Samples of hundred drops or thousands drops says absolutely nothing.
My caveat with that is definitely true
if people are talking about just raw numbers. A couple hundred or thousand drops can be more than enough for statistical tests though that I mention above or some of the testing Leeloo, myself, and others have done in my signature as well as just above this. A good example of this what
excavator enhancer testing where 70 pulls was enough to detect a statistical difference, you didn't need 10,000+ pulls. Hit rate is different than TT per pull obviously. If for instance I wanted to formally detect if the hit rate between two "runs" decreased from 32 to 28% in a statistical test, I'd need
a sample size of 800 claims. Bump that down to 30 vs 28% and you'd at least 3160 samples to detect that difference.
I mention that because this problem is exactly something we go over when teaching intro statistical analysis courses. If you try to just get a massive sample size without any measure of variation, it's still a largely meaningless stat that you can maybe spitball with a little. If you do try to formally analyze a huge sample size, you also run into issues where you might detect a slight difference as "significant" that isn't really relevant. Plus, it's time intensive and expensive. Instead you want a representative sample size where sometimes 30-100 samples is plenty, other times you need thousands. On the opposite end of the spectrum, you can get people with a tiny sample size with no statistical measures claiming this coin must be rigged because they flipped it four times and got heads each time.
Usually in classes we'll be navigating that very subject with students and how you reach a minimum required sample size needed for statistical tests to be accurate representations of what the "true" average in a population. Sometimes in these discussions on PCF though, we'll get people just sticking to a mantra of saying you can't say anything unless you have over 100k samples. That completely ignores what's being taught on basically day 1 statistics. That's not to say you or anyone else here is falling into that pitfall, but that just seemed like a good cue to remind people about how sample size is navigated.
