D&D General DPR Calculations Wut?

[TwoSix]

Some can take the enemy completely out of the fight. Suggestion for example. That's neither DPR, nor eHP. It's elimination with a single spell, while doing no damage.

No, it’s an increase of eHP to infinity for that encounter.

 

[TwoSix]

I did not understand this.

Basically, when your chance to hit is really low, advantage almost doubles your hit chance. But since your chance to hit is so low, even a doubling of that chance is only equivalent to a +1.

When you hit chance is really high, advantage can't increase it much more, so it's also only worth a +1. The benefit of advantage when your chance to hit is really high is that makes your miss chance go from "maybe once in a session" to "maybe once in an entire campaign".

 

[Tigris]

But if you can get beyond the precision and judgement calls, it's probably the more useful in a broad context.

I am not sure about this part, fully agree with the rest.

In the end not being precise makes it for me too unreliable to be useful in many cases, of course you can do some studies with it, but to make a really good study, and remove all bias etc. is so much work its rarely done in the end. And there are famous examples where there is like a 95% prediction candidate A wins the election and then still B wins.

Basically at the extremes it's less valuable. If you need a 20 to hit, rolling an extra die for advantage doesn't make you much more likely to succeed. If you need a 2 to hit, advantage provides a huge bonus, but you were going to succeed 95% of the time anyway, so it's not that helpful.

What? When you need a 2 to hit advantage does not give you a huge bonus. It gives you the smallest possible bonus bigger than 0. And when needing a 20 to hit the bonus is exactly the same. (And comparatively its almost double).

 

[tetrasodium]

Basically at the extremes it's less valuable. If you need a 20 to hit, rolling an extra die for advantage doesn't make you much more likely to succeed. If you need a 2 to hit, advantage provides a huge bonus, but you were going to succeed 95% of the time anyway, so it's not that helpful.

When the system is designed so that on average a pc only needs an 8 or so before magic items lower that to 7 or less advantage is enormous. Tack on things like the rather common reroll 1s & 2s and you start really borking assumptions like die a averages and what were previously good enough and ballpark shortcuts in the math right out the window.

As was already pointed out it makes a miss implausible and the reroll1s/2s makes the die a average so much higher that it's almost silly to pretend stuff like d6=3.5 is anything but misleading in the extreme

Edit. I messed up the any dice code and it's only trolling the first 1 for each d6. Adding the first 2
Would raise the peak further. Zzzzzz

 

[Zardnaar]

Typo. Smites tend to get

Okay? The plural of anecdote is not data. Divergence occurs. That doesn't mean statistics are bad or wrong.

If your doctor tells you that someone with your symptoms, medical history, and test results has a 90% chance of death within two years unless you get treatment, do you tell your doctor, "Okay but those statistics don't ALWAYS reflect reality!"? Or would you, y'know, listen to them? Because, even though statistics cannot give guaranteed results, they really really do actually work most of the time and if you have a 90% chance of death without treatment FOR GOD'S SAKE SEEK TREATMENT.


Uh...I have no idea what you're saying with this. Like that...is a tautology. When you smite more...you smite more? There must be something missing here.


No, they aren't. It's that tactics and strategy are narrower things. They require more assumptions. They require context.

Notice, though, that you write off any possible mathematical analysis as "white room". As usual, a thought-ender, a conversation-ender, a "that's Just A Horrible Bad Thing, never ever do that". No. You are--flatly--wrong. You are now "white rooming" ABOUT white rooms!


I don't understand this either. You make it sound like it's not a thing...and then you say it is a thing? And common at that? So...which is it?


"I had a person argue a stupid thing, therefore ALL MATHEMATICAL ANALYSIS IS GARBAGE" is not a warranted argument.

Typo smites get used more when you crit.

 

[Zardnaar]

Where white room fails.

Fireball vs Sorcerous Burst.
D8 vs d10. D10 better right. Well SB also has non fie options (20% of MM is resistant/immune) to fire.

On a critical hit say lvl 11 its 3d8vs 3d10. + 5 for dragon sorcerer.

Where the math breaks down is the exploding dice. On crits 6d8 vs 6d10 dice are rolled. Throw in empower spell though.

I'm sure you can math it out but outside a while room on paper fireball might still be better (margibally). I would still use SB due to non fire options. On paper the extra damage wins in reality it often doesnt matter due to overkill. A monster dies in 2 or 3 hits regardless of d8 or d10.

In a real game however SB crits are funny and thats also when you empower them. The explosion happens often enough it turns that two or 3 hits into 1 or 2. On paper firebolt might still be dealing more damage. But its not killing stuff any faster while SB is.

Its also useless for comparing danage to non damage. Control is really good its fairly useless if everyone does it though.

 

[FrogReaver]

I am not sure about this part, fully agree with the rest.

In the end not being precise makes it for me too unreliable to be useful in many cases, of course you can do some studies with it, but to make a really good study, and remove all bias etc. is so much work its rarely done in the end. And there are famous examples where there is like a 95% prediction candidate A wins the election and then still B wins.

No forecast is perfectly accurate. Doesn't mean they are too unreliable to be useful.

 

[FrogReaver]

Where white room fails.

Fireball vs Sorcerous Burst.
D8 vs d10. D10 better right. Well SB also has non fie options (20% of MM is resistant/immune) to fire.

On a critical hit say lvl 11 its 3d8vs 3d10. + 5 for dragon sorcerer.

Where the math breaks down is the exploding dice. On crits 6d8 vs 6d10 dice are rolled. Throw in empower spell though.

All of that can be mathematically modeled as a long term average over time. If you mean to say the long term average doesn't always align with a high impact, extremely low probability event occurring in a small sample size, then I agree, but that doesn't mean the long term average math is breaking down or not being modeled correctly. It's simply the nature of a long term average that you seem to be pushing back against.

 

[FrogReaver]

Basically, when your chance to hit is really low, advantage almost doubles your hit chance. But since your chance to hit is so low, even a doubling of that chance is only equivalent to a +1.

When you hit chance is really high, advantage can't increase it much more, so it's also only worth a +1. The benefit of advantage when your chance to hit is really high is that makes your miss chance go from "maybe once in a session" to "maybe once in an entire campaign".

There's a different perspective on this is well, the one being discussed isn't wrong but it's not the only perspective that is 'true'.

Another perspective is to look at the proportional difference. You are describing the absolute difference.

As an example:
Proportional Difference would say a 5% chance to hit becomes a 9.75% chance to hit with advantage. That's almost a 100% increase in accuracy and thus almost a 100% increase in DPR.

Depending on the context of question one wants to answer it may be more revealing to focus on the proportional increase rather than the absolute.

 

[Zardnaar]

All of that can be mathematical modeled as a long term average over time. If you mean to say the long term average doesn't always align with a high impact, extremely low probability event occurring in a small sample size, then I agree, but that doesn't mean the long term average math is breaking down or not being modeled correctly. It's simply the nature of a long term average that you seem to be pushing back against.

What you cant math out is player experience and the situation if it pays to use empower or smite.

As a player in cant math out the odds of landing a save or suck vs mist of the MM. I know a few critters wisdom saves for example off by heart (assassins, goblins).

A know throwing an AoE targeting wisdom is more likely yo have multiple fails on the roll. Sometimes it will whiff sure. The chance of everything flanuking tends to outweigh the odds of everything succeeding more often than not though.

DPR is a useful guideline. Its more when people invent specific scenario and think math proves them right when the actual problem is the invented scenario.

See my thread on chromatic orb for example. Some people just wouldn't move on the 65% accuracy assumption. We were hitting 85-95% in a real game (plus rerolls).

I did it before the major youtubers.


Basically I was right. You dont get much pushback now if you say CO us really good at least as far as damage dealing spells go.

Also I dont value damage that highly in a vacumn. Its the total package I"m looking at. I like defenses as well. If youre on 0 hp or flunked a wisdom save damage is 0.