kaayru
24/07/2006, 20h26
Basic cooking for stats
http://www.atitd.net/forum/showthread.php?t=10153
Comme d'hab, pas obligé de traduire tout le texte :D Traduisez juste une partie, et petit à petit tout sera traduit, si on s'y met à plusieurs :)
J'ai eu un petit soucis avec le tableau, si quelqu'un sait comment le faire apparaitre convenablement, comme sur le forum atitd.net, hésitez pas !
There are two types of cookpots which you can use when cooking for stats. The only reason to use an iron cookpot is to intentionally nerf your stats. We shall then assume that you're using a copper cookpot.
The game divides all the ingredients in the meal into two sets, which we will call set A and set B. Set A consists of all ingredients whose quantity is tied for the most in the meal. Set B consists of everything else in the meal. Set A is by definition non-empty, though set B could be empty. If the number of ingredients in set A exceeds your cooking level, the game reports the flavors as muddled, and the recipe doesn't work.
Each ingredient corresponds to a point in R^2. Each set B ingredient will react with the set A ingredient to which it is closest. This could cause a set A ingredient to react with several set B ingredients, or none at all, depending on how the meal is designed. If more than one set B ingredient reacts with the same set A ingredient, the meal has a heavy penalty to duration.
If no more than one set B ingredient reacts with a particular set A ingredient, then the set B ingredient will contribute its stats to the meal as follows. The contribution will be (1000 - the distance of the set B ingredient from the set A ingredient it reacts with), times the potency of the set B ingredient, times the base stats of the set B ingredient (-1, 0, or 1 to each stat), provided that the distance between the ingredients is less than 1000. If the distance is at least 1000, the set B ingredient contributes nothing to stats.
Most ingredients give -1 to exactly two stats, though some give -1 to only one stat. It is about as common for an ingredient to give +1 to one stat as to two, and some ingredients give +1 to nothing. I do not know of any ingredients which give -1 to nothing, nor of any which give +1 to more than two stats or -1 to more than two stats.
Set A ingredients contribute a very small amount to stats. With iron cookpots, set B ingredients weren't contributing much either, which is why the stats were so low. For most purposes, the stats of the set A ingredients can be ignored.
While stats are based on distances, just as duration is, they're not perfectly correlated. Depending on the distances from the set A ingredients they react with, one set B ingredient may contribute quite a lot, while another contributes nothing, regardless of the duration of the meal as a whole.
The ratios of the ingredients has no effect on stats, apart from that the ratios determine how the ingredients are partitioned into sets A and B. If a set B ingredient is expensive, you'll probably want to use only one of it, though you can use it to fill servings if it is cheap. In particular, the small portion of duration which depends in the ratios of the various ingredients has no effect on stats.
After all ingredients have contributed their stats, a small positive amount is added to each stat, to make positive stats slightly more positive, and negative stats slightly less negative. Unless the final stats are near zero, that usually disappears into the rounding. The game then multiplies by some constant and then divides each stat by the square root of its magnitude. The stats are then rounded to the nearest integer (or floor or ceiling or integer part or something) to get the final stats.
The constant which the game multiplies by could be taken to be any arbitrary positive constant, by rescaling potency. For various theoretical reasons, however, there is a canonical scale for potency, so it's better to fit the constant around that.
As for the square root, something that gives +25 before taking the square root would give +5 after, while something that gives -25 before would give -5 after. It's not quite right to say that the game merely takes the square root, as that could involve taking the square root of negative numbers.
The square root is why stats don't add like some might expect. If you take four +4 pairs, then they were each about +16 before taking the square root, so they add to about +64, and then when the square root is taken, the observed stats are +8 and not +16. Due to rounding, the observed stats could well end up being near +8 but not equal to it, such as +7 or +9, but could not reach +16.
Here's the table of the base stats of my list of 24 common ingredients:
food str dex end spd con foc per
cabbage 1 -1 -1 1
carrots 1 -1 1 -1
leeks 1 -1 1 -1
onions 1 1 -1 -1
honey 1 -1 -1 1
oil -1 1 1 -1
coconut -1 1 1
c. juice 1 -1 -1
garlic -1 -1
camel 1 -1 -1 1
mutton 1 -1 1 -1
milk -1 1 -1
g. fish 1 -1 -1 1
g. cabbage -1 -1
g. carrots -1 -1
g. onions 1 -1 -1 1
g. garlic 1 -1 1 -1
chromis 1 1 -1
tilapia 1 -1 -1
catfish 1 -1 -1
perch -1 1 -1
carp -1 1 1 -1
abdju 1 -1 1 -1
dates -1 1 -1
I'm not 100% sure that the cheapest ingredients there are exactly right. The fish, grilled foods, dates, and garlic certainly are, though.
You can get the base duration for pairs of those 24 ingredients (which is already 1000-distance, so you don't need to rescale for that) here:
http://www.atitd.net/forum/showthread.php?t=10081
- Quizzical
http://www.atitd.net/forum/showthread.php?t=10153
Comme d'hab, pas obligé de traduire tout le texte :D Traduisez juste une partie, et petit à petit tout sera traduit, si on s'y met à plusieurs :)
J'ai eu un petit soucis avec le tableau, si quelqu'un sait comment le faire apparaitre convenablement, comme sur le forum atitd.net, hésitez pas !
There are two types of cookpots which you can use when cooking for stats. The only reason to use an iron cookpot is to intentionally nerf your stats. We shall then assume that you're using a copper cookpot.
The game divides all the ingredients in the meal into two sets, which we will call set A and set B. Set A consists of all ingredients whose quantity is tied for the most in the meal. Set B consists of everything else in the meal. Set A is by definition non-empty, though set B could be empty. If the number of ingredients in set A exceeds your cooking level, the game reports the flavors as muddled, and the recipe doesn't work.
Each ingredient corresponds to a point in R^2. Each set B ingredient will react with the set A ingredient to which it is closest. This could cause a set A ingredient to react with several set B ingredients, or none at all, depending on how the meal is designed. If more than one set B ingredient reacts with the same set A ingredient, the meal has a heavy penalty to duration.
If no more than one set B ingredient reacts with a particular set A ingredient, then the set B ingredient will contribute its stats to the meal as follows. The contribution will be (1000 - the distance of the set B ingredient from the set A ingredient it reacts with), times the potency of the set B ingredient, times the base stats of the set B ingredient (-1, 0, or 1 to each stat), provided that the distance between the ingredients is less than 1000. If the distance is at least 1000, the set B ingredient contributes nothing to stats.
Most ingredients give -1 to exactly two stats, though some give -1 to only one stat. It is about as common for an ingredient to give +1 to one stat as to two, and some ingredients give +1 to nothing. I do not know of any ingredients which give -1 to nothing, nor of any which give +1 to more than two stats or -1 to more than two stats.
Set A ingredients contribute a very small amount to stats. With iron cookpots, set B ingredients weren't contributing much either, which is why the stats were so low. For most purposes, the stats of the set A ingredients can be ignored.
While stats are based on distances, just as duration is, they're not perfectly correlated. Depending on the distances from the set A ingredients they react with, one set B ingredient may contribute quite a lot, while another contributes nothing, regardless of the duration of the meal as a whole.
The ratios of the ingredients has no effect on stats, apart from that the ratios determine how the ingredients are partitioned into sets A and B. If a set B ingredient is expensive, you'll probably want to use only one of it, though you can use it to fill servings if it is cheap. In particular, the small portion of duration which depends in the ratios of the various ingredients has no effect on stats.
After all ingredients have contributed their stats, a small positive amount is added to each stat, to make positive stats slightly more positive, and negative stats slightly less negative. Unless the final stats are near zero, that usually disappears into the rounding. The game then multiplies by some constant and then divides each stat by the square root of its magnitude. The stats are then rounded to the nearest integer (or floor or ceiling or integer part or something) to get the final stats.
The constant which the game multiplies by could be taken to be any arbitrary positive constant, by rescaling potency. For various theoretical reasons, however, there is a canonical scale for potency, so it's better to fit the constant around that.
As for the square root, something that gives +25 before taking the square root would give +5 after, while something that gives -25 before would give -5 after. It's not quite right to say that the game merely takes the square root, as that could involve taking the square root of negative numbers.
The square root is why stats don't add like some might expect. If you take four +4 pairs, then they were each about +16 before taking the square root, so they add to about +64, and then when the square root is taken, the observed stats are +8 and not +16. Due to rounding, the observed stats could well end up being near +8 but not equal to it, such as +7 or +9, but could not reach +16.
Here's the table of the base stats of my list of 24 common ingredients:
food str dex end spd con foc per
cabbage 1 -1 -1 1
carrots 1 -1 1 -1
leeks 1 -1 1 -1
onions 1 1 -1 -1
honey 1 -1 -1 1
oil -1 1 1 -1
coconut -1 1 1
c. juice 1 -1 -1
garlic -1 -1
camel 1 -1 -1 1
mutton 1 -1 1 -1
milk -1 1 -1
g. fish 1 -1 -1 1
g. cabbage -1 -1
g. carrots -1 -1
g. onions 1 -1 -1 1
g. garlic 1 -1 1 -1
chromis 1 1 -1
tilapia 1 -1 -1
catfish 1 -1 -1
perch -1 1 -1
carp -1 1 1 -1
abdju 1 -1 1 -1
dates -1 1 -1
I'm not 100% sure that the cheapest ingredients there are exactly right. The fish, grilled foods, dates, and garlic certainly are, though.
You can get the base duration for pairs of those 24 ingredients (which is already 1000-distance, so you don't need to rescale for that) here:
http://www.atitd.net/forum/showthread.php?t=10081
- Quizzical