stats question

[Birdjaguar]

The real question is: did he get the drops?

 

[Hygro]

Ok so I fixed it. I legit was multiplying by my iterations instead of powering
I'm embarrassed.

The multiplication give me this linear result that caused me to disbelieve it, I guess that part was good, but I disbelieved the wrong part

With a proper exponential curve it all makes sense.
50% chance is 317 hours.
99.9% chance is 3166 hours.
To get ruby to round the results to 1 takes 17,155 hours.

Mental crisis averted.

Special thanks to everyone who participated, called me out, did your own calculations, kept it real. Bonus to @AdrienIer who saved me from an extra dumb mistake of lazily willing to conflate a REALLY SMALL NUMBER with convergence, just because my computer would choose to do that.


You, too, would wonder what you didn't know about math if something had 50% odds at half the time to 100%, instead of 75% at that time.

 

[Arwon]

A confounding factor here may be that computers are only pseudo random, so the theoretical odds may not be the actual odds?

 

[uppi]

A confounding factor here may be that computers are only pseudo random, so the theoretical odds may not be the actual odds?

Depends on how much dependency the PRNG injects between drops. If they had used a cryptographically secure PRNG, the odds should be match almost exactly the theory. I very much doubt that, however. Given the age of the game, it could be anywhere between "good enough, doesn't matter" and "homebrew with serious flaws, so you can easily calculate in advance". Though, if it was the latter, I suspect these people would know by now.

In absence of any detailed knowledge about the PRNG involvement, I would assume that it works as intended and can be assumed to be ideal.

 

[Hygro]

A confounding factor here may be that computers are only pseudo random, so the theoretical odds may not be the actual odds?

The RNG always felt a little "sticky" in Diablo 2, although I always chalked that up to how random shouldn't actually look like noise and will look like these almost-patterns. I wasn't worried about that in making a prediction. The weakest links are in the drops-per-boss and bosses-per-hour.

 

[EgonSpengler]

Oh I see, you mean you can't divide up the time halfway as part of some progress bar.

That's a good way of putting it, yes.