I think this has to be evaluated for each propellant combination and situation individually. I was involved in a horizontal ground test program at AeroAstro in the mid 90's for a LOX/RP-1 (kerosene) engine that was designed to be LOX lead to start. After a very short burn was computer aborted due to an accidental trigger of the "high thrust" alarm, the next burn ended up hard starting the engine because there was some unexpected fuel pooled in the chamber from the abort. It turns out that the pooled Kerosene, when chilled to cryogenic temperatures, can become explosive in the presence of LOX. This was something most of had not heard of except for some of the real old guys who had worked with RP-1 for a long time. There's an old NASA paper describing some detonation testing that had been done for RP-1/LOX here: NASA Detonation Research Thus when the engine was restarted, the pool cooled when the LOX from the next start hit it (which according to our gray beard can cause RP-1 to form a gel with LOX) and then our igniter lit and the gelled mix detonated. But we were using LOX to regeneratively cool the faceplate of the injector so we wanted the pre-chill that the LOX lead gives, so we kept this timing and changed our inspection criteria between runs. BTW, our "High Thrust" alarm was triggered because we initially set this very conservatively at just 20% over MEOP. We ended up having some small combustion instabilities which weren't completely eliminated with the first revision resonator cavities we drilled, so we saw ~ 30% pressure spikes above MEOP. The system was designed to take 3x MEOP for this early test, so we had room to move our trigger to 50% and that's what we did for the second run.
For a NTO/Hydrazine engine we recently tested at Stellar Exploration, we elected to run with an NTO lead because spilled hydrazine is harder to clean up/decontaminate than NTO, which readily evaporates and is dispersed.
At a third organization, we were testing a LOX/methane engine and had elected to use LOX lead for the ignition to avoid building up potentially explosive methane gas in the test cell. That engine still found a way to blow up, it just wasn't due to methane buildup in the test cell.