Where did you come up with that information? I couldn't find much in the way of criticism about random.org who is used to generate the random numbers for the wishing well (there is no way to know if that is used for in-game use, but I doubt it because I expect the return time for the hundreds of thousands of necessary random numbers coming from an external site would not be practical. So more on that later).
random.org is headed by Dr Mads Haahr, School of Computer Science and Statistics, Trinity College, Dublin 2, Ireland - this according to their web page. According to their site they use some factor of atmospheric noise to determine the random number instead of the typical programmed way (again more on that later). I wouldn't understand the science of the picking up of atmospheric noise and turning it into a seed value for a random formula but conceptually it sounds more random than how it's normally done.
I could only find one blog about someone saying that generator is skewed. They generated some numbers between 1 and 10 and kept track of the results. They found some numbers picked more than others.
Do you know what the sample size (iterations) was? 100. I nearly fell out of my chair. That is such a statistically insignificant amount that it's beyond laughable. I can't believe he posted it as valid information. You need thousands of iterations at least to have enough of a sample to have any idea of the distribution.
Now concerning random number generators. And with a disclaimer of not knowing exactly how atmospheric noise is used in this case. It's always been that the proper phrase was pseudo random number. This is because a computer is not capable of generating a true random number. It always requires a seed value to the formula and that seed value can be predictable and if you could 'freeze' every bit and reproduce it you would get the same random number. It's a calculation which doesn't change, only the seed value does depending on how you want to approach it the seed is changed every time or only periodically. I've written my own random number generators several times for the programming exercise, and I found it very interesting.
Usually the microseconds are taking from the computer clock and then ran through a modulus formula. This results in a predictable range of values but no (practically) predictable exact value. I used different modulus parameters in experiments and found they all pretty much worked on the same more-or-less. But I used thousands of repetitions sometimes letting the computer run all night doing it and found that I could get a very even spread across the range. Of course it would never be 100% even but it sure was close enough for computer games.
Based on my research and experience I see nothing that tells me you can't trust random.org. And concerning the core BB code I expect they are using a standard bif (built in function) inherit to the language which would have been well tested before the compiler or dll's or whatever programming structure is being used before BB incorporated it. Of course any idea of how the code might be structured is pure conjecture but I'd bet they didn't hand-write a pseudo random number generating routine from scratch.
Which leads to a quick (hopefully) mention of probabilities because this thread has elements of that. Remember there are two distinct questions that some people try to make the same question.
1) What is my chance of getting my 5% drop this time?
2) What is my chance of getting at least 1 5% drop if I play the adventure x times?
They are not the same question. Question 1 is easy. 5%. Each time you open the mail to see your loot it's 5%. It will never change.
Question 2 is about a sample over time where only 1 out of x needs to work. Just to keep it simple as you increase the number of attempts the chance that one of those attempts will have dropped what you wanted. And as you make x bigger the chance gets bigger. But it will never reach 100%. It just can't. Why? because in every pass there is a chance it won't drop. Therefore no matter how many attempts are made you will never be 100% sure to get one. If you could manage to make hundreds of thousands of attempts you could get to 99.99999999% but there will always be that .000000000000001 that will trip you up.
Put it this way. What's my chance of flipping a coin and it coming up heads? 50%. easy.
Let's say I flipped a coin 9,999 times and by some miracle they were all heads. What' my chance of flipping the coin and it coming up heads. 50%. That hasn't changed.
But if I ask "what's my chance of flipping a coin 10,000 and always getting heads" - that's a very different question and the odds are very much against it.
Hope any of this helps or at least might have been interesting on some level