When we're taught MO theory in basic chemistry, a rule-of-thumb hammered into our heads is that an orbital takes up to 2 electrons with opposite spins. Sigma or pi bonds are represented by stable bonding orbitals. Orbitals sometimes have just the one electron (radicals) but are much more stable with the pair.
While this justifies the occurrence of one-electron bonds, like that of the dihydrogen cation, it does not smoothly explain the occurrence of three-electron bonds like in nitric oxide and dioxygen. If we draw a straightforward MO diagram for nitric oxide with the "two electrons per orbital" rule, we end up with a lone electron in a antibonding pi orbital. "Antibonding" seems to imply the lone electron is sitting at its host atom, but that's not the case. Where does this lone electron go? Does it "merge" with one of the bonding sigma or pi orbitals and if so how do we know which one? Is this just a shortcoming of MO theory?