Again, I think this article is flawed, unless I am misunderstanding run expectancy:
With no outs and a runner on first in a five run/game environment, the average team is expected to score 0.950 runs. Moving the runner to second base on a steal yields a run expectancy of 1.192 runs, meaning that the change in run expectancy from that steal was 0.242 runs. The same move with two outs, a move often lauded because the runner is placed in scoring position for the current batter, only yields 0.09 runs on average. In total, in this run-scoring environment, the average stolen base was worth 0.175 runs. A worthwhile move if successful, but hardly worth making a fuss about teams emphasizing.
The run expectancy of .950 of a runner on first includes all the SB and small ball scenerios to move him to 2b.
Don't they get run expectancy by averaging the number of runs that actually scored in a given scenerio?
Example.. Stubbs leads off and singles. Steals 2b. Steals 3b. Scores on a sac fly.
He started the inning at 1b, so that run gets credited to the run expectancy of a runner on 1b, no outs.
I think in order to accurately calculate this, you'd have to average out "station to station" run expectancy with a runner at 1b.
Then calculate run expectancy when the runner is at 1b and attempts to steal 2nd.
Both these run expectancy numbers might be different than the values expressed in the article.. A runner that just stole 2nd might score more frequently than a slow runner that just got a double (or maybe not).
I just think the saber communities analysis to get this 68%-74% number is not complete. If they wish to make a bold statement such as "speed is overrated", as the author does, they need more precise data collection, not a sloppy, back of the envelope calculation.