Bill Zebub
“It’s probably Matt Mercer’s fault.”
2d6 averages 7 when both attacks hit. The probabilities that one attacks out of two hits is higher than a that of a single attack landing, but the probabilities that both attacks hit are lower than that of a single attack landing a hit.
If we assume that the first attack hits in both cases, d6 averages 3.5+stat, d10 averages 5.5+stat. Then, the second attack might or might not hit, so let's say your target number is 11 on the d20 to hit the AC, that's 1 chances out of 2, averaging to 1.75 damage. The lower the TN, the more this second attack contributes to average damage.
Then again it's possible that only the second attack hits, in which case TWF nets a 3.5 damage over one attack at d10 (because there's no bonus to damage from stat). So my maths are not the greatest but eyeballing it overall, two attacks at 1d6, with only the first adding stat bonus to damage, probably washes out with one attack at 1d10+stat bonus.
Oh, I see.
The only difference is that the standard deviation is higher for the two attacks, which is what you are trying to describe. But expected DPR is still average damage * chance to hit.
Really the only time it matters is when there is a reason you need to hit beyond that average damage, such as rogue sneak attack or preventing regeneration, in which case dual is superior.