Once you factor in crit chance it’s possible for the advantage version to outperform the TWF with no mod damage on 2nd attack version.

An interesting possibility I hadn't considered. We're back to 2014PHB crits, so sneak attack would get increased. Let's look at the math.

Assuming a hit-but-not-crit chance of 60% and a crit chance of 5% on a single die (so 35% chance to fail).

W=Weapon damage

S=Sneak Attack damage

B=Bonus damage

Rolling once with advantage gives us:

No hits = .35*.35 = 12.25%

Crit = 1 - (.95*.95) = 9.75%

One hit = (1 - .1225 - .0975) = 78%

Expected damage = (W+S+B) * 0.78 + (2(W+S)+B)*.0975 =

**0.975W+0.975S+0.8775B**
TWF gives us:

On-hand attack:

Expected Damage (W+S+B) * 0.6 + (2(W+S)+B) * 0.05 =

**0.7W+0.7S+0.65B**
Off-hand attack:

Sneak attack can only be applied if not already done. 35% chance the on-hand attack was a miss and sneak attack can be applied here.

Expected damage = (W + 0.35S) * 0.6 + 2(W + 0.35S) * 0.05 =

0.7W+(0.21+0.0175)S =

**0.7W + 0.2275S**
Total expected TWF damage is

**1.4W +0.9255S+0.65B**
Advantage attack expects to do 5% more Sneak Attack and 22.75% more Bonus damage - DEX mod, magic, and other static numerical bonuses.

TWF attack expects to do 42.5% more weapon damage.

Best TWF without any feat expenditure is a short sword, so 3.5 damage on a hit (and the hit chances are already worked in). So that's 1.4875 expected damage.

First couple of levels this edges tows TWF. Late game this heads towards advantage. At it's biggest discrepancy, a 20th level rogue (10d6 sneak) with a +5 DEX and +3 weapon we get:

(35 * 0.05) + (8 * .2275) - 1.4875 = 1.75+1.82-1.4875 =

**2.0825**
So at it's most for a 20th level character it's around a 2.1 HP per round expected difference. You are right, but in practical terms they are pretty much the same.

Thanks, I hadn't considered that option.