In pairwise accretion -- that is, the formation of bigger planets hierarchically from smaller ones -- the typical encounter speeds are comparable to the escape velocities of the dominant bodies, and the colliding bodies are of comparable size. This is a far cry from the two most commonly considered regimes: hypervelocity bullets hitting a much larger target (a.k.a. impact cratering), and the perfect mergers (sometimes with a disruption threshold) assumed in nearly every N-body simulation of planetesimal/planetary growth. The best studied case is the Moon's formation by a giant impact. Direct modeling of pairwise accretion -- which is not always accretionary -- has led to several new models that will be presented: (1) the origin of chondrules by relatively slow collisions between similar-sized molten planetesimals (Asphaug et al. 2011), (2) the origin of the lunar farside highlands by a low-velocity 'big splat' (Jutzi and Asphaug 2011), and (3) the stripping of mantles from asteroids and planets (for instance Mercury) when they almost but not quite accrete with a larger body (Asphaug 2010).