06-15-2004, 08:51 AM
Hi unrealshadow13,
First off, let me state that we did experiments with small sample size (50-50) to prove that 350 mf was better than 450. Yes, at such a small sample size, we really could prove it statistically. So, you truly need AT LEAST 1000 runs to be able to prove anything about diminishing drops. And mind you, even if you throw together a sweet little excel-sheet with an easy-to-fill-out form and the necessarry scripts to make it run, it'd still take at least 2 minutes to record every run.
Other issue: implementation of such a counter. It has to be stored on your character. That is a breeze, but now think about it: what would reset that counter? If you killed anything else than the given boss? This would be silly, you could simply kill a blood lord on the way down. Or killing another "counter"-ed monster? This would lead to allow players to do meph-pindle runs without "penalizing" them for the runs. IF blizzard truly wanted to screw around with that, they'd have to store 2^n counters (if n is the number of monsters whose kills they'd want to count), to be able to store the "kills on a given set of those monsters". Jeez. Half the character file would be "run counters". Highly unpractical.
Yes, none of this proves anything. We're dealing with people who created a spell that fires of dozens of separate shards, calculating the damge of each shard separatelly, so they COULD have done all this. It is just, well, rather unlikely. If, however, you're like some of those people who are unwilling to accept stochastical reasoning, go ahead, and spend the next 5k+ minutes running Meph. Just one last thing about statistics and sample sizes: at such a large sample size, the secondary error (that is: your mistake in recording what dropped) will be so large, that the data you collected during that week-long playing session will be all but worthless.
At least you'd have some cool items :P
Caaroid
ps: learn to tell apart "with" from "at".
First off, let me state that we did experiments with small sample size (50-50) to prove that 350 mf was better than 450. Yes, at such a small sample size, we really could prove it statistically. So, you truly need AT LEAST 1000 runs to be able to prove anything about diminishing drops. And mind you, even if you throw together a sweet little excel-sheet with an easy-to-fill-out form and the necessarry scripts to make it run, it'd still take at least 2 minutes to record every run.
Other issue: implementation of such a counter. It has to be stored on your character. That is a breeze, but now think about it: what would reset that counter? If you killed anything else than the given boss? This would be silly, you could simply kill a blood lord on the way down. Or killing another "counter"-ed monster? This would lead to allow players to do meph-pindle runs without "penalizing" them for the runs. IF blizzard truly wanted to screw around with that, they'd have to store 2^n counters (if n is the number of monsters whose kills they'd want to count), to be able to store the "kills on a given set of those monsters". Jeez. Half the character file would be "run counters". Highly unpractical.
Yes, none of this proves anything. We're dealing with people who created a spell that fires of dozens of separate shards, calculating the damge of each shard separatelly, so they COULD have done all this. It is just, well, rather unlikely. If, however, you're like some of those people who are unwilling to accept stochastical reasoning, go ahead, and spend the next 5k+ minutes running Meph. Just one last thing about statistics and sample sizes: at such a large sample size, the secondary error (that is: your mistake in recording what dropped) will be so large, that the data you collected during that week-long playing session will be all but worthless.
At least you'd have some cool items :P
Caaroid
ps: learn to tell apart "with" from "at".