D&D 5E [GUIDE] I Fought the Law and Won: The Rogue Guide

Yunru · Apr 25, 2019

Another nice one is if you also pick up Warcaster, and... some way to be less squishy (AT maybe?).

Esker · Apr 25, 2019

Yunru said:
Another nice one is if you also pick up Warcaster, and... some way to be less squishy (AT maybe?).

Yeah, if you're building around booming blade you're not making *that* much use of your swashbuckler features, so you may as well get find familiar and mirror image with your subclass and free up a feat later on for warcaster.

vostygg · Apr 26, 2019

Yunru said:
Does it though?
Booming Blade does +3.315 more damage if the extra damage triggers every time that you hit.
The break even point is if you trigger the damage 38.61% of the time (not even factoring in the action denial possibilities).

So unless Booming Blade triggers less than 38.61% of the time, Booming Blade provides higher DPR for Swashbucklers on average.

(I kinda want to work out a build that uses Shield Master to guarantee they'd have to move to make an effective attack.)

My experience is that Booming Blade triggers far less than 38.61% (closer to 10%), but YMMV. That's why I consider it highly situational, and why I stand by the statement you quoted that "TWF provides higher non-situational DPR for Swashbucklers on average". By the way, your TWF Barb 2 / Fighter [Champion] 5 / Rogue [Swashbuckler] 13 crit-fishing build is the gold standard, as far as I'm concerned, in terms of cases where TWF actually makes sense.

EDIT: I'm also curious about the math behind your 38.61% threshold. By my own computations, the break-even point is closer to 33.25%. That's the point at which Booming Blade comes out ahead in terms of average DPR across all levels.

Esker · Apr 26, 2019

vostygg said:
EDIT: I'm also curious about the math behind your 38.61% threshold. By my own computations, the break-even point is closer to 33.25%. That's the point at which Booming Blade comes out ahead in terms of average DPR across all levels.

I get 36.8%. Guessing the 38.6 was a typo.

Math:

At level 5, the rider does an average 9 damage. Without the rider, TWF is ahead by 3.315. The average rider damage is (probability) * (rider damage), i.e., 9 * p. Set 9 * p = 3.315 to get p = 0.368.

Esker · Apr 26, 2019

vostygg said:
By the way, your TWF Barb 2 / Fighter [Champion] 5 / Rogue [Swashbuckler] 13 crit-fishing build is the gold standard, as far as I'm concerned, in terms of cases where TWF actually makes sense.

Where's the synergy between crit-fishing and TWF? Seems to me in a crit-fishing build you want to pack a lot of damage into a single attack and then find ways to get advantage on that attack (which is easier than getting advantage on all attacks in a turn).

Speaking of which, in the discussion above we always assumed there was no source of advantage. But if you can get advantage (particularly if you are able to get advantage on your first attack, as from a familiar taking the help action), BB becomes better than TWF. One reason why arcane trickster is a better single target DPR choice than swashbuckler, since they have easier access to advantage (and free rein of races, since they can get find familiar and booming blade from their class).

vostygg · Apr 26, 2019

Esker said:
Where's the synergy between crit-fishing and TWF?

http://www.enworld.org/forum/showthread.php?553760-Repeatable-DPR-Kings

Simply put, more attacks means more chances to crit.

When I assume a 60% hit probability, and plug this into my DPR calculator program, I get the following, which is the best I've seen for TWF:

----------------------------------------------
Critfish TWF Barb2 / Fghtr 5 / Rogue 13 - DPR
----------------------------------------------

Level 1: 8.75
Level 2: 12.44
Level 3: 14.96
Level 4: 14.24
Level 5: 14.89
Level 6: 16.78
Level 7: 25.17
Level 8: 28.56
Level 9: 28.56
Level 10: 32.81
Level 11: 33.63
Level 12: 37.86
Level 13: 37.86
Level 14: 42.09
Level 15: 42.70
Level 16: 46.91
Level 17: 46.91
Level 18: 51.12
Level 19: 51.12
Level 20: 55.33

vostygg · Apr 26, 2019

Esker said:
I get 36.8%. Guessing the 38.6 was a typo.

Math:

At level 5, the rider does an average 9 damage. Without the rider, TWF is ahead by 3.315. The average rider damage is (probability) * (rider damage), i.e., 9 * p. Set 9 * p = 3.315 to get p = 0.368.

I see. I am multiplying the frequency with which we expect the rider to trigger by the average rider damage at each level and adding that to the Booming Blade DPR at each level. I am then adjusting the frequency with which we expect the rider to trigger until the average DPR across all levels for Booming Blade exceeds that of TWF. That becomes the break-even point.

Example - Level 5 DPR assuming a 33.25% chance of triggering the rider damage

Regular DPR: 0.6 * (avg(2d8) + 4) = 0.6 * (9 + 4) = 0.6 * 13 = 7.8
Crit DPR: 0.05 * avg(2d8) = 0.05 * 9 = 0.45
Regular Sneak Attack DPR: 0.6 * avg(3d6) = 0.6 * 10.5 = 6.3
Crit Sneak Attack DPR: 0.05 * avg(3d6) = 0.05 * 10.5 = 0.525
Rider Damage DPR: 0.6 * (0.3325 * avg(2d8)) = 0.6 * (0.3325 * 9) = 1.7955

Total DPR: 16.8745

Repeat this for every level from 1 to 20 to get the Booming Blade DPR adjusted for rider damage. Then compare it to the TWF DPR at each level to get the DPR difference at that level, and take the average of these differences. Keep adjusting the chance that the rider damage will trigger until the average of the differences is close to zero, and you will have the break-even point. Feel free to suggest corrections if you think I've made a mistake somewhere.

BTW: When I add the Dual Wielder feat to the TWF build, the break-even point shifts up to about 52.5%, which is to say that the Booming Blade rider has to trigger at least 52.5% of the time in order for Booming Blade DPR to exceed TWF DPR on average across all levels.

Esker · Apr 26, 2019

vostygg said:
Simply put, more attacks means more chances to crit.

In general, if you hold total dice damage constant, one chance to crit yields the same expected added DPR as two chances to crit on weaker attacks. I see at the link that this build relies on constantly applying reckless attack and holding back sneak attack unless you crit. It's only because you're doing that that having the second attack actually buys you something re: crits. But I guess if you have constant advantage, elven accuracy, and a wide enough crit range that you can expect to crit most turns, then it's probably worth it.

I don't think this build works as written though: Elven Accuracy doesn't work with STR-based attacks, and rage only works with STR-based attacks. It may have been based on an obsolete UA wording of EA.

EDIT: Oh, I see that the thread predates the release of Xanathar's and therefore of the official Elven Accuracy feat. So it definitely is based on the UA version.

Esker · Apr 26, 2019

vostygg said:
I see. I am multiplying the frequency with which we expect the rider to trigger by the average rider damage at each level and adding that to the Booming Blade DPR at each level. I am then adjusting the frequency with which we expect the rider to trigger until the average DPR across all levels for Booming Blade exceeds that of TWF.

It's not the break even point though if you're counting cases where the leader goes back and forth as you level as favoring TWF. In any case, this is splitting hairs. I agree with you that the rider will go off less than 1/3 of the time.

EDIT: I may have misunderstood you. I thought you were saying that BB had to exceed TWF at every level, but now I think maybe you were saying that you averaged across levels. That's at least not biased, but I don't think it makes sense to average over levels. For one thing, that assumes that you have the same amount of play time at every level, which isn't really reasonable. For another, the value of one point of damage means something very different at a high level compared to a low one (not to mention the variable value of the control effect and the variable value of a bonus action).

vostygg · Apr 26, 2019

Esker said:
It's not the break even point though if you're counting cases where the leader goes back and forth as you level as favoring TWF. In any case, this is splitting hairs. I agree with you that the rider will go off less than 1/3 of the time.

EDIT: I may have misunderstood you. I thought you were saying that BB had to exceed TWF at every level, but now I think maybe you were saying that you averaged across levels. That's at least not biased, but I don't think it makes sense to average over levels. For one thing, that assumes that you have the same amount of play time at every level, which isn't really reasonable. For another, the value of one point of damage means something very different at a high level compared to a low one (not to mention the variable value of the control effect and the variable value of a bonus action).

I edited my post to expound a bit more on the specific math I am using. Feel free to cross-check.