D&D 5E The impact of overkill damage

[jgsugden]

... As DPR increases your average damage increases which increases your overkill. However, at the same time, doing more DPR increases your chance of landing a killing blow which decreases your overkill.

No. It does not.

DPR is a per round metric. If I have three attacks that deal 40 damage, 10 damage, and 10 damage; one attack that deals 60 damage; or three attacks for 20 damage each, I'll have the same DPR (60). However, each of those scenarios, with identical DPR totals, is going to have insanely different interactions with overkill.

Overkill is relevant on a PER ATTACK basis. Not a DPR overall basis. If you fix DPR at approximately the same level, as I did in my earlier posted example, you can see just how massive an impact overkill has on efficient damage when you have one higher damage attack compared to a greater number of lower damage attacks.

Your argument that that we do not land the killing blow all of the time is mostly irrelevant to the evaluation. It doesn't matter that we're mixing our damage in with ally damage against a foe. Sometimes we'll be landing the first blow, and sometimes others will be softening a foe up for us. We're still going to see a spread of times we're attacking a totally healthy foe, a wounded foe, and a foe on death's door.

I referenced this phenomena earlier and discussed the one biggest element where it does make an impact - primarily for rogues and monks. They tend to attack early in the first and surprise rounds when the enemy is likely unwounded, giving them more effective damage on the first round due to less overkill on round one. But just as a lead off batter's RBI total is likely to be less overall because there is nobody on base during their first at bat, but RBI total for second plate appearance and on is much closer to other players, the effective damage on other rounds shows a steep drop off due to overkill. It is of limited impact, overall.

I've recorded and analyzed actual game table data for years and years - in 5E and prior editions - and I can tell you, based upon the totality of that analysis, that DPR is an overdramatized metric that has far less to do with the efficacy of a PC by itself than it is given credit for doing, and that an efficient PC build that deals less DPR can easily do FAR MORE effective damage than the higher DPR build when you concentrate damage in a single attack.

 

[Ancalagon]

I made no claim that you should disregard roleplaying considerations for tactical ones.

But you said that "if the enemies don't take into accounts Fireball, the DM isn't being tactical". The DM could be very tactical - tactics that his NPC foes know of. "Overwhelm the enemy with sheer numbers" is an excellent tactic if the other side doesn't have a way to hit masses of combatants. It just happens that in this case it was a terrible idea...

 

[Ashrym]

It's not possible to be less effective in combat by doing more damage. More damage is clearly better regardless of overkill arguments.

Overkill does exist and does reduce effective DPR through lost damage. This is more prominent in bigger damage from less attacks than lower damage with more attacks. IE a fighter with 3 attacks will have better effective DPR than a rogue with 2 attacks when DPR calculation works out to be equal because that last attack is more likely to target a fresh opponent.

I find DPR calcs are over-rated and can be misleading because of it. Not as much from overkill as the fact that a combat never lasts long enough to see those averages anyway; especially with swingy spike damage.

Damage is definitely worth looking at, but understanding what mitigates that damage is also important.

 

[MarkB]

It's not possible to be less effective in combat by doing more damage.

Yeah, but this isn't about doing more damage. It's about doing the same overall amount of damage but in different portion sizes.

 

[FrogReaver]

Yeah, but this isn't about doing more damage. It's about doing the same overall amount of damage but in different portion sizes.

i propose that if overkill only “matters” when DPR is the same or nearly so that my point is made and overkill can be safely disregarded in DPR discussions.

 

[Ashrym]

Yeah, but this isn't about doing more damage. It's about doing the same overall amount of damage but in different portion sizes.

It's still not possible the net 0 increase in all scenarios because overkill can only apply to some attacks.

It's TtK (turns to kill) instead of DPR also getting looked at, but TtK is still dependent on damage. HP increases with CR rapidly to give benefits to higher damage.

More damage > less damage regardless of mitigating arguments. It's possible to situationally net 0 gain from overkill but never lose ground and gains in other scenarios will exist for an overall net gain.

The real challenge is comparing non-damage to damage in alternative abilities. ;-)

 

[MarkB]

i propose that if overkill only “matters” when DPR is the same or nearly so that my point is made and overkill can be safely disregarded in DPR discussions.

You're free to propose that, but it's not what I said, and it doesn't make any sense either. The point isn't that it only matters under those circumstances, it's that it matters even under those circumstances.

 

[Fanaelialae]

But you said that "if the enemies don't take into accounts Fireball, the DM isn't being tactical". The DM could be very tactical - tactics that his NPC foes know of. "Overwhelm the enemy with sheer numbers" is an excellent tactic if the other side doesn't have a way to hit masses of combatants. It just happens that in this case it was a terrible idea...

No, the monsters may be behaving tactically (based on the information they have), but the DM most certainly is not. The DM is choosing tactics that they know to be poor, because it's what they believe the monsters would do. Which is a perfectly valid choice for a DM to make, but is not based on sound tactics.

It's entirely possible that the DM can make a valid choice from both a RP and tactical standpoint. For example, hobgoblins who've encountered the party and know to expect fireballs.

 

[Ovinomancer]

I'm going to assume your math is 100% right for a moment (I haven't checked it yet). Let's talk about what the conclusion means. You showed an effective 1 DPR difference due to overkill. I would consider that result as supporting my claim that overkill doesn't really matter. I've not claimed that it doesn't have any impact afterall, just a minimal impact that can be safely ignored.

On a side note - your work shows something else very interesting. Overkill depends on your chance for a killing blow and average damage. You said that yourself. But let's think about this for a moment. As DPR increases your average damage increases which increases your overkill. However, at the same time, doing more DPR increases your chance of landing a killing blow which decreases your overkill.

Yes, I used your assumption set, which I think is incorrect, which has the effect of reducing the impact of overkill. Besides, the point I was making was how to evaluate overkill damage, not that it would reveal a huge discrepancy. The larger the damage delta is, the more effect it has, though.

It's useful for other examples as well. If we continue to use your assumption on killing blows of 1/4 of attacks are killing blows and combat lasts three rounds, let's look at two fighters, one who takes a +STR ASI and the other who takes GWM. We'll set the AC for the target at 13 so as to be well within the GWM range.

Fighter 1, STR 20, level 5, 2 attacks, dmg(F1) = 2d6+5 = 12, %hit = 80%

Fighter 2, STR 18, level 5, 2 attacks, dmg(F2) = 2d6+14 = 21, $hit = 50%

The calculated X' for Fighter 1 is 8.93, so per round damage is 17.86. For Fighter 2, X' is 9.75, for a DPR of 19.5 The DPR difference here, due to overkill using your assumption of frequency, is less than 2. Since, as you say here, a difference of 1 doesn't really matter, does less than 2 matter? In other words, using GWM within the determined bandwidths is brought to close to no difference. If we really want to be precise, we'd add one more attack of damage to Fighter 2 to represent the first bullet of GWM, but that occurs once in 6 attacks, so it's 1/6th of the damage -- that's still +3.5 DPR, though, which really just proves that it's the extra granted attack that makes the difference and not the +10.

Interestingly, the DPR delta actually get better for the GWM fighter at higher ACs because they lose less due to overkill. They also hit less often. At a hit% of .6/.3, X' delta drops to 0.817.

As an update, the OAD calculate should take the rounded down value for X. This is because X can have fractional values but target hp cannot. The difference this makes is very small, but improvements are improvements.

 

[jgsugden]

@FrogReaver - Look, it has been shown over and over and over and over and over and over again that if you ignore overkill when looking at DPR, your DPR results will be misleading as an estimate of how effective a PC is in combat. If you're comfortable with the misleading data, more power to you.

I really suggest you stop posting and start collecting data. At your next couple dozen games, record the sessions and go back through the combats after the game and calculate how damage is dealt, how much overkill there is, and how much effective damage each PC deals both in total and on average per target (so that a fireball dealing 28 to 10 foes is evaluated both as 280 damage and separately as only 28). You'll find the effective damage dealt by a GWM, a rogue, a sharpshooter or other "high hit" PCs is offset quite a bit by overkill. You'll see that in actual play, once you collect enough data, it tends to follow the trends I showed with my example earlier in this thread.