a) The very first article I read emphasized that overspeed should not be a drastic jump, but rather a slight increase–to avoid screwing up the runner's mechanics. We always kept our overspeed in the range of .5sec/100 or so–about as much as if the runner had a steady wind at her/his back (except you can't count on steady winds on a regular basis!).
b) I misspoke–good catch. I meant strength development *throughout* the range of motion, meaning particularly at the limits. The middle part of the ROM is not typically an injury problem area, but the development of flexibility–extending the ROM–without developing strength concurrently is a recipe for injury. PNF develops flexibility and strength at the ends of the ROM. (However, I don't want to get sidetracked into a PNF thread.)
c) The arms have been shown to (very) slightly lead the legs in sports movement. There is a neurological correspondence between elbows and knees–it's not simply a question of physical or skeletal distance between the two joints. "[A] shorter lever just means [it's] easier to run"–well, yeah. In particular, it's easier to run at a higher level of steps/sec.
d) "[Y]ou have [fewer] injuries thru sheer luck or undertraining"–or through optimal, many-sided training, perhaps? Going through ten years without any hamstring strains suggest that it might not be a question of sheer luck–although, statistically, that is possible. Fielding conference champs/all-Americans in the sprints on a fairly consistent basis suggests that undertraining may not have been a major factor, either–though I could be wrong there, too.
"[T]he objections to overspeed or assisted running are that it is impossible to increase leg turnover or stride rate while being pulled!!!! You have to increase stride length to maintain the pace." If I may inject some data into the discussion: we actually did tests on sprinters who were being pulled (vis-a-vis their top unaided speed), and they emphatically *did* increase stride rate while being pulled! This may have been because they were coached to emphasize that increase in rhythm while being towed. Their unaided-top-speed stride rates were also increased somewhat, both immediately following towed reps and in the long term (course of a season, and from year to year).
In any case, I agree with your final statement: "Why not stick to 1K or 800m repeats at faster than race pace for 2-3 [times?] the distance? It's much more specific." It's definitely more specific! I don't think I would make overspeed a *major* part of a distance runner's training; I guess my question was just: Has anybody had actual experience with giving distance runners assisted work, either i) for top-speed development, or ii) longer reps like the miles/two-miles mentioned above?
And by the way, thanks for the response, danimal9!