That's the number I always have in mind: 200. I'm not sure how I settled on that number but it's in my head now and I have found it to be a nice, conservative, effective guideline.
I can start to be influenced before 200 samples. Specifically if I'm researching a trend and getting 50-50 results after 100 samples, I'll probably just scrap that research and move on to something else.
But generally I wait until I have a compelling result after a good 200 samples before I start backing a trend with cash. (And I'll still continue to watch the results very closely).
|