Andor's Trail

fiernaq wrote:Number of kills can be found within the save files. Otherwise, just count the measure item found in the Extraordinary and legendary items thread. In the case of the RoLs that would be Small Rocks. That's why you'll see in my signature the abbreviation "SR" for Small Rocks. Do note, however, that the gargoyles can also drop small rocks (though at a much lower drop rate) so in order to be accurate, you would have to have 2 piles of small rocks on the ground... one pile for rocks picked up while killing gargoyles and one pile for rocks picked up while killing the trainers that also drop the RoLs. At this point I have about 290 small rocks from the trainers and about 50 small rocks from the gargoyles which don't count towards the measure. Apparently, I need about 2210 more small rocks to have killed 10,000 of the trainers

...happy hunting, I wish luck for you.

I don't know if this has been addressed here, but to me, the drop count is disturbingly misleading and severely disingenuous.

The fact of the matter that as the number of monsters you kill approaches infinity, the probability that a rare item will drop approaches one is true. But it has been suggested several times here that the graph made to illustrate this point is linear; that is that the drop rate is additive. It's not. The rate is multiplicative.

What this means is that the chances of getting a legendary item in ten thousand kills is still less than two in three (~63.214%). In fact, you can't expect a legendary drop with a scientific level of certitude (five sigma) in even one hundred thousand kills (although for all practical purposes, ~99.995% should be enough for anyone).

If each monster has a 1/10,000 chance to drop an item you still have precisely 1/2 chances to get the item per kill. Either it drops or it doesn't. Statistics has been proven time and time again to be an extremely faulty logic that is only given a semblance of realism on disproportionately large groups of numbers. On smaller groups it doesn't even try to be accurate, merely claiming to give you a decent guess as to what the outcome will be. So yes, the whole "1/10,000 drop chance" thing is entirely misleading because by the rules of randomness you have just as much chance of never getting the item to drop as you do of having it drop every single time. If this were not true then you would not be using the true rules of randomness but merely an artificial mostly random set. In other words, you could kill a single monster and have the item drop or you could kill your 9 gazillionth monster and still not have it drop... both are equally valid options for you. That, for the group of AT players as a whole, the drop rate evens out towards 10k is purely based on multiple people having their own random number within an unterminated range, which, by the way, is the key. There is no termination on the range. You could get unlucky and not have it ever drop even after 900 billion kills.

.... I don't have the slightest clue what you meant by that first sentence. The drop chance is one in ten thousand. The chance the item will be dropped is not one in two. How did you manage that leap of logical discontinuity? I'll grant what I think you may have meant by that is that your options are binary in nature - either you will obtain the drop or you will not, but so what? That does absolutely nothing to alter the validity of the mathematics in question.

Statistics are not logic, nor a form of logic. That aside, it has not ever been proven to be faulty at predicting the outcomes of any random occurrence in which all factors of the occurrence are known - precisely the opposite; it is the most relied upon tool that exists to make predictions of any sort, even in situations in which it is not designed to handle, simply because of the ease and consistency with which you get accurate results. To say that, "Statistics has been proven time and time again to be an extremely faulty logic that is only given a semblance of realism on disproportionately large groups of numbers. On smaller groups it doesn't even try to be accurate, merely claiming to give you a decent guess as to what the outcome will be," is simply a lie. Regardless, what is that we're dealing with here if not precisely what it was you claimed statistics is best at? Is ten thousand not a large enough sample size to be handled ably by statistical analysis?

I find that we agree that the one in ten thousand drop chance is inaccurate, but for completely different reasons. The "rules of randomness" say nothing of the sort - you're back to the one chance in two argument, but this time you quite clearly and succinctly defined it to actually mean one chance in two, which is just flat-out wrong. I gave the math to determine the likelihood of a drop among a population of players in which each player kills ten thousand mobs in hope of obtaining a legendary drop. Almost two thirds of the players, by the time they kill their ten thousandth mob, will have acquired the legendary drop at least once - but more than a third will not. Statistically after all is said and done, the average number of mobs that each player must kill is ten thousand, but that is not to say that the statistical likelihood of getting a drop by your 161,124th kill (which does approximate five sigma ~99.99999%) is equal to the likelihood of getting a drop on your first kill (0.01%).

Just dropped first ROLS to me, 651 rocks/260 polished gems. Working on 2nd now.

Lvl.30

Its been a long time since that statistics class I took and barely passed, but for the Legendary items, as I understand it, the 1/10,000 chance of it dropping takes effect any time you kill one of the monsters that could drop it. So even if you kill 10,000, its still just the 1/10,000 chance, each of those 10,000 times, which is why it could go on for longer.

The "you get it or you don't" isn't the odds, just the number of outcomes (2 possible).

I had never heard of statistics being unreliable, unless the way they were gathered was unreliable.

Heh, this argument is sounding familiar...

Statistics sounds good because it can work for the greater group but for individuals it is faulty 99.99999~% of the time. Just ask the guy who won the lottery. I highly doubt that anyone has ever played the lottery as many times as there are chances to win making the odds significantly higher for any individual than they are for the group. Or you can ask anyone who hasn't won the lottery and who has stopped playing the lottery and their statistics are precisely 0 out of however many times they played which regardless of the fraction you use to represent it still comes out to 0. In other words, for those individuals, the chances are 0 and 1/100,000 (to be generous) while the chances for the entire group of players could be 1/1,000,000,000. This is why statistics are faulty. I could discuss the weather or stock or sports or any number of other places where statistics are commonly used and the chances that reality matches what statistics predicts is extremely low.

Yes, I used the realistic individual options as expressed in terms of chances happening and no I didn't account for the minuscule chances that the game would bug out or your phone crash or a natural disaster might occur. Purely in terms of game mechanics there are precisely two possible outcomes: the item drops or the item doesn't drop. On any given kill your chances are precisely 1/2 because there are no other options. Because it is a random roll from 1 to 10,000 there is just as much chance to roll 10,000 ones as there is to roll every single number from 1 to 10,000 or whatever other combination you can think of. We, as humans, have a difficult time understanding randomness because in our minds random means not having the same number repeat more times than would seem random. In other words, if you flipped a coin 10 times you would expect that at least 2 or 3 times would be different from the other 7 or 8 purely because to you, that is random. Flip a coin 10 times, however, and have it land heads up all 10 times and someone is going to question either the coin or the coin flipper as to why that happened. On the other hand, true random would mean the coin could just as easily go half and half or all of one side. Expand the number out to 100 and you wouldn't expect more than a handful of any one number but true random could just as easily have 50 or even all 100 of the same number.

Statistics claims that if you kill the monster once you have 1/10,000 chances for the item to drop. Statistics also claims (as you said) that out of 10,000 kills you have about 2/3 chances for the item to drop. A true random 1-10,000 roll claims that it could drop from any of them or none of them just as easily. This is why, for a realistic viewpoint, the chances of the item dropping is the exact same thing as the number of possible outcomes or 1/2 in this case.

I still think you're mixing up "odds" with "outcomes", and they're not the same. But there's no point arguing it further since it doesn't matter enough to me. So, happy hunting.