Before I wrote further, I consulted a webpage , talking about chance, probability and odds.
It seems it is so confusing a topic that, some of the entries found in dictionaries can give inconsistant defintions, so no wonder that the non mathematician can be bewildered by it all.
So like the, do this task and get 'X'% bonus (and what is this % based on?) Is it really as big as it sounds?
Now the first question might be, what exactly is meant by 85% chance?
Now to answer what you last wrote.
It is in keeping with one aspect of my answer to Blackadder's comment.
How accurate are the complainees records, and further to this, how scientific is the data?
Though what they discovered may have some merit and be worth triggering an investigation.
It's sounds like it is just a bunch of 'this is what happened to me', rather than an audit by a professional statistician.
For one thing, players cannot conduct multiple trials, and there is no need to continue, once you have done the upgrade,and also is the total number of attempts made on upgrades a reasonable sample for a reliable analysis?
One thing that could be done is, take a hundred cards mark them up pass and fail, then not forgetting to put back a fail card ,and reshuffle the deck,see how many draws it takes to get a pass card.Rinse and repeat.
Remembering, that although pass and fail cards give a specific outcome, each of its kind is unique, so drawing a fail card may not be the same one.
Presumably this is what the program is trying to emulate, maybe it does, maybe they believe it does, but we know sometimes the game gets bugged.
Forgot to add.
My intial comment was based on the fact, it was like you were waiting for the secret instructions such that you would be guaranteed a 100% success rate with an 85% chance.
Given how long it takes to gather the materials, If the program is bugged, it makes sense to wait, otherwise it perhaps doesn't .