Quote:
Originally Posted by Lokean
Omni, let me recount for you the tale of the most terrifying re-enactment veteran I've ever met.
Bear in mind that I am English, so when we do reenactments there are two main versions: Everyone gets pissed and falls over one another, then gets more pissed afterwards; Everyone wears real armor and wields real weapons that have been blunted or wrapped in leather, broken bones are not unheard of, everyone has fun and then goes home to nurse torn muscles and bruises the size and colour of a Black Magic aubergine (that's the roundish purple egglpant, to you americans).
She did the second sort, she was five foot three and looked like nothing so much as a china dolly. For any relevant physical stat you felt like selecting, she was probably below average. However, nobody ever liked to see her standing even roughly opposite them in the battle line, because she used a two handed sword that was taller than her. And she would run, screaming blue murder, at the opposing battle line and use the weight of the sword and momentum of her charge to throw herself, followed by the spinning arc of death that was the returning upswing of the sword, clean through the ranks...
The point I'm getting at here is that what mediates whether you can actually wield a weapon is less to do with your physical capacity in everyday life (she was weak, clumsy and always exhaused after the fights) than with your skills with the weapon in question.
Further, I'm afraid the natural logarithm is another one of those functions that only exists for positive, real numbers. So if you want to use it you'll make it impossible to use a weapon if you have three points less in the required stats, since this would drop the argument of the function below zero. And you can't even go that low, since you have to keep the argument of the function above one, such that the output is not negative, such that the geometric mean can be used.
In reply to End and JimMorrison, it's actually highly dependant on the value assigned to the weapon as a multiplicative factor, (the arms value, according to Omni). If we assume that the intent is that being just barely capable with the weapon should mean that the 'average fighter' remains average we find that the arms value of the weapon should initially be set (for playtesting) as the arithmetic mean of the requirements for the weapon. This actually means that it suddenly becomes extremely beneficial to the person that invests in a weapon's primary stats, since the natural logarithm increases more rapidly than simple addition at low values.
Endo's two examples actually come out as follows:
Weapon: Warhammer (requires str 8, dex 4, arms value 6)
Psyche 10
Perception 8
Arms (str 7, dex 4) = 3.25 (if his strength was only 6 he couldn't even use the weapon, since ln(e-2) is less than zero)
Geometric Mean = 6.38
Weapon: Warhammer (requires str 8, dex 4, arms value 6)
Psyche 7
Perception 4
Arms (str 10, dex 8) = 12.99
Geometric Mean = 7.14
|
It sounds to me like she had a high Psyche score too
Ok, so given these suggestions, how about this...
Code:
str\
\
dex----------weapon-------|
/ |
skl< |
\ |
per----------tactics------|---attack
/ |
psy/ |
|
random roll---------------|
this is basically a method called 'Path Analysis' used in mathematical modeling. 'weapon' and 'tactics' are called 'Moderators' because they moderate the effect of raw stats and skills into attack.
* str/dex/skl will be a Gompertz Function with an intercetpt of 1, upper bound of 1.5 or 2, and lower bound of 0.
Gompertz Curve
* skl/per/psy will be a natural exponent
* random roll is a.... random roll
the three can be averaged geometrically, and then perhaps multiplied by the 'arms' value.
this makes the weapon have minimum requirements for use, less than required is penalized (down to 0 effect), while more than required provides a bonus (maybe up to 1.5 or 2 times effect)
tactics on the other hand has no requirement, but is a penalty for too low tactics, while an unbound but exponentially more expensive bonus. I was thinking that ideally this bonus should be in practice rarely higher than the maximum weapon bonus (of 1.5 or 2).
I'll need to work out just _how_ the str/dex/skl sigmoid curve is computed, to prevent gaming the system (it may require some weighting method, where weighting of the stat/skl is tied to its requirement level somehow)
Lastly, note that 'skl' enters the computation twice (in weapon handling and tactics). I thought this was intuitive, since training to use a weapon involves not only handling it, but the movement and footwork that goes into placing the energy moving through the weapon onto your opponents body. In fact, it is a little hard to separate out str/dex/skl/tactics/muscle memory and whatever, so in a sense the 'weapon' moderator is also a part of the 'tactics' moderator. 'skl' then serves as the conceptual link between them, and str/dex versus psy/per highlights the muscle memory versus the instinct/perception that guides muscle memory. The basic point is that now skill is more important than other stats. It isn't doubly important, rather it simply influences two separate calculations, so has more weight.
despite the slightly complex path analysis going into this formula, it is still easy to implement. the formula and procedure is universal, all that needs to be assigned is weapon minimum requirements and some 'arms' or 'attack' value. This is what I'm going for, a universal and intuitive formula that doesn't require ad hoc treatment for different weapons.
How does this look? Does it satisfy the issues you guys were bringing up?