While it is true that it should be easy to deflect, it is only easy to deflect if you can detect it. The essay assumes that the projectiles would be pushed to a significant fraction of the speed of light. At that speed there is no time to launch any type of deflection.
The energy required to accelerate a big object to even half the speed of light is prohibitive. Even if the process was feasible, it would be a vast cosmic event bound to emit some radiation, which would still arrive much sooner than the object itself.
Well you don't have to accelerate it all at once. Remember, this is space war, and we're playing for keeps. Spending a millennium to accelerate a rock up to .5c might be worth the trouble.
If you take longer to accelerate it, that just increases the probability of detection by other means. It still doesn't change the argument that it should be much cheaper for the defender to detect and deflect your projectile than it was for you to launch it in the first place. Not to mention that an unmapped interstellar gas cloud, or an undetectable asteroid, is all it takes to deflect the projectile a few millionths of an arcsecond and make it miss its target.
