS). That post featured the average stolen base times of players on the Royals and Mets who had at least six timed stolen base attempts of second base during the 2015 season. A few questions related to this data were raised on Twitter by a couple people, including the Managing Editor of Baseball America, J.J. Cooper. One of J.J.’s questions in particular mentioned the following:
We thought it appropriate to address this question so as to help clarify things for everyone.
First, we can all agree that Jarrod Dyson and Billy Hamilton are two of the fastest men in baseball, and BIS’s data bears that out. Of baserunners that met the criteria of having at least six timed attempts of second base in 2015, Dyson and Hamilton were second and third fastest at 3.48 and 3.49 seconds, respectively. Rico Noel of the Yankees finished first at 3.42 seconds. League average is 3.67 seconds. So the assumption that Dyson should be faster than league average and similar to Hamilton is a valid one.
That brings us to the apparent discrepancy between the magnitude of BIS’s timer data and what scouts have anecdotally reported. There are two likely sources for this discrepancy. The first is the start of the clock. BIS begins recording stolen base times from when the pitcher makes his first move toward home plate. Scouts, however, typically start recording their stolen base times from when the runner makes his first move toward second. The lag between when the pitcher starts his move to the plate and when the runner first breaks for second explains most of the reason for BIS’s times looking slower than a scout’s times. The scout starts his clock later, so his recorded time will look faster.
The second possible reason for the discrepancy is related to the method by which the data is collected. Scouts using a stopwatch have been shown to have a split second lag between when they see a player begin his move and when they actually press the start button on the stopwatch. However, that same lag does not exist on the back end of the stolen base attempt because at that point the scout can anticipate the runner hitting the bag. BIS, on the other hand, breaks down the video using frame-by-frame software to minimize human error in recording the time. So, again, because of the lag, a scout’s stopwatch time is likely to look faster than a time recorded via frame-by-frame.
The times that J.J. Cooper referenced are not necessarily inconsistent with the times that BIS has collected. They are just measured on a different scale and with different technology. BIS goes to great lengths to ensure the quality and accuracy of our data, and we hope that providing this explanation helps to clarify our data collection methodology for everyone.