In my description of linear weights two Staturdays ago, I made an error. Tom Tango of The Book Blog pointed out that I had the wrong run values for stolen bases and caught stealing. Here is the corrected version:
Event Run Value
Now, some might say that steal attempts are leveraged - that is, assigning a run value to steals is misleading, because you attempt steals when you have a chance to impact your probability of winning, not just your probability of scoring. In that case, you can do a few mathematical shenanigans and come to the conclusion that the 0.175 leveraged runs you get from a typical steal is more like .20 "random" runs. You can do the same thing with the caught stealing value and substitute - 0.49 for the -0.467.
Thanks to Tom for pointing out the error.
Also: LINK DUMP!