Arguably, the strong-field approximation (SFA) has been the most successful analytical approach to strong-field phenomena. For photodetachment of negative ions it is practically exact. For above-threshold ionization (ATI) of atoms, it predicts most observed features at least qualitatively. Actually, the term SFA refers to the first two terms of a Born-series expansion of the ionization amplitude with respect to the binding potential. The first term describes the direct electrons, which do not feel the potential any more after they have been liberated. The second term incorporates one single such interaction. A stationary-phase evaluation of the amplitude allows one to interpret ATI in terms of classical electron orbits in the laser field, with zero, one, or more scattering events if the electron trajectory revisits the position of its parent ion. The first-order Born term allows for one act of rescattering, which, however, does not have to be at the first revisit. Forward rescattering makes a very important contribution yielding, e.g., the low-energy structure. The significance of terms of higher order is an open question. For example, several acts of forward scattering may precede the final off-axis rescattering There is preliminary evidence that, for a Coulomb potential, the Born series does not converge. Healing this deficiency may require some form of renormalization where forward rescattering is included to all orders.