SacredWitness just probably made a mistake in calculation. The 25% is more or less correct. More importantly, advantage doubles the chance to make critical hit and nearly negates the chances to critically miss. Vice versa disadvantage. This is very often forgotten part of that feature.
Sorry to be picky, but it's not more or less. It's correct lol.
In order to calculate an advantage for a roll, you just take the roll percentage evaluated with a max value of 1, which for 50% is 0.5. Then you reformulate what you are looking for: "Finding the chance you succeed an advantage (to succeed at least one roll of two with 50% chance each), it's the same as finding the chance you won't have two failures".
The chance not to have two failures is 1 - the chances to have two failures. So: 1 - ((1 - chance to succeed) * (1 - chance to succeed)).
Which in the case of 50% is: 1 - (0.5 * 0.5) = 0.75 => 75%.
In the case of a 60% roll it's: 1 - (0.4 * 0.4) = 0.84 => 84%.
I don't get your way of reasoning. In the case of a 10% roll it's: 1 - (0.9 * 0.9) = 0.19 => 19%. This is an increase of 9% (or 90% depending on what you are considering to be the base). What's messing with numbers are criticals. You really can't reduce advantage or disadvantage to one precise or flat number. It's not that simple. What is important that we agree the bonus is significantly huge
I don't see why critical messes with numbers. What I am calculating here is the percentage increase of chance to hit in case you are having an advantage. And my calculation is accurate according to probability laws.
And yes, if you have a 10% chances to hit, when you have an advantage, it makes it 19% indeed. I am not talking about dice results here, just probability to hit in terms of percentage, which is what they give us in game. That's why we get 56% or 87% in the tooltip sometimes, because it's the raw calculation of our chances to hit.
Besides, on a computational standpoint whether you check 2 random numbers against their respective targets or you calculate the final probability and check only one random number against it, it remains the same. We don't know if they actually 'roll' 2 dice for advantage/disadvantage. The only thing we see is the result.