Halls of Mist: super-simple skills?

[buzzkill]

... for example it is similar to taking the probability of a disjunction, to De Morgan's laws in logic, to harmonic means, and to resistance in parallel.

!Bravo!

[Mikko Lehtinen]

Just to point out that this is equivalent to the multiplicative vulnerabilities Nick and buzzkill describe. Yours is an elegant description, but it can be useful to have the end effect rolled into one number to make it clear to the player.

Avernum uses this system for all resists and even for armour class.

Since all my skills are percentage based, I could use this system with them all. At the moment I think most items just add bonuses straight to the skill score. (For many skills it doesn't matter if the success percentage grows to 100%. For some skills I even have effects for scores beyond 100%. Weapon skills beyond 100% start increasing your critical chance.)

[Nick]

Yes, it's really the damage fractions that are multiplicative.

Note that for all my praise of it, Sil uses a system where you take 1/(n+1) of the damage with n sources of resistance, as this fit my purposes a little better.

Funnily enough, the O system (which FA started out with) is an approximation to this.

The percentage system makes intuitive sense (after some thought), but it is less elegant and messier to code than the 1/(n+1).

[half]

They stack multiplicatively.

I agree that this is an excellent way for resistances and some other things to work. It is a bit tricky to work out what to call it. Multiplicatively kind of implies that two 50% resistances make a 25% resistance. The method you use comes up fairly often in systems, for example it is similar to taking the probability of a disjunction, to De Morgan's laws in logic, to harmonic means, and to resistance in parallel. I like to call it 'co-multiplicative', since you take the complements of the numbers, multiply these and take the complement of the result. Effectively it is that the complement is multiplicative (in this case 'vulnerability').

Note that for all my praise of it, Sil uses a system where you take 1/(n+1) of the damage with n sources of resistance, as this fit my purposes a little better. I think that your system is more general and often better.

[Scatha]

Sleeping monsters do not move or attack. Instead, on each of their turns, nearby sleeping monsters force you to make a Stealth check. If the noise you make disturbs them enough (the threshold depending on the monster) they wake up.

Your character and all her Stealth equipment have their own Stealth skill, and they all roll their skill checks separately. You avoid disturbing a monster if any of the checks succeed.

Just to point out that this is equivalent to the multiplicative vulnerabilities Nick and buzzkill describe. Yours is an elegant description, but it can be useful to have the end effect rolled into one number to make it clear to the player.

[buzzkill]

They stack multiplicatively. It's best to think of it as how much damage gets through:
If you have a 40% RFire item and a 30% RFire item, you take 60% damage because of the first, then 70% of that because of the second. So in total you take 0.6x0.7 = 0.42 or 42% damage, and overall you have 58% RFire.

What he said. It eliminates the need for a hard cap as each additional source of stealth would give a diminished return. 50% + 50% =75% add another 50% and your only up to 87%. another 50%=93%, not that 50% stealth boosts should be at all common. I've always thought that the % resist system is superior to V's.

[Nick]

Could you elaborate a bit? How do resists stack in FA?

They stack multiplicatively. It's best to think of it as how much damage gets through:
If you have a 40% RFire item and a 30% RFire item, you take 60% damage because of the first, then 70% of that because of the second. So in total you take 0.6x0.7 = 0.42 or 42% damage, and overall you have 58% RFire.

In the case of Stealth, I guess you would think of each item reducing the noise you make by a fixed percentage, which kind of works; I actually think your idea makes a lot of intuitive sense, though.

[Mikko Lehtinen]

Would this be an elegant implementation of Stealth?

Sleeping monsters do not move or attack. Instead, on each of their turns, nearby sleeping monsters force you to make a Stealth check. If the noise you make disturbs them enough (the threshold depending on the monster) they wake up.

Your character and all her Stealth equipment have their own Stealth skill, and they all roll their skill checks separately. You avoid disturbing a monster if any of the checks succeed.

[Mikko Lehtinen]

Wouldn't it be easier to just use percentages like FA does with resists?

Could you elaborate a bit? How do resists stack in FA?

If you mean that boots have their own Stealth percentage, and both boots and their owner make their own skill check, yes, that would work. "More turns between Stealth checks" is just another way of saying that. Which one would be easier to understand?

[buzzkill]

Wouldn't it be easier to just use percentages like FA does with resists?

[Mikko Lehtinen]

I figured out how to handle Stealth items in a world with Stealth capped at 90%.

They don't give bonuses to your Stealth score. Instead, each stealth equipment "increases the average time between Stealth checks by one turn".

This is handled as another random chance to pass Stealth checks without actually rolling Stealth. 1/2 chance for one item, 2/3 for two, 3/4 for three, and so on.

This would make stealth items about as good for warriors with Stealth 20% and Rogues with Stealth 90% -- much like in the original system. Stockpiling stealth items would work without breaking the game.

[Mikko Lehtinen]

Before reading Derakon's really helpful posts I did some very quick and dirty numbers myself. Seems I was quite close to reality.

Stealth skill    Stealth chance
20               99%
17               98%
13               95%
10               90%
9                87.5%
7                80%
4                60%
2                36%
0                0%

I'm surprised that Stealth skill would probably work quite well with pure, unmodified 1d100 skill checks. It's easy to give race/class combinations similar percentages to what they have in the original system. (Rogue is the only class the develops Stealth with experience.)

Equipment that increase Stealth are the only problem. There would either need to be a cap at 90% or some sort of diminishing returns. It would not be necessary to have any Stealth equipment -- we don't have any Saving Throw or Magic Device equipment, either.

At the moment Grippli Rogue has Stealth 12 (about 93%) at level 50. Grippli Ranger has stealth 7 (80%) at any level. A cap at 90% would not break anything too badly.

[ekolis]

"<<" operator is bit shift operator. How does that result exponents is beyond me. Some magic is applied there.

Bit shift does exponents (but only powers of 2) because shifting a number to the left one bit is equivalent to doubling it.

[Derakon]

I suffered through the forum's search feature to find the post where I described how stealth works and another one where I described the values of various stealth levels.

[Mikko Lehtinen]

Shift happens.