few years ago I developed a formula for stolen bases I viewed as superior to a simple tally or a percentage. I decided that
because a player must score in two outs, or before the third out is made and because that player needs to advance four bases
to score, that a stolen base was only half as valuable as an out made through a failed steal attempt because the player advances
¼ of the bases he needs with a stolen base but makes ½ of the outs he can expend when he is caught. And ½ is twice as much
as ¼. So I invented what I thought at the time was a unique statistic that I called SBR for "stolen base rating." It simply
calculated stolen bases – 2 x caught stealing.
once I started reading sabermetrics, I discovered a more advanced statistic of basically the same type created by Total Baseball. They also called their statistic SBR, for "stolen base runs." But where theirs outshined mine
was in that it actually calculates how many runs a player produces by virtue of his stolen base attempts. Total Baseball figured
that a stolen base contributes 0.3 runs to a team's effort and each time caught stealing costs the team twice as many potential
runs, or 0.6. This statistic has been criticized by some in the vanguard of statistical research but as far as I can see still
holds value. Read “Stolen bases” under “Misunderstood Conceptions” in the myths section of this product
for other explanation.
= (0.3 x stolen bases) – (0.6 x failed steal attempts)
Long ago Bill James introduced runs created as perhaps the most valuable measurer of a player’s ability to hit
for power and to get on base. It has long been respected. ARC is simply “adjusted runs created.” It is fairly
simple and combines two successful, if occasionally maligned, statistics – runs created and stolen base runs - to provide a cumulative offensive evaluation of a player, of hitting for bases, power and base running.
ARC = Runs created + Stolen base runs
= (On base percentage x total bases) + (((0.3 x stolen bases) – (0.6 x failed steal attempts))
OPS (On base percentage plus slugging percentage) was once hailed as the great new statistic. It has recently been
dropped on the scale of respectability by statisticians. It is too simplified and inaccurate for their tastes. But a little
research shows that for all its flaws, there is still a good deal of value in OPS. While it does not give enough emphasis
to on base percentage, they say, it still manages to have a correlation coefficient of about .96 to .98 with Bill James’
statistic, runs created. That value is very high and means that but for a very small deviation in select cases (that should
not affect the totals very much) it falls right on par with runs created. And it is easier to calculate. So why not use it?