Red ... have you tried to open 6-7 umbite keys in raw? ... 4 is not enough for sure (probability of myth not higher then 15%)... I personally noticed notable difference between opening stacks of 105+ and less then 60 keys (umbite wise)... I am going to test my patience and accumulate this time 300 keys and will not try umbite chest unless I have 7 keys at least
Went with 10 that's where I came up with 70/30. By the time I collect enough to try again the next expansion will probably be out lol.
I remind you all again. It doesn't matter whether you unpack one key or ten thousand, as each key is unpacked, the virtual dice are rolled, then compared against the loot table for Dirt Encrusted Keys. The loot table is weighted so that 90% come out gold, 9% platinum, and 1% umbrite. Period. There is no method you can use to get more umbrites, it's purely random.
In general this is correct but devil usually in details ... if you not opening 100 keys at once you not rolling for your 1% (from my counts of 400 keys its more like 2%). Random basically set for your maximum chance but with lesser stuck you are hitting for minimum
Exactly, well said! Just because someone gets X umbrite keys from N dirt encrusted keys, it doesn't mean that X/N is the probability of getting X umbrite if you open N keys.
And where did you study in the fields of statistical probability again? If your sample size is 30 I would agree with you. We're talking a sample size of 1000 keys spanned out into sets of 100 over a several day timeframe in which case that is exactly what it means. As your sample size increases the accuracy increases. Formula of total probability through sampling = Positive result/total sample. Which has been concluded to be within the realm of .5-1% as are the results for many. As with any random sample you will have "flyers" but the average remains the same. Your point of contention is the exact opposite of what he was saying. Nobody has guaranteed 1000 keys = 10 umbrite. We are stating the likelihood to get what you want. Which are slim. To have an EXACT chance would require a look at the coding which nobody here has access to. In short X/N IS the probability.
it is all random i have opened 155 keys total and got 6 plat and 4 umbrel keys for my 4 umbrel keys i got 3 wands and one charm. and of course i am not all access so i can not use the wand
I suggest you read what I wrote. The key words were "Just because" at the beginning. I agree that 30 is a rubbish sample size. 1000 isn't great either. The chances of getting X umbrite from 1000 keys assuming that opening a key is a binomial trial with a probability of success of 0.01: P(X = 9) : 12.56% P(X=10) : 12.57% You can see from this that even though we *know* that our umbrite chances are 1%, the chances of getting 9 or 10 are almost exactly equal. According to your statement "In short, X/N IS the probability", if I opened 1000 keys and got 9 umbrite, I would claim that the chance of getting an umbrite is 0.9%, despite this being wrong. Therefore I stand by my statement (this time with numbers): just because I got 9 umbrite keys from 1000 dirt encrusted keys, doesn't mean that the probability of getting an umbrite is 0.9%. I'm perfectly willing to accept that the chance of an umbrite key is 1%, but I would base that number on the propensity of people to use nice round numbers, rather than as the result of the limited statistical information so far. In short I didn't study statistics and probability in a place that used words like "flyers", I studied in a place that allows me to calculate that I'd need to open 69 keys to have a greater than 50% chance of getting at least one "1%" umbrite. [As an aside, I've opened 137 keys and got 0 umbrite. The chances of that happening using the figures above is about 25%.]
Drake, you have been wrong throughout this whole discussion. You have no idea whatsoever of how the mathematics of random number generators works. If you open 100 keys, it is doing a random roll FOR EACH KEY. It doesn't matter if you open one, wait a week, then open another key. It doesn't matter if you open a huge stack at once. Each key is the result of its own, unique, random roll. The RNG doesn't care if you do them as you get them, one at a time, or in batches of a thousand, just like it doesn't care if you are naked and dancing with clown makeup on as you open them. There is NOTHING a player can do to change the odds of getting more umbrite keys. The RNG rolls a number. That number is compared to the loot table. The loot table is weighted, so perhaps rolls of 1-90 are always a gold key, rolls on 91-99 are always a platinum key, and on a roll of 100 it gives you an umbrite key. Anything else is rank hocus-pocus superstition. You might as well sacrifice a chicken on your keyboard.
How many times can I explain that it's a separate "dice roll" for each key you open, no matter how many you open, period? Your chances of getting a particular type of key are controlled by that roll and the loot table. If your "roll" is 1-89 or so, you get a gold key. If it's 90-99, you get a platinum key. If you "roll" 100, you get an umbrite key. I sometimes open the keys as I get them, one at a time. Sometimes I buy keys, and when you open them you open all in your stack. I've opened stacks of around 50 and stacks as large as 300. The results are extremely consistent, and follow the assumed loot table I used in the prior paragraph. I'm tired of explaining this to someone who acts like a concrete-headed bovine. It's like talking to the cinderblock wall of your superstitious B.S.
All right - let me ask you differently ... did you take a science class in the school? Do you familiar with term "experiment"? do you familiar with terms "statistically meaningful result"? Theory good only if experiment proves it .. otherwise its' philosophical blah-blah-blah So far my experiments (and results from guildies) tell me this with stacks less then 100 - solid result less then 1% With stacks of 10 (many people do that) - many report results 0.5-0.7% with stacks 150-200 - solid results way above 1% (its more like 1-3%) That is experimental results - and I am sorry if they not fit into your theory
Did you take statistics? How about any computer science? I've worked in I.T. development since 1982. Your stupidly small sample sizes cannot tell anything whatsoever. It is statistically possible to open 300 keys and get 300 gold, but it's unlikely, just like it's possible to flip a coin 1000 times and get all heads. Open a dirt encrusted key >> ask the RNG for a value from 1-100 >> take that value and look it up on the loot table LOOT TABLE Roll-----Result 1-89-----Gold Key 90-99-----Platinum Key 100-----Umbrite Key
Well I have Ph.D. in molecular biology so yes I do familiar with statistics but a lot more I trust to experiments. Your problem that you might have wrong assumption to begin with. You are assuming that programmatically /rand is true random numbers. All what I am saying - theoretically you "should" see same numbers no matter what do you open 100 keys 1 by 1 or 100 keys as a stack ... but experiment give you somewhat different numbers which mean there is factor that we don't take in account
If you have a PhD in molecular biology, presumably you have taken at least one class on designing experiments. So what's your experimental methodology? What different sizes of stacks have you tried? What is the relationship between size of stack and probability of umbrite keys, is it a discrete or continuous function? How many trials are you running? Have you presented your methodology to other players so they can test it out and have they achieved statistically equivalent results? What's the confidence interval for your results? Or, let's look at it from a non-mathematical perspective. Did the devs spend time creating a function specifically designed to reward those who either buy or hoard a large quantity of keys, or did they just use a loot table that gets queried the same way every time a key is opened? As I said in a previous post, 1000 keys opened isn't anywhere *near* enough to gain statistically significant results. Let me know when you've opened 100,000 keys one at a time, and a further 100,000 keys 200 at a time, and what was the quantity of umbrite keys for each experiment.
What p value are you assuming for your experiment? There is no reason whatsoever that a developer will go out of their way to make more work for themselves. That means that the random roll is a call to whatever the random-number function is in the language library they are using. These are, indeed, pseudo-random, but the standard algorithm generally used for a random number generator is a linear congruential generator, which produce values from an extremely large field that can't be determined to be non-random without major computer effort. In other words, our human observations cannot distinguish the numbers produced from true randomness. So, you open a dirt encrusted key. The developer's code makes a call to the random function in the language library, a number that is more than random enough for any normal human usage is generated, then the code compares this to the loot table.
I will... No in Russia we do not take extra classes in Grad School for experiment design ... this this study during junior years . What we do care is correlation between results .. if you have 5 experiments and all give you same result within 20% margin - then you may consider to have correct number ... If you getting 3 out of 5 within 20% range and rest 2 falls out - then you need more experiments to get your numbers rights And here I am not talking about theories or codes designs - just naked number of experimentation... explanation of experiment is completely different beast
For quality experiment in biochemistry you need 5 experiments with same condition that produce results within 20% of each other to be considered "correct results to be published" with 3 results within 20% of each other and 2 falls out - you need perform more experiments but generally speaking 10-20% of results that falling out is acceptable
now in my inventory 1810 gold keys, open only platinum and umbrite keys. 100-150 platinum keys = 29-33 epic relict, 20 mith altars 19(20?) umbrite keys 10\9(10) pandora boxes and golden ring Mith bow - hope dies last. Maybe it makes sense to write in a support? = (People pulls myth with 1-10 keys, and two dozen shots into the void ....
