D&D General DPR Calculations Wut?

[FrogReaver]

Some of those are incomplete.

Control is often going to be DPR, because the type of control also allows others hit more accurately and/or crit automatically, allowing greater DPR to occur. In those situations, it also becomes future enemy turns removed.

The modeling for that in the framework I just provided seems straightforward.

1. Convert the Control component into enemy turns removed.
2. Convert the DPR-Support component into enemy turns removed.
3. Add them.

That’s the entire point of a unified currency - everything reduces to the same unit, so the combination is clean and lossless.

 

[Blue]

You should have asked me...I agree we have been doing calculations since the beginning of time.

Its why I know that a d6 is 3.5, not 3 or 4.

The math assumptions change per edition, but statistics is statistics.

I'm with this. I aced a probability and statistics class back in college because I'd taught it to myself playing D&D. The effort over years with something I enjoyed frankly was greater than I probably would have put into the class otherwise.

 

[FrogReaver]

I actually understood that.

That puts you in the 99 percentile. Congrats.

 

[EzekielRaiden]

I actually understood that.

If you want to know how much good advantage would do, and you know your chance to succeed (expressed as a decimal), call it P, then you can always calculate the new probability with P+P(1-P). This has its peak at P=0.5, new chance 0.75. It's a parabola, so it declines slowly close to 0.5, and quickly at the edges. Naturally, 0≤P≤1.

Frex, at P=0.9, we get 0.9+0.9(.1), or 0.99. Definitely a great chance to succeed! But also, technically, only worth about a +2 bonus to your chance to succeed. (Same goes for 0.1, as it increases to 0.19).

 

[SkidAce]

If you want to know how much good advantage would do, and you know your chance to succeed (expressed as a decimal), call it P, then you can always calculate the new probability with P+P(1-P). This has its peak at P=0.5, new chance 0.75. It's a parabola, so it declines slowly close to 0.5, and quickly at the edges. Naturally, 0≤P≤1.

Frex, at P=0.9, we get 0.9+0.9(.1), or 0.99. Definitely a great chance to succeed! But also, technically, only worth about a +2 bonus to your chance to succeed. (Same goes for 0.1, as it increases to 0.19).

I did not understand this.

 

[Sacrosanct]

I'm with this. I aced a probability and statistics class back in college because I'd taught it to myself playing D&D. The effort over years with something I enjoyed frankly was greater than I probably would have put into the class otherwise.

I, on the other hand, struggled in statistics class because my basic algebra brain had a hard time figuring standard deviation and p-values.

 

[Tigris]

I, on the other hand, struggled in statistics class because my basic algebra brain had a hard time figuring standard deviation and p-values.

Well the thing is, this is not statistics or at least not what on an university level people understood about statistics. Its just probability and combinatorics. So one does not need p values and standard deviation.


And for many people thats a huge difference. I know several people who are good at math but struggle with probabilities and also many people who are good at math and probabilities but dont like statistics at all (especially when it turns into p hacking etc.)

 

[FrogReaver]

Well the thing is, this is not statistics or at least not what on an university level people understood about statistics. Its just stochastics and probability. So one does not need p values and standard deviation.


And for many people thats a huge difference. I know several people who are good at math but struggle with probabilities and also many people who are good at math and probabilities but dont like statistics at all (especially when it turns into p hacking etc.)

Yea. I liked Math and Probability. I wasn't a huge fan of statistics early on because it wasn't precise and rarely could be related back to anything more than contrived examples. 'Treat it like a normal distribution if any of these hold, it will be close enough'. When you have this distribution, that never really comes up this clean naturally, it means X.

Or, what constitutes a good p-value and other results depending on the specific field you are working in.

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

 

[Maxperson]

I did not understand this.

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.

 

[SkidAce]

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.

Ahhhh.