The nature of the excited-state double proton transfer in 7-azaindole (7AI) dimer-whether it is stepwise or concerted-has been under a fierce debate for two decades. Based on high-level computational simulations of static and dynamic properties, we show that much of the earlier discussions was induced by inappropriate theoretical modelling, which led to biased conclusions towards one or other mechanism. A proper topographical description of the excited-state potential energy surface of 7AI dimer in the gas phase clearly reveals that the stepwise mechanism is not accessible due to kinetic and thermodynamic reasons. Single proton transfer can occur, but when it does, an energy barrier blocks the transfer of the second proton and the dimer relaxes through internal conversion. Double proton transfer takes place exclusively by an asynchronous concerted mechanism. This case-study illustrates how computational simulations may lead to unphysical interpretation of experimental results.