I haven't seen a single post about damage after level 20! Jokes aside, assuming you meant <20th level here's one for you. A level 1 fighter can do an average of 7.55 damage per round using a greatsword without damage boosting feats. This assumes a hit rate of 65% and no opportunity attacks are taken.
Oh - right. Well, I have actually seen at least one post about damage above level 20. But ok, what I meant to say was, these discussions tend to be theoretical, and a bit ... I'd say inane, but that would be insulting to those who enjoy them. Um, it would be inane for me because I play a level 6 rogue, and he may reach level 8 or something before the campaign ends.
But I'm not blind to optimizing. He does .. let's see, first strike (bugbear) (2d6), precise strike (squire of solamnia) (1d8), sneak attack (3d6), booming blade (1d8), hunters mark (1d6), plus weapon damage of 1d6+6. Oh, and in case the target is draconic in origin, another 2d6. And if it moves, another 2d8.
To me, that feels like a lot at level 6. I'm sure someone can beat that, easily. But the thought exercise itself seems rather bland to me. Obviously not so bland that I'm not here taking part in the discussion. But it would be more interesting - to me - if it had some practical angle. Like, instead of asking for the highest DPR for a warlock or something - ask for a solid damage dealer for a particular campaign. That never seems to happen.
Rollback Post to RevisionRollBack
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
That's exactly it, though. The small sample sizes are prone to that deviation so an average from which we constantly deviate become unreliable.
So? A higher average rate is still a higher average rate, it's just that the lower the number of rolls, the swingier the results will be.
"So?"
Any single fight a character enters lasts a few rounds. They make maybe 6 attacks. The sample size per fight is so small that the DPR average has a radical deviation in each fight.
The average damage over a day or a week or 20 levels has no ability to offset a single battle. The practical application of calculating DPR doesn't really exist.
That's not to say larger differences in DPR aren't relevant. It's the way we talk about smaller differences that get silly knowing how far off from the average each fight can be.
Any single fight a character enters lasts a few rounds. They make maybe 6 attacks. The sample size per fight is so small that the DPR average has a radical deviation in each fight.
Any single fight a character enters lasts a few rounds. They make maybe 6 attacks. The sample size per fight is so small that the DPR average has a radical deviation in each fight.
7 ± 5 is still more than 6 ± 5.
On the scale of a single fight, DPR does not directly matter --- it only makes a difference if it changes the number of turns the enemy gets.
On the scale of a single fight, DPR does not directly matter --- it only makes a difference if it changes the number of turns the enemy gets.
So, what you're saying is that DPR has a chance of mattering. Generally speaking that's true -- outside of extreme conditions (such as comparing characters of significantly different level), differences in damage just correspond to changes in the probability of various outcomes, rather than introducing new outcomes that wouldn't even be possible for a lower damage character -- but differences in probability do still matter.
When an opponent has 5 hp it doesn't matter if we hit them for 9 or 90 damage.
If a group trying to drop an ogre might have damage matter if they drop the ogre in time to stop an attack from that ogre, but...
Each die roll has an equal chance of landing on any given roll. So, for example, there if the same chance to roll a 1 on a d8 as there is a 4 or a 5 or an 8.
Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average.
So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20.
So we have to not be throwing extra damage away dropping a creature, and we have to accomplish dropping opponents to prevent an action they would have had against the lower DPR, and we have to make enough attacks against those opponents for the law of large numbers to gravitate towards the averages before average DPR kicks in.
The randomness on the low sample size of attacks per combat interferes with the other stipulations for calculated DPR to be effective.
It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages.
That's why DPR is unreliable. It's based on the unreal expectation / flawed premise that these averages are consistently achievable when they obviously are not.
Looking in the Ranger threads, the term "DPR" is thrown around along with numbers.
What does "DPR" mean? Both literally and figuratively, because it seems like a metric.
Just it's just another one of those gamer terms, that mostly multiplayer types tend to bring over in RPG's where it doesn't belong. It means Damage per Round similar to Damage per Second. It doesn't belong in a role playing game any more than "tank", "buff/de-buff" or "no sell". Nerf can be relevant I guess.
Just my opinion of course.
Every one of those except what I assume to be slang is in dnd though.
tank: certain barbarian, fighter, paladin, or multiclass builds
When an opponent has 5 hp it doesn't matter if we hit them for 9 or 90 damage.
If a group trying to drop an ogre might have damage matter if they drop the ogre in time to stop an attack from that ogre, but...
Each die roll has an equal chance of landing on any given roll. So, for example, there if the same chance to roll a 1 on a d8 as there is a 4 or a 5 or an 8.
Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average.
So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20.
So we have to not be throwing extra damage away dropping a creature, and we have to accomplish dropping opponents to prevent an action they would have had against the lower DPR, and we have to make enough attacks against those opponents for the law of large numbers to gravitate towards the averages before average DPR kicks in.
The randomness on the low sample size of attacks per combat interferes with the other stipulations for calculated DPR to be effective.
It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages.
That's why DPR is unreliable. It's based on the unreal expectation / flawed premise that these averages are consistently achievable when they obviously are not.
That is very sketchy math.
"Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average." - I suppose that is true if you are talking about a d8 and consider 4 and 5 "that average".
"So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20." - Also true and also does not matter. Rolling a 4 and rolling a 5 is the same here as rolling a 1 and rolling an 8.
"It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages." - Denying that dpr is important is like going to a slot machine instead of a roulette table because both are random. You can make your choices but it is more logical both to want a higher dpr and to choose the roulette table. (making the reasonable assumption that the roulette has a higher average return)
When an opponent has 5 hp it doesn't matter if we hit them for 9 or 90 damage.
If a group trying to drop an ogre might have damage matter if they drop the ogre in time to stop an attack from that ogre, but...
Each die roll has an equal chance of landing on any given roll. So, for example, there if the same chance to roll a 1 on a d8 as there is a 4 or a 5 or an 8.
Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average.
So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20.
So we have to not be throwing extra damage away dropping a creature, and we have to accomplish dropping opponents to prevent an action they would have had against the lower DPR, and we have to make enough attacks against those opponents for the law of large numbers to gravitate towards the averages before average DPR kicks in.
The randomness on the low sample size of attacks per combat interferes with the other stipulations for calculated DPR to be effective.
It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages.
That's why DPR is unreliable. It's based on the unreal expectation / flawed premise that these averages are consistently achievable when they obviously are not.
That is very sketchy math.
"Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average." - I suppose that is true if you are talking about a d8 and consider 4 and 5 "that average".
"So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20." - Also true and also does not matter. Rolling a 4 and rolling a 5 is the same here as rolling a 1 and rolling an 8.
"It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages." - Denying that dpr is important is like going to a slot machine instead of a roulette table because both are random. You can make your choices but it is more logical both to want a higher dpr and to choose the roulette table. (making the reasonable assumption that the roulette has a higher average return)
It's not sketchy. It's how probability works. Each number on the die has an equal chance to be rolled. I used a d 8 example.
D8 -- 12.5% chance for each number. Average is 4.5 so a 4 or a 5 approaches the average, which is a 25% chance. Any other number diverges from the average, which is a 75% likelihood.
D10 -- 10% chance for each number. Average is 5.5 so a 5 or a 6 approaches the average, which is a 20% chance. Any other number diverges from the average, which is an 80% likelihood.
That's before the d20 attack.
So any instance we roll the die it's actually more likely to diverge from the average than to approach that average. The law of large numbers will gravitate towards the averages, which cannot happen because the sample size is too small, which is the point.
This is also mitigated by static bonuses instead of random bonuses, which is why static bonuses are better.
When an opponent has 5 hp it doesn't matter if we hit them for 9 or 90 damage.
If a group trying to drop an ogre might have damage matter if they drop the ogre in time to stop an attack from that ogre, but...
Each die roll has an equal chance of landing on any given roll. So, for example, there if the same chance to roll a 1 on a d8 as there is a 4 or a 5 or an 8.
Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average.
So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20.
So we have to not be throwing extra damage away dropping a creature, and we have to accomplish dropping opponents to prevent an action they would have had against the lower DPR, and we have to make enough attacks against those opponents for the law of large numbers to gravitate towards the averages before average DPR kicks in.
The randomness on the low sample size of attacks per combat interferes with the other stipulations for calculated DPR to be effective.
It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages.
That's why DPR is unreliable. It's based on the unreal expectation / flawed premise that these averages are consistently achievable when they obviously are not.
That is very sketchy math.
"Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average." - I suppose that is true if you are talking about a d8 and consider 4 and 5 "that average".
"So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20." - Also true and also does not matter. Rolling a 4 and rolling a 5 is the same here as rolling a 1 and rolling an 8.
"It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages." - Denying that dpr is important is like going to a slot machine instead of a roulette table because both are random. You can make your choices but it is more logical both to want a higher dpr and to choose the roulette table. (making the reasonable assumption that the roulette has a higher average return)
It's not sketchy. It's how probability works. Each number on the die has an equal chance to be rolled. I used a d 8 example.
D8 -- 12.5% chance for each number. Average is 4.5 so a 4 or a 5 approaches the average, which is a 25% chance. Any other number diverges from the average, which is a 75% likelihood.
D10 -- 10% chance for each number. Average is 5.5 so a 5 or a 6 approaches the average, which is a 20% chance. Any other number diverges from the average, which is an 80% likelihood.
That's before the d20 attack.
So any instance we roll the die it's actually more likely to diverge from the average than to approach that average. The law of large numbers will gravitate towards the averages, which cannot happen because the sample size is too small, which is the point.
This is also mitigated by static bonuses instead of random bonuses, which is why static bonuses are better.
The thing is that since, as you noted, it is equally likely to be greater than the average as be less. Therefore, with the small sample size, you are just as likely to see disproportionately high results as low ones. Since you are using larger dice, those above average results will be higher. Also, as I noted, you don't need to roll average to see average results. If you roll a 1 and an 8, that is equivalent to rolling a 4 and a 5.
It's not sketchy. It's how probability works. Each number on the die has an equal chance to be rolled.
While true, that doesn't mean averages don't matter. Let's consider a simple example: we have two combatants, both wielding longswords; the only difference is that one of them has a strength of 15 and the other has a strength of 16. They're fighting against guards (AC 16, HP 11)
The strength 15 character has a 40% chance to hit for 1d8+2 and a 5% chance to crit for 2d8+2; total expected dpr 3.15.
The strength 16 character has a 45% chance to hit for 1d8+3 and a 5% chance to crit for 2d8+3; total expected dpr 3.97 (26% higher)
But... you're right that dpr isn't the best measure. Instead, I wrote a stupid program to try and kill a guard a thousand times, and recorded how many attacks it took to kill a guard.
When an opponent has 5 hp it doesn't matter if we hit them for 9 or 90 damage.
If a group trying to drop an ogre might have damage matter if they drop the ogre in time to stop an attack from that ogre, but...
Each die roll has an equal chance of landing on any given roll. So, for example, there if the same chance to roll a 1 on a d8 as there is a 4 or a 5 or an 8.
Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average.
So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20.
So we have to not be throwing extra damage away dropping a creature, and we have to accomplish dropping opponents to prevent an action they would have had against the lower DPR, and we have to make enough attacks against those opponents for the law of large numbers to gravitate towards the averages before average DPR kicks in.
The randomness on the low sample size of attacks per combat interferes with the other stipulations for calculated DPR to be effective.
It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages.
That's why DPR is unreliable. It's based on the unreal expectation / flawed premise that these averages are consistently achievable when they obviously are not.
That is very sketchy math.
"Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average." - I suppose that is true if you are talking about a d8 and consider 4 and 5 "that average".
"So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20." - Also true and also does not matter. Rolling a 4 and rolling a 5 is the same here as rolling a 1 and rolling an 8.
"It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages." - Denying that dpr is important is like going to a slot machine instead of a roulette table because both are random. You can make your choices but it is more logical both to want a higher dpr and to choose the roulette table. (making the reasonable assumption that the roulette has a higher average return)
It's not sketchy. It's how probability works. Each number on the die has an equal chance to be rolled. I used a d 8 example.
D8 -- 12.5% chance for each number. Average is 4.5 so a 4 or a 5 approaches the average, which is a 25% chance. Any other number diverges from the average, which is a 75% likelihood.
D10 -- 10% chance for each number. Average is 5.5 so a 5 or a 6 approaches the average, which is a 20% chance. Any other number diverges from the average, which is an 80% likelihood.
That's before the d20 attack.
So any instance we roll the die it's actually more likely to diverge from the average than to approach that average. The law of large numbers will gravitate towards the averages, which cannot happen because the sample size is too small, which is the point.
This is also mitigated by static bonuses instead of random bonuses, which is why static bonuses are better.
The thing is that since, as you noted, it is equally likely to be greater than the average as be less. Therefore, with the small sample size, you are just as likely to see disproportionately high results as low ones. Since you are using larger dice, those above average results will be higher. Also, as I noted, you don't need to roll average to see average results. If you roll a 1 and an 8, that is equivalent to rolling a 4 and a 5.
You are just as like to see high as low. My claim is the average is unreliable short term like a fight. Where high and low matter is that last attack that drops an opponent. A large number can do the same damage as an average number hitting 0 hit points, but an average number cannot do the same damage as a high number or a low number accomplish the same as a high or average number.
A 1 and an 8 does average the same as a 4 and a 5 but you're still missing the likelihood of hitting that. If I roll a 1 there's a 12.5% chance to roll an 8 next and an 87.5% chance to roll any other number than an 8 so there's an equal chance for another 1 as there is for a matching 8.
The law of large numbers means if we roll a d8 800 times we'll probably roll each number 100 times and then we'll have our 4.5 average.
The more rolls we make the more we gravitate towards that average. The less rolls we make the greater the risk of deviation from that average. That's one of the inherent flaws we have with calculating DPR. An average calculation requires a lot of rolls and we do make enough rolls to make a claim to the accuracy of achieving that average.
The other big inherent flaw is trying to assume all the variables impacting DPR.
A monk is closer to average than a bard because a lot of attacks equals a lot of rolls. Sneak attack has a lot of d6's eventually and that would also become closer to average.
The thing is that since, as you noted, it is equally likely to be greater than the average as be less. Therefore, with the small sample size, you are just as likely to see disproportionately high results as low ones. Since you are using larger dice, those above average results will be higher. Also, as I noted, you don't need to roll average to see average results. If you roll a 1 and an 8, that is equivalent to rolling a 4 and a 5.
You are just as like to see high as low. My claim is the average is unreliable short term like a fight. Where high and low matter is that last attack that drops an opponent. A large number can do the same damage as an average number hitting 0 hit points, but an average number cannot do the same damage as a high number or a low number accomplish the same as a high or average number.
A 1 and an 8 does average the same as a 4 and a 5 but you're still missing the likelihood of hitting that. If I roll a 1 there's a 12.5% chance to roll an 8 next and an 87.5% chance to roll any other number than an 8 so there's an equal chance for another 1 as there is for a matching 8.
The law of large numbers means if we roll a d8 800 times we'll probably roll each number 100 times and then we'll have our 4.5 average.
The more rolls we make the more we gravitate towards that average. The less rolls we make the greater the risk of deviation from that average. That's one of the inherent flaws we have with calculating DPR. An average calculation requires a lot of rolls and we do make enough rolls to make a claim to the accuracy of achieving that average.
The other big inherent flaw is trying to assume all the variables impacting DPR.
A monk is closer to average than a bard because a lot of attacks equals a lot of rolls. Sneak attack has a lot of d6's eventually and that would also become closer to average.
The average is unreliable in very short fights. That is correct. However, the average damage will tell you what you can expect from using that ability over multiple fights. You don't need to get the exact average to see the effects of a higher dpr.
This argument is inane. Building a character that can do more damage per round means they are more likely to deal more damage than another character. Ignoring DPR because there's probably is throwing the baby out with the bath water. To increase your dpr you have to increase your chance to hit or the amount of damage you do. Both of which will make your character better at dealing damage than non optimized characters reliably.
The only valid criticisms that I can think of when using DPR as a metric to judge effectiveness of damage dealing are:
non-standardized and unclear assumptions people use; a lack of ability to show if the DPR is based on one big attack or multiple smaller attacks; a lack of ability to show the variation on the damage rolls; the lack of ability to differentiate between more hit chance and higher damage.
Even with these shortcomings it still is an excellent indication of the effectiveness of a character's damage dealing in combat. To think otherwise is just silly.
The words Buff/de-buff are nowhere in any core D&D books. Which is my point.
But you realise there are buffs, and debuffs, in all D&D books, right? No matter what we chose to call them.
Rollback Post to RevisionRollBack
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
Is this really how some measure the worth of a character? It does more damage? This is why many believe modern D&D is approached more like a video game than a TRPG in which actual characterization should signify more. This is why, while I do play in two 5E games, I run a different version at my table, and why whenever I introduce newcomers to the hobby I use this or an older edition. Because I want them to experience a TRPG. What playing one felt like before the advent of video games. And not a game that has become the table-top equivalent of one in which people spend hours building their avatar before they then hit START.
It isn't for no reason that the Wikipedia page about power gaming, min maxing, or optimization is scathing in its criticism of the practice. It reduces characters to flavorless manques that are more about numbers than about who they are and the game to little more than a beat 'em up with fantasy coating. It's an insult to any serious DM who puts in hours of work to produce solid world building that is supposed to serve as more than a mere backdrop to fights.
Is this really how some measure the worth of a character? It does more damage? This is why many believe modern D&D is approached more like a video game than a TRPG in which actual characterization should signify more. This is why, while I do play in two 5E games, I run a different version at my table, and why whenever I introduce newcomers to the hobby I use this or an older edition. Because I want them to experience a TRPG. What playing one felt like before the advent of video games. And not a game that has become the table-top equivalent of one in which people spend hours building their avatar before they then hit START.
It isn't for no reason that the Wikipedia page about power gaming, min maxing, or optimization is scathing in its criticism of the practice. It reduces characters to flavorless manques that are more about numbers than about who they are and the game to little more than a beat 'em up with fantasy coating. It's an insult to any serious DM who puts in hours of work to produce solid world building that is supposed to serve as more than a mere backdrop to fights.
It's not unreasonable.
I've played (not often) with people who took roleplaying seriously, going all in on character aspects that really were not relevant to anyone but themselves, and it wasn't ... I'm sorry, but it just wasn't enjoyable. Hearing this one guy go on, at length, about his characters absolutely awful past (I respect how imagininative he was about creating a background story that hurt to listen to - but there's a limit to how much time at the table should be dedicated to this) - was just painful.
And meanwhile, D&D and other games beside rely to a great extent on fighting monsters. Having a measure of how good a character is at that, and also letting that weigh in on how good the character is, isn't insane. I sorta checks out. You roll a rogue, you kinda wanna stab some kidneys, and you want that to hurt. If it doesn't, that detracts from the value of that character.
Rollback Post to RevisionRollBack
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
Is this really how some measure the worth of a character? It does more damage? This is why many believe modern D&D is approached more like a video game than a TRPG in which actual characterization should signify more. This is why, while I do play in two 5E games, I run a different version at my table, and why whenever I introduce newcomers to the hobby I use this or an older edition. Because I want them to experience a TRPG. What playing one felt like before the advent of video games. And not a game that has become the table-top equivalent of one in which people spend hours building their avatar before they then hit START.
It isn't for no reason that the Wikipedia page about power gaming, min maxing, or optimization is scathing in its criticism of the practice. It reduces characters to flavorless manques that are more about numbers than about who they are and the game to little more than a beat 'em up with fantasy coating. It's an insult to any serious DM who puts in hours of work to produce solid world building that is supposed to serve as more than a mere backdrop to fights.
Way to misrepresent what others say. I don't think anyone in this thread has said that high DPR makes for a better overall character. The only argument for DPR that has been made is that DPR is a good metric for evaluating damage output and that characters with high DPR are more effective damage dealers in combat.
Optimizing a character is not all about combat and it does not make it flavorless. If anything it makes the character more flavorful because it becomes better at its role, usually at the expense of something else. If your character's role is to be the face of the party then you often forfeit some combat or exploration abilities to be better at social interactions. If it's role is to deal damage then you should consider your DPR so that you can do the most damage per round possible even if that means you aren't as good at other things.
If you enjoy playing a fighter that can't hit the broad side of a barn or a rogue who can't sneak past a guard then play an unoptimized character to your heart's content and leave me out of it. I want my characters to do the things they're created to do.
Which is why DPR isn’t all that useful in reality.
Wisea$$ DM and Player since 1979.
If you specifically define dpr in a way that isn't useful, it isn't useful. That's why people mostly don't define dpr that way.
Oh - right. Well, I have actually seen at least one post about damage above level 20. But ok, what I meant to say was, these discussions tend to be theoretical, and a bit ... I'd say inane, but that would be insulting to those who enjoy them. Um, it would be inane for me because I play a level 6 rogue, and he may reach level 8 or something before the campaign ends.
But I'm not blind to optimizing. He does .. let's see, first strike (bugbear) (2d6), precise strike (squire of solamnia) (1d8), sneak attack (3d6), booming blade (1d8), hunters mark (1d6), plus weapon damage of 1d6+6. Oh, and in case the target is draconic in origin, another 2d6. And if it moves, another 2d8.
To me, that feels like a lot at level 6. I'm sure someone can beat that, easily. But the thought exercise itself seems rather bland to me. Obviously not so bland that I'm not here taking part in the discussion. But it would be more interesting - to me - if it had some practical angle. Like, instead of asking for the highest DPR for a warlock or something - ask for a solid damage dealer for a particular campaign. That never seems to happen.
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
"So?"
Any single fight a character enters lasts a few rounds. They make maybe 6 attacks. The sample size per fight is so small that the DPR average has a radical deviation in each fight.
The average damage over a day or a week or 20 levels has no ability to offset a single battle. The practical application of calculating DPR doesn't really exist.
That's not to say larger differences in DPR aren't relevant. It's the way we talk about smaller differences that get silly knowing how far off from the average each fight can be.
7 ± 5 is still more than 6 ± 5.
On the scale of a single fight, DPR does not directly matter --- it only makes a difference if it changes the number of turns the enemy gets.
So, what you're saying is that DPR has a chance of mattering. Generally speaking that's true -- outside of extreme conditions (such as comparing characters of significantly different level), differences in damage just correspond to changes in the probability of various outcomes, rather than introducing new outcomes that wouldn't even be possible for a lower damage character -- but differences in probability do still matter.
When an opponent has 5 hp it doesn't matter if we hit them for 9 or 90 damage.
If a group trying to drop an ogre might have damage matter if they drop the ogre in time to stop an attack from that ogre, but...
Each die roll has an equal chance of landing on any given roll. So, for example, there if the same chance to roll a 1 on a d8 as there is a 4 or a 5 or an 8.
Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average.
So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20.
So we have to not be throwing extra damage away dropping a creature, and we have to accomplish dropping opponents to prevent an action they would have had against the lower DPR, and we have to make enough attacks against those opponents for the law of large numbers to gravitate towards the averages before average DPR kicks in.
The randomness on the low sample size of attacks per combat interferes with the other stipulations for calculated DPR to be effective.
It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages.
That's why DPR is unreliable. It's based on the unreal expectation / flawed premise that these averages are consistently achievable when they obviously are not.
The words Buff/de-buff are nowhere in any core D&D books. Which is my point.
That is very sketchy math.
"Each of those rolls has a 1 in 4 chance of rolling that average and a 3 in 4 chance of rolling above or below that average." - I suppose that is true if you are talking about a d8 and consider 4 and 5 "that average".
"So if a fight is 2 rounds long and a fighter has 2 attacks each round that's 4 chances that each roll will hit that 1 in 4 chance. That's before also applying the d20." - Also true and also does not matter. Rolling a 4 and rolling a 5 is the same here as rolling a 1 and rolling an 8.
"It's akin to gamblers thinking they have a system to crack the odds when the system will continue to be random at any point in time. The randomness is too high for the number of chances to meet those averages." - Denying that dpr is important is like going to a slot machine instead of a roulette table because both are random. You can make your choices but it is more logical both to want a higher dpr and to choose the roulette table. (making the reasonable assumption that the roulette has a higher average return)
It's not sketchy. It's how probability works. Each number on the die has an equal chance to be rolled. I used a d 8 example.
D8 -- 12.5% chance for each number. Average is 4.5 so a 4 or a 5 approaches the average, which is a 25% chance. Any other number diverges from the average, which is a 75% likelihood.
D10 -- 10% chance for each number. Average is 5.5 so a 5 or a 6 approaches the average, which is a 20% chance. Any other number diverges from the average, which is an 80% likelihood.
That's before the d20 attack.
So any instance we roll the die it's actually more likely to diverge from the average than to approach that average. The law of large numbers will gravitate towards the averages, which cannot happen because the sample size is too small, which is the point.
This is also mitigated by static bonuses instead of random bonuses, which is why static bonuses are better.
The thing is that since, as you noted, it is equally likely to be greater than the average as be less. Therefore, with the small sample size, you are just as likely to see disproportionately high results as low ones. Since you are using larger dice, those above average results will be higher. Also, as I noted, you don't need to roll average to see average results. If you roll a 1 and an 8, that is equivalent to rolling a 4 and a 5.
While true, that doesn't mean averages don't matter. Let's consider a simple example: we have two combatants, both wielding longswords; the only difference is that one of them has a strength of 15 and the other has a strength of 16. They're fighting against guards (AC 16, HP 11)
But... you're right that dpr isn't the best measure. Instead, I wrote a stupid program to try and kill a guard a thousand times, and recorded how many attacks it took to kill a guard.
Strength 15
Strength 16
Which is actually... 25% faster, rather than 26%. Oh no, DPR is a terrible and useless measure!
You are just as like to see high as low. My claim is the average is unreliable short term like a fight. Where high and low matter is that last attack that drops an opponent. A large number can do the same damage as an average number hitting 0 hit points, but an average number cannot do the same damage as a high number or a low number accomplish the same as a high or average number.
A 1 and an 8 does average the same as a 4 and a 5 but you're still missing the likelihood of hitting that. If I roll a 1 there's a 12.5% chance to roll an 8 next and an 87.5% chance to roll any other number than an 8 so there's an equal chance for another 1 as there is for a matching 8.
The law of large numbers means if we roll a d8 800 times we'll probably roll each number 100 times and then we'll have our 4.5 average.
The more rolls we make the more we gravitate towards that average. The less rolls we make the greater the risk of deviation from that average. That's one of the inherent flaws we have with calculating DPR. An average calculation requires a lot of rolls and we do make enough rolls to make a claim to the accuracy of achieving that average.
The other big inherent flaw is trying to assume all the variables impacting DPR.
A monk is closer to average than a bard because a lot of attacks equals a lot of rolls. Sneak attack has a lot of d6's eventually and that would also become closer to average.
The average is unreliable in very short fights. That is correct. However, the average damage will tell you what you can expect from using that ability over multiple fights. You don't need to get the exact average to see the effects of a higher dpr.
This argument is inane. Building a character that can do more damage per round means they are more likely to deal more damage than another character. Ignoring DPR because there's probably is throwing the baby out with the bath water. To increase your dpr you have to increase your chance to hit or the amount of damage you do. Both of which will make your character better at dealing damage than non optimized characters reliably.
The only valid criticisms that I can think of when using DPR as a metric to judge effectiveness of damage dealing are:
non-standardized and unclear assumptions people use; a lack of ability to show if the DPR is based on one big attack or multiple smaller attacks; a lack of ability to show the variation on the damage rolls; the lack of ability to differentiate between more hit chance and higher damage.
Even with these shortcomings it still is an excellent indication of the effectiveness of a character's damage dealing in combat. To think otherwise is just silly.
But you realise there are buffs, and debuffs, in all D&D books, right? No matter what we chose to call them.
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
Is this really how some measure the worth of a character? It does more damage? This is why many believe modern D&D is approached more like a video game than a TRPG in which actual characterization should signify more. This is why, while I do play in two 5E games, I run a different version at my table, and why whenever I introduce newcomers to the hobby I use this or an older edition. Because I want them to experience a TRPG. What playing one felt like before the advent of video games. And not a game that has become the table-top equivalent of one in which people spend hours building their avatar before they then hit START.
It isn't for no reason that the Wikipedia page about power gaming, min maxing, or optimization is scathing in its criticism of the practice. It reduces characters to flavorless manques that are more about numbers than about who they are and the game to little more than a beat 'em up with fantasy coating. It's an insult to any serious DM who puts in hours of work to produce solid world building that is supposed to serve as more than a mere backdrop to fights.
It's not unreasonable.
I've played (not often) with people who took roleplaying seriously, going all in on character aspects that really were not relevant to anyone but themselves, and it wasn't ... I'm sorry, but it just wasn't enjoyable. Hearing this one guy go on, at length, about his characters absolutely awful past (I respect how imagininative he was about creating a background story that hurt to listen to - but there's a limit to how much time at the table should be dedicated to this) - was just painful.
And meanwhile, D&D and other games beside rely to a great extent on fighting monsters. Having a measure of how good a character is at that, and also letting that weigh in on how good the character is, isn't insane. I sorta checks out. You roll a rogue, you kinda wanna stab some kidneys, and you want that to hurt. If it doesn't, that detracts from the value of that character.
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
Way to misrepresent what others say. I don't think anyone in this thread has said that high DPR makes for a better overall character. The only argument for DPR that has been made is that DPR is a good metric for evaluating damage output and that characters with high DPR are more effective damage dealers in combat.
Optimizing a character is not all about combat and it does not make it flavorless. If anything it makes the character more flavorful because it becomes better at its role, usually at the expense of something else. If your character's role is to be the face of the party then you often forfeit some combat or exploration abilities to be better at social interactions. If it's role is to deal damage then you should consider your DPR so that you can do the most damage per round possible even if that means you aren't as good at other things.
If you enjoy playing a fighter that can't hit the broad side of a barn or a rogue who can't sneak past a guard then play an unoptimized character to your heart's content and leave me out of it. I want my characters to do the things they're created to do.
Wikipedia...