This topic does come up every now and then. I can super guarantee that the percentages are very accurate. If you are certain that this is not the case, please record a video of it failing to be accurate over a significant number of contiguous crafts.
If you can provide me with proof that it is not actually accurate, I will rewrite the code for it.
Neither are off, all the numbers are perfect. I’m generally pretty hesitant to make statements like this but I’ve tested it so much that I can very confidently say that the numbers are right.
Because this keeps coming up, I keep testing it. In total, I have now done a total of 500 recorded crafting tests on the live version of the game. My luck variance was -0.02%. Anything within 1% and I would call that accurate. My results were 50x more accurate than I would need to be happy with the system.
I have also done simulation testing. This uses the exact same functions as the actual crafting but instead of just doing it once, it runs each craft 1 billion times. This was accurate to a number than can only reasonably be represented with scientific notation.
I’m not sure what the base instability on an item like that would be off the top of my head sorry.
Just to be clear, if an item says 99% success chance, it can still get a fracture on the first try. It’s ultra rare but still possible.
Yea, just to be clear, I’m not defending the system itself. Just the implementation of the system. This is something that we are taking very seriously and have been working on updates to the system for quite a while now. They weren’t quite ready to incorporate into this patch so we pushed them back.
I take a little exception with the XCOM comparison because you have a real world representation of the scenario that makes the percentage seem ridiculous. That’s not what’s happening here at all.
devtrackers.gg