Most stars are born in binary and multiple systems. Nearly all very close binaries with orbital periods P < 5 days are in triples, initially suggesting that the tertiary companion played a role in the dynamical hardening of the inner binary. However, several properties of these hierarchical triples are inconsistent with dynamical models. I will share how recent observations coupled to population synthesis models resolve this apparent discrepancy. The second mystery involves binaries across intermediate separations a = 10 – 100 AU. Early adaptive optics surveys discovered a substantial excess of these binaries in low-density star-forming regions like Taurus and Upper Scorpius compared to the galactic field population. Field stars represent an ensemble average of different star-forming environments, and thus it was predicted that dense star-forming regions must preferentially disrupt intermediate-period binaries in order to compensate for the observed excess in low-density environments. However, recent observations have shown that even the dense Orion Nebula Cluster manifests a similar or even larger binary fraction across intermediate separations compared to the field. I will demonstrate how an overlooked selection bias fundamentally alters the inferred primordial binary period distribution, yielding significant implications for star formation and the initial mass function.