Not confirmed yet, but something like this requires a lot to really confirm. Even assuming it brings the odds down to as low as 1/512 that's still very high. Do you know how long someone has to work at this to experimentally verify the odds? They actually need to have encounter a fair amount of shinies to prove it, and even then you could still claim they got really lucky. Because you can just get terrible data even if you take it correctly.
Though everyone anecdotal evidence seems to agree that it does work. And that's good enough for me.