[0bfus]

One thing to note is the reported mindset of the program. It doesn't really talk about their expectations around plausability of this, but rather that they seem to be targeting success through large numbers even at low probabilities. They might not care to anticipate/correct anomalous "course deviations" but rather write those off as a loss completely, hoping that just a few of many initially accelerated succeed.

[paulmd]

It would be impossible/unlikely to correct any course deviations anyway. First off, even if the beam is collimated to within a degree or something, a degree at even a million miles is a pretty big spot, and you'd probably not be able to hit any specific unit. Even if you collimated it more finely - good luck hitting the target.

The other problem is that the beam would ideally go straight between earth and Alpha Centauri, and the sail just rides the beam. The Earth does rotate around the sun but the parallax this provides (over 3-6 months) is pretty low relative to the distance travelled by the probes over that time. If you try to do selective course corrections (i.e. bump one side harder than the other) then you're going to need a super fine beam that is going to miss frequently and probably hit the target when you hit.

This is pretty much inherently a shotgun approach. You don't get a mid-course correction on the birdshot that misses its target. Or rather, it's like putting out balloons into the wind and letting them drift.