It's rather simple to explain with a banal hypothesis: that the epidemic began just slightly earlier than initially thought. It just failed to take hold, maybe by pure chance, or because the virus didn't infect a superspreader, or because the initial strains had not yet acquired a tiny mutation that would have made them infectious enough to propagate further.

It's just a question of probabilities. There will be more people infected near the origin of the disease before a pandemic is detected. Some of them will travel and infects others, starting new foyers, but they will all be smaller than the starting location. The pandemic is therefore more likely to be detected at the starting point, because more people will have been affected right there.

If your hypothesis is based on an exclusive, direct trace, how comes that you may still maintain it, if you have to allow for an unaccounted for and untraced span of time and spread, including an untraced roundtrip around the globe? How do you still establish the link between the lab and a location as specific as the wet market? This is a matter of basic science and hypothesis checking. If you propose a direct link, any unrelated, conflicting data point invalidates your hypothesis. (Incidentally, Lombardy, where most of the Italian samples came from, was also one of the regions hit hardest by the Alpha variant.)

P.S.: Mind that the allegations are not about a loose probability, but about a very specific scenario, involving the lab and the outbreak at the wet market at a certain date.