| 1 person marked this as a favorite. |
Have seen next to no "Damage isn't important" discourse. Have seen loads of "DPR is incredibly misleading and not very useful".
This thread makes me think the OP conflates DPR and damage.
"DPR is incredibly misleading and not very useful".
And I think a lot of people are mislead by this assertion, thinking that DPR calculations are actually misleading and not helpful.
| 5 people marked this as a favorite. |
"DPR is incredibly misleading and not very useful".And I think a lot of people are mislead by this assertion, thinking that DPR calculations are actually misleading and not helpful.
The problem arises because some people try to take them out of the context the belong in.
DPR calculators are immensely helpful when evaluating two specific character options.
For example, how will Vicious Strike (formerly Power Attack) impact my characters abilities.
I'm not going to do all the math, or find the break points, but you could use it to evaluate how much the enemy's AC needs to be above your to hit value such that Vicious Strike is an increase in damage per round over making two regular strikes in a turn.
However, people could mistakenly come to a conclusion that oh, they should make 3 attacks in a round because that leads to a greater DPR (but it will be miniscule because of poor chance to hit) rather than doing something like moving away from the enemy or raising a shield.
DPR is a useful tool within a certain context. The problem is that people often try to use the tool where they shouldn't.
The truth is that people who overly focus on DPR forget the wider context of the game, and that damage isn't the only thing you should care about. But the decriers of DPR forget that dealing HP damage is the primary and common way of dealing with threats to the party (in combat). So it makes sense that you should evaluate how options impact your ability to do that.
However the truth is in the middle ground. Understanding the limits of DPR as a tool, and knowing that DPR isn't the only thing you should concern yourself with. And I suspect many people realize this. It's just those voices in the middle ignore the extremes at both ends and decide not to bother trying to correct either of them.
DPR calculators are immensely helpful when evaluating two specific character options.
For example, how will Vicious Strike (formerly Power Attack) impact my characters abilities...
The vast majority of calculations I've seen done merely calculate average. Now that's useful, but not the end of the story. Distribution is important. Multiple attacks vs. vicious strike is a good example; because multiple attacks derives more damage from crit hunting (and lower chance of complete miss), you can get situations where it's calculated average is higher but the most probable damage amount you get in any given round is lower.*
Using "average only" to decide what feats or strategies to use is particularly problematic given that you may only make 9-12 attack rolls per session. Thus, basing your feat choice or combat strategy decision solely on the average is to somewhat fall for the small numbers fallacy. I mean, it's a definite step up from no information at all, but it's not necessarily giving you the optimal strategy for every situation or every play style.
*For those who don't understand this, here is a toy example: imagine two feats each of which lets you roll a d20 and apply some damage. Feat #1 says: a roll of 1-18 causes 1 damage, a roll of 19-20 causes 22 damage. Feat #2 says: a roll of 1-20 causes 3 damage. The average damage of Feat #1 is 3.1, which is higher than Feat #2's average of 3. So all those calculations people run? They're going to tell you to pick Feat #1. Every time, all the time. No matter what. Because they only consider the average and base everything on it. But if two PCs are using these two feats head-to-head, the PC who picked feat #2 is going to do more damage most of the time i.e. on most rolls. So depending on your play strategy and the sort of combatants you face, Feat #2 could often be your preferred choice. The average is not the end-all, be-all of metrics.
Yeah, this is good example of the problem and applying it beyond where it's useful (even though it might look like it's useful). Because the effect is so extreme of feat #1 in the example, it's not really a useful comparison because we need also need to look at the probability distribution (which is not a part of typical DPR). And that is where people will run into a difference between actual play experience and expectations. Yes, that chance to deal 22 damage is amazing, but it happens only 10% of the time. 90% of the time that will deal less damage. And in many cases the 22 damage would be overkill, or not occur in the fight at all.
And that's why some people calculate DPR over not just one 3 action turn, but over say 4 rounds of combat because it can help to illustrate these issue.
Agreed 100%
Distribution is important.
Not really. Distribution is important if you're supposed to 1 or 2-shot enemies or if you're the only one to deal damage. So it's slightly relevant during the first levels and quickly becomes useless. Average is really the important value.
Your example works only because it's extreme but you won't find actual play example that reproduce it.
| 3 people marked this as a favorite. |
Distribution is important.Not really. Distribution is important if you're supposed to 1 or 2-shot enemies or if you're the only one to deal damage. So it's slightly relevant during the first levels and quickly becomes useless. Average is really the important value.
Your example works only because it's extreme but you won't find actual play example that reproduce it.
Distribution determines standard deviation from the average.
If you have a low standard deviation you can expect the damage you will do to be pretty consistent. But the further the SD is from the average the less reliable that average is about telling you the actual damage you will do.
Distribution determines standard deviation from the average.
If you have a low standard deviation you can expect the damage you will do to be pretty consistent. But the further the SD is from the average the less reliable that average is about telling you the actual damage you will do.
Distribution doesn't impact the number of attacks you will have to make to down an enemy. Maybe your numbers will be random, but the actual important values (time to kill and such) won't be affected once you hit level 3 (at very low level, they are because of the really low hit point pool of level -1 and 0 enemies).
Distribution doesn't impact the number of attacks you will have to make to down an enemy. Maybe your numbers will be random, but the actual important values (time to kill and such) won't be affected once you hit level 3 (at very low level, they are because of the really low hit point pool of level -1 and 0 enemies).Distribution determines standard deviation from the average.
If you have a low standard deviation you can expect the damage you will do to be pretty consistent. But the further the SD is from the average the less reliable that average is about telling you the actual damage you will do.
Wouldn't that only be 100% the case when minimum damage and maximum damage time to kill is the same?
But to your point I believe what makes one PC's damage distribution matter less is that time to kill is never in a vacuum. There are usually 4 PCs contributing to that time to kill value.So really for any party its the average damage of the entire party in a given the round as they choose to apply it that matters, never just the average damage of one pc. When the standard deviation of the party average damage varies the time to kill value that is where there is something to discuss.
Bluemagetim, it's not quite about deviations from the average because these are not single peaks. A single attack is trimodal (with a 0 damage peak at miss, a damage distribution at hit, and another damage distribution at crit). While two attacks has something like 9 distinct possible outcomes (because damage is the same regardless of chance to hit, outcomes like Hit-Miss and Miss-Hit just add together), 8 of them being damage distributions, with some overlap. But this is why 'average' is not so useful; any time you have bimodal or multi-modal distributions, the "average" will be some value that sits between peaks and thus will be an outcome that you...
Point taken, Looking at SD of damage on hit isnt considering outcomes other than success and so it wouldn't be representative of more than what happens on hit alone.
because of this it would be a poor method to compare for example weapons with deadly or fatal to ones without or compare classes with different proficiency with their attacks or key stats at certain levels.It could be useful to compare damage where those things are held constant and the hit distribution is the only difference.
| 1 person marked this as a favorite. |
Distribution is interesting, and can be useful to talk about in terms of how it effects TTK.
But one problem is that generally speaking SD isn't really a factor in how features are designed.
Like with weapons, a d12 weapon has a wider distribution than the d6 weapon... but that's mostly just a quirk of the larger die size. The minimum for both is the same, and the average and peaks are higher on the d12, it's just better but also happens to come with more variance.
There aren't a lot of situations in PF where you get to make meaningful choices about your damage distribution on its own.
So while distribution is a factor in actual play, it's somewhat harder to theorycraft around because it's just a thing that happens, imo.
Distribution is interesting, and can be useful to talk about in terms of how it effects TTK.
But one problem is that generally speaking SD isn't really a factor in how features are designed.
Like with weapons, a d12 weapon has a wider distribution than the d6 weapon... but that's mostly just a quirk of the larger die size. The minimum for both is the same, and the average and peaks are higher on the d12, it's just better but also happens to come with more variance.
There aren't a lot of situations in PF where you get to make meaningful choices about your damage distribution on its own.
So while distribution is a factor in actual play, it's somewhat harder to theorycraft around because it's just a thing that happens, imo.
Agreed, and you're rarely going to have the extreme kinds of distributions that were given as an exaggerated example, to illustrate in an obvious way how it makes a difference. The difference does exist at less extreme ends, but usually isn't going to be terribly obvious.
Damage discussions always interest me.
The DPR calculations work in a game like WoW because the encounters are fairly static and can be planned for. Even in that game certain classes that rely on standing still have their DPS severely reduced in battles with highly mobile enemies.
PF2 has even more variables for DPR and TTK. I was surprised how much a high initiative impacted TTK and DPR. High initiative gives you more opportunities to do damage often leading to killing the target before others get to go or taking the majority of the hit points. Ranged capability also impacts DPR and TTK depending on the battlefield. A melee martial who has to spend a few rounds to close to battle will often do less DPR in fights where a caster can unleash ranged AoE by a huge number because everything will be heavily annihilated prior to the martial getting to start their attack sequence.
The damage variables are many in the DPR and TTK calculation in these games. It changes a lot as you level whereas casters start with a fairly low DPR and TTK, then reach the point where their DPR and TTK is substantially higher than martials and can reach levels where martials are barely needed as the enemies are already dead before they even close the distance.
An assassination team tried to surprise and attack our group, a few eclipse burst were dropped and the martials did clean up duty on the massively wounded targets. There weren't enough hit points left for a high DPR. The spike damage from the casters was untouchable as far as martial DPR goes.
It might be reversed in a boss fight where a fighter or barb lands a few crits on a boss and that fight is over before a caster can do much damage, especially if the boss saves.
DPR is very situational. TTK can often be a matter of having the highest initiative and dealing the most damage to a target leaving it barely standing for the next martial or caster.