The author has been a vocal critic of missile defense systems for decades. While that doesn't negate his arguments, it suggests a potential motivation for his artificial definition of success.
Edit: As the comment below points out, the rocket's trajectory could be changed so as to increase casualties. While that's true, it won't do so on average for several reasons. A rocket is only targeted if the system believes it has a high probability of striking a populated area. Since the expected losses are high, average will be better. In fact, average should be much better because the distribution of population density is so skewed. A small percentage of land is in dense regions. In this particular case, there's the added advantage that the most populous cities are relatively far away from Gaza so early impact is better.