@lucisferre, you're 100% correct here. I certainly don't ignore 300-600ms differences but I only look for them after I've exhausted other, perhaps larger, gains

@lucisferre, you're 100% correct here. I certainly don't ignore 300-600ms differences but I only look for them after I've exhausted other, perhaps larger, gains. As for how many times the test was run, it was randomized but but somewhere between ~60-100 times.