Poll of the Day > I used to be good at math, these pills are rough

(+) Yeah! (+)

I'm trying to figure this out.

Ok, you have some STRENGTH (S) and some enemy TOUGHNESS (T); there are some fiat numbers in place. If your S = their T your chance to hit is 50%. (roll a 4+)

If your S is over their T, your chance to hit is 66% (roll a 3) and if under it's 33% (roll a 5+). Same goes for double, If your S is twice their T your chance to hit is a 2+ (83%) and if half it's 16% (roll a 6)

...

Yes, this is the rules of Warhammer 40k

...

Anyways.

What I'm trying to figure out is how to granularize this... but it seems almost impossible to granularize; In the rules as written if your S is 4 and their T is 4 you hit at 50%, 4=4' But all ranges between 4.01 and 7.99 would ALSO be in the 3+ category.

I want to granularize it so that when you are at half you still need a 6, if you are equal it's still 50/50 (4+) and double it's still a 2+ (83%) with the 3+ and 5+ options being in there somewhere...

Yaknow?

Math is dumb. Half of it isnt even real!

You lost me after "flat numbers".

Lokarin posted...

In the rules as written if your S is 4 and their T is 4 you hit at 50%, 4=4' But all ranges between 4.01 and 7.99 would ALSO be in the 3+ category.

I want to granularize it so that when you are at half you still need a 6, if you are equal it's still 50/50 (4+) and double it's still a 2+ (83%) with the 3+ and 5+ options being in there somewhere...

did you actually state what you're looking for here? you say "I want to granularize it so that... [everything is still the same]".

Sahuagin posted...

did you actually state what you're looking for here? you say "I want to granularize it so that... [everything is still the same]".

My brain fell out

What I want. Ok, a roll of 1 doesn't count since it's an autofail. I want to take these data points: A 2 = 83% success, or 0.8333, 3 = 0.6666, 4=0.5, 5=0.3333 and 6=0.1666, since that's the normal pattern.... and make a formula so that I could figure out what would happen if you rolled a, say 4.1 or a 2.5 or whatever.

Lokarin posted...

My brain fell out

What I want. Ok, a roll of 1 doesn't count since it's an autofail. I want to take these data points: A 2 = 83% success, or 0.8333, 3 = 0.6666, 4=0.5, 5=0.3333 and 6=0.1666, since that's the normal pattern.... and make a formula so that I could figure out what would happen if you rolled a, say 4.1 or a 2.5 or whatever.

I think part of the problem is that the in-betweens are not defined. they could slope in all kinds of ways (possibly infinite ways).

it looks like there are sites that can do this for you. eg: mycurvefit.com

it seems like it's possible to do this linearly, and the resulting formula is y = (x - 1) / 6

I'm not sure if you have something backwards though saying that 1 = 0% chance of success but 2 = 83% chance of success. it makes more sense if 2 is an 83% chance of failure, and then 1 = 0 still fits linearly.

(1 - 1) / 6 = 0/6 = 0
(2 - 1) / 6 = 1/6 = 0.166...
(3 - 1) / 6 = 2/6 = 0.33...
(4 - 1) / 6 = 3/6 = 0.5
(5 - 1) / 6 = 4/6 = 0.66...
(6 - 1) / 6 = 5/6 = 0.833...

even if you have something slightly different in mind this is probably a good enough start that you could modify it as needed. the key thing here I think is that all of the %s are just x/6 values.

(unless you're looking for this in terms of S and T...?)

Sahuagin posted...

I think part of the problem is that the in-betweens are not defined. they could slope in all kinds of ways (possibly infinite ways).

it looks like there are sites that can do this for you. eg: mycurvefit.com

it seems like it's possible to do this linearly, and the resulting formula is y = (x - 1) / 6

I'm not sure if you have something backwards though saying that 1 = 0% chance of success but 2 = 83% chance of success. it makes more sense if 2 is an 83% chance of failure, and then 1 = 0 still fits linearly.

(1 - 1) / 6 = 0/6 = 0
(2 - 1) / 6 = 1/6 = 0.166...
(3 - 1) / 6 = 2/6 = 0.33...
(4 - 1) / 6 = 3/6 = 0.5
(5 - 1) / 6 = 4/6 = 0.66...
(6 - 1) / 6 = 5/6 = 0.833...

even if you have something slightly different in mind this is probably a good enough start that you could modify it as needed. the key thing here I think is that all of the %s are just x/6 values.

(unless you're looking for this in terms of S and T...?)

Basically, I wanna take the basic warhammer 40k math and make it, like, so you have stats...

ya, i just wanted to make sure something with decimals worked out... using y = (x-1)/6 then if a 4.2 were rolled it would have a success rate of 0.5333, which is in line with what I was trying to figure

...

yaknow when I was back in school I coulda done stuff like this in my sleep :<