Rolling one attack with advantage is worse than rolling two normal attacks (because we have two attack rolls either way, but only one vs two chances of dealing dmaage).
However, what of the inverse situation? Is it better to roll two attacks with disadvantage, or one normal attack?
Since this will likely depend on target number, assume an unmodified rolls of 8, 12 and 15 are needed to hit.
The main reason I’m asking this question is for a homebrew feat that allows trading one attack for disadvantage on two attacks, but this could also come up in RAW, if you had disadvantage and multi attack, the decision would be between shoving your opponent prone and taking one attack, or taking two with disadvantage.
Best Answer
2 attacks at disadvantage is better only when a single roll has more than 50% chance to hit
It's probably easiest to talk about this through calculating average damage based on a 'standard' hit chance. This standard hit chance is the chance to hit with a single roll. To make calculations easy and understandable, let's set the damage dealt on a hit to 1, changing this number does not at all affect the results but it saves me some type work. I'm not taking critical hits into account, they do change the result but only slightly in favor of a single normal attack. Accounting for crits is a bit rough because their influence depends on more variables, like the size of the damage die compared to your damage bonus and any additional features like Sneak Attack or Hunter's Mark that add more dice to your damage rolls. This answer is meant as a general guideline for how these different attacks affect general gameplay and unless some character has a crazy combination of features that heavily benefit from that small chance to crit it holds up pretty well.
On a single standard attack with a potential damage output of 1 your average damage is simply the hit chance, for example if you have 80% chance to hit for 1 damage your average damage will be 0.80x1=0.80.
On a single attack at disadvantage you pick the lowest roll, so in essence both rolls need to 'hit'. If normally you would have a 80% chance to hit, you now have a 0.8x0.8=0.64=64% chance to hit. So you have a 64% chance to hit for 1 damage, your average damage in this case is 0.64.
On two attacks at disadvantage you could do some math to figure out the average hit chance but since we're talking about average damage output this is equivalent to simply doubling the average damage from a single attack at disadvantage. So with a normal 80% chance to hit, brought to 64% due to disadvantage, you have an average damage of 2x(0.8x0.8)=2x(0.64)=1.28.
Now we can generalize this. For a single attack your average damage is P, with P being the standard hit chance for an individual roll. For two attacks at disadvantage this average damage is then 2*(P*P) or 2P^2. Plotting these 2 functions yields:
The straight line is the single straight roll, and as you can see it is worse than attacking two times at disadvantage as long as the normal hit chance is above 50%.