My other thread about comparing monk weapons got me to looking more
closely at the "accepted" method of comparing weapons in general, and
I'm curious where the formula comes from.
What I've seen everywhere is:
(damage x 2 + bonus) / delay
The bigger this is, the better the weapon. For secondary weapons the
bonus is always zero. Fine so far. Now, I've also seen this referred
to as the "modal damage", which feels like it must be a misnomer.
Mathematically, the "mode" of a set of data is the value that appears
most often. What should be more interesting, I'd expect, is the
"mean", i.e. the average amount of damage per hit. (But "average" is
actually an ambiguous term, which can in fact refer to the mean, the
mode, or even the median (the "middle value" in a sorted list of the
damage numbers). But I digress.)
Being both curious and mathematically inclined, I wanted to go back
and understand where this formula comes from. If a weapon is listed
as doing, say, damage 9 with delay 24, what does this really mean?
The "24" means it strikes every 2.4 seconds if autoattack is on,
right? But what damage does it do when it hits? Is it a random value
from 1 to 9, like rolling a 9-sided die? (And if my damage bonus is,
say, +3, it would then be from 4-12 instead of 1-9?) But if that's
the case, the average (mean) damage would be 5 (plus my damage bonus),
and the formula for comparing weapons would be
[((damage + 1) / 2) + bonus] / delay
which isn't at all like the accepted formula. I also found some
discussion giving a formula for maximum damage from a weapon, and
that formula involved STR and weapon skill. But if those can affect
max damage, why don't they enter into calculation for average damage?
(The claim was max damage = the damage number x (STR + skill) / 100.)
Is it just that people's STR and weapon skills vary so much that
computing average damage is considered to be too complicated, and the
accepted formula is a simplification that "comes close enough"? E.g.,
I could see the formula being pretty close for a character with STR +
skill = 400, But at level 30 my STR + skill is only about 270-280,
which could make a big difference in comparing weapons if my average
damage is being scaled by that amount and the formula thinks it's 400.
I could try doing some parses of my own, but my logs don't record
which weapons I'm using (other than the hints given by verbs like
"strike" or "crush"), which makes it hard. (Somehow when I switch
weapons during a fight -- e.g. because I was trying to get some
skillups with 2HB but the fight's going badly so I'm switching to
my higher-skill 1HB or HTH weapons -- I'm too busy to spam a note
into the log to record the weapon change. :-)
So, the short form of the question is: What is the formula for
finding the average (mean) damage per hit? Alternatively, what is
the range of possible damage for a single hit, and are all values
in that range equally likely?
-- Don.
---------------------------------------------------------------------
-- Sukrasisx, Monk 30 on E. Marr Note: If you reply by mail,
-- Terrwini, Druid 17 on E. Marr I'll get to it sooner if you
-- http://www.iCynic.com/~don remove the "hyphen n s"
Empty - 30 Apr 2004 00:21 GMT
> So, the short form of the question is: What is the formula for
> finding the average (mean) damage per hit? Alternatively, what is
> the range of possible damage for a single hit, and are all values
> in that range equally likely?
There is none, AFAIK.
Once upon a time, I wrote a Monk weapon comparison script for use when
choosing weapons. It's not perfect(notably 2hb damage bonuses are wonky at
some levels, all the formulae are from the severely dated Monkly Business
article), but it is still better than nothing in most cases.
Here it is:
http://www.emptiedout.com/cgi-bin/eq/monk-calc
~Empty

Signature
'You're not friends. You'll never be friends. You'll be in love till it
kills you both. You'll fight, and you'll shag, and you'll hate each other
till it makes you quiver, but you'll never be friends. Love isn't brains,
children, it's blood... blood screaming inside you to work its will. I may
be love's bitch, but at least I'm man enough to admit it.'
Spike
Remington Stone - 30 Apr 2004 00:23 GMT
Don Woods said:
}So, the short form of the question is: What is the formula for
}finding the average (mean) damage per hit? Alternatively, what is
}the range of possible damage for a single hit, and are all values
}in that range equally likely?
This may have some information on that:
http://p202.ezboard.com/fpaladinsofnorrathgeneral.showMessage?topicID=19301.topic
[65 Coercer] Zinphandel Chianti <Prism> (Gnome) Ayonae Ro
Remington Stone - 30 Apr 2004 00:30 GMT
Remington Stone said:
}Don Woods said:
}}So, the short form of the question is: What is the formula for
}}finding the average (mean) damage per hit? Alternatively, what is
}}the range of possible damage for a single hit, and are all values
}}in that range equally likely?
}This may have some information on that:
}http://p202.ezboard.com/fpaladinsofnorrathgeneral.showMessage?topicID=19301.topic
Ah, and here's more:
http://forums.crgaming.com/eqbb/viewtopic.php?t=52612
[65 Coercer] Zinphandel Chianti <Prism> (Gnome) Ayonae Ro
Don Woods - 30 Apr 2004 01:44 GMT
> }}So, the short form of the question is: What is the formula for
> }}finding the average (mean) damage per hit? Alternatively, what is
[quoted text clipped - 6 lines]
>
> http://forums.crgaming.com/eqbb/viewtopic.php?t=52612
Thanks for the pointers!
-- Don.
---------------------------------------------------------------------
-- Sukrasisx, Monk 30 on E. Marr Note: If you reply by mail,
-- Terrwini, Druid 17 on E. Marr I'll get to it sooner if you
-- http://www.iCynic.com/~don remove the "hyphen n s"
rja75@bellsouth.net - 30 Apr 2004 19:06 GMT
<snip>
>which isn't at all like the accepted formula. I also found some
>discussion giving a formula for maximum damage from a weapon, and
[quoted text clipped - 5 lines]
>computing average damage is considered to be too complicated, and the
>accepted formula is a simplification that "comes close enough"?
I'm not a matematician, but I think the reason the formula is kept
simple is because the character comparing two weapons isn't going to
have the numbers for his skill and strength changed (at least not by
much) when comparing one sword to another. The damage, the delay, and
the damage bonus are much more important. It's sort of like the
formula for a car's stopping distance being so many car lengths when
moving at a certain speed; while the current wind speed will affect
the actual stopping distance, there's not much need to factor it in
even if someone had a 100% accurate formula.
Another factor is raid buffs often put a melee character at his max
strength any way. That is to say, while the weapon's STR stat may
affect the damage output of the weapon when the character is buffless,
it makes no difference when he is buffed to max STR before wielding
the weapon. And since melee characters use their weapons often, the
formulas are also assuming the characters are at their maximum skill,
or soon will be once they have been in a few fights.