Originally Posted by Rojo
The standard sabremetric view is that stolen bases are largely no more than a nice add-on, given a high success rate. But are exceptions allowed for Coleman/Hamilton speed? Yes, a .310 OBP is too low for a 35sb player or even a 40sb or 50sb player. But a 110sb player?
I guess I always take that comment with a grain of salt.
Sure, it's better to have a 340 obp than a 310 obp. No one is arguing that.
But what's the impact? Over 600 plate appearances, that's getting on base 18 more times.. I've brought this up with Stubbs sometimes. Speed and power can compensate for a lower obp in Stubbs case (who I think has pretty much averaged 90 runs scored per season).. Billy obviously won't have Stubbs power, but the steals will put him in scoring position pretty often when he does reach base.
It's hard to quantify.. If Billy gets into scoring position X% more often than the average player (which causes him to score more frequently).. at some point, the speed makes up for lower OBP.. I can't really figure out what X% is.. Obviously caught stealings factor in also. But it seems that at some point, enough steals will offset getting OBP 18 fewer times in this example.
The standard argument about needing to be successful stealing 70% of the time is also flawed in my opinion. That's based on run expectancy numbers.
I have done the math at one time.. but the problem with that exercise is that the run expectancy for a guy at 1b also includes the benefit of stolen bases, bunts, productive outs and other "Smallball" activities which advance the runner to 2b. (It also includes the negative aspects like caught stealing).
That's a flaw when calculating the result using run expectancy, which is essentially a markov chain.. Maybe I am wrong here (someone please correct me).. I have not had the time to dig deeper into this.