and infield hits/drag bunts.

Plus the very threat of that makes the infielders play closer in, helping the speedy player get a few extra hits through the infield. Tough to quantify, though, even if you had good positioning data.

a high career BAPIP would be a good proxy to quantify the whole effect, right?

It gets messy though. There are so many factors that go into a high BABIP. Like hitting the ball hard. Jack Cust has a .337. Then you've got the very skilled hitters, like Boggs, Carew, and Gwynn. 2 of the 3 were fast when they were younger, but even as old players (and Boggs his whole career) they had high BABIP.

What is the typical Fenway boost to BAPIPs?

The specificity of drag bunts is what I don't get. Drag bunts are about placement as much as speed. Beating out a bunt to 3B is speed (or a deep 3B) but if you can get the ball in the triangle between P, 1B, 2B then pretty much any non-Molina should be able to beat it out.

But no matter.

As the article sort of notes, the combo of top defender and top baserunner is maybe not as common as we might think. Certainly a lot of the game's great base-stealers were pretty average defenders -- Brock, Henderson, Raines, Coleman. Even Ichiro had only 3 seasons of 8+ runs of base-running (b-r) and only one season above 1 dWAR. Of course I guess Ichiro's infield hits are folded into Rbat so he adds speed value there but the article doesn't seem to separate that speed impact. So, for his career, Ichiro has added about 9 wins via his combo of baserunning and defense or less than one win per year. (I don't think he's including Rdp here ... if he is, Ichiro adds another 5 wins!).

The article's premise isn't really about trying to measure speed's impact, it's "OK, Billy Hamilton is essentially pure speed, can he stick in the majors." By the end he decides that, if he can match Bourn's roughly 2 wins a year by speed, then Hamilton only needs to hit like Brendan Ryan to be an average player. I've got my doubts about that but sure, it's theoretically possible.

I thought Tony Gonzalez was sometimes lazily called "Speedy" and he did have 12 triples in 1963 but his SB numbers were not Billy Hamilton in any way.

I’m mostly certain when I suggest that speed produces runs in three main ways: by means of defensive range, baserunning (both via the stolen base and other manner of advancement), and infield hits/drag bunts.

What about avoiding double plays?

It's hard to sort out double plays because there's too much noise in the data. Fly ball hitters don't hit into many DPs, but they're often big, slow sluggers. Speedy guys who hit lots of ground balls sometimes do. And, of course, batting order matters (opportunities).

#3, exactly, Edgar Martinez had a career BABIP of .335. He could clear the bases on a single, which could be a rocket off the wall. BABIP probably has more to with the ability to hit the ball hard on a line than anything, unless you're Ichiro and can outrun a ball 3 feet in front of the catcher :)