English

I haven't had time to look into KC yet, however I changed rounding multiple times to match border cases better. If you have the patience to try this formula / recipe with different versions of kc and one version is closer to the ingame results I could investigate my changes and report back. I personally think +/- 1% is sufficient enough, but I know it is important for some users.

arc

Im not too worried about anything other then dmg, speed, parry for weaps for example and they seem to work fine, kc had the precraft more accurate then the ingame preview after all :p
we'd have to make recipes with decimals like .01 .99 to see where things drop off or rise in the result.
but should we really bother?

Below are the precraft exact and postcraft exact of kc20170120, kc20170208, kc20180119, kc20180128 and kc20190520
I checked kc20160518 (the current main download) but it misses postcraft formula fuctionality.
https://www.mediafire.com/file/w1d2olizftcmmrw/KC-IG_PreC-PostC.z ip/file

most of them differ a little in decimals.
The first 3 version are closest to sapload.
I have no idea how lightness works.

Formula i used:
;Generic Long axe
[data]
Durability = 100 * (1 + (((<precraft> / 100) * 0.75) + 0.20))
Lightness = 6 * (2 - (<precraft> / 100))
SapLoad = 2500 * (<precraft> / 100)
Damage = (27 + (<quality> - 1)) * (1.200 + (2.400 - 1.200) * (<precraft> / 100))
Speed = (2.272 + (4.545 - 2.272) * (<precraft> / 100)) * 6
DodgeModifier = -20 + (0 - -20) * (<precraft> / 100)
ParryModifier = -10 + (10 - -10) * (<precraft> / 100)
AdversaryDodgeModifier = <precraft> * (15 - 30) / 100 + 30
AdversaryParryModifier = <precraft> * (-20 - 0) / 100 + 0

Last I checked ... these were all correct.  I had a question on one to check out, but I forgot what it was.

https://app.ryzom.com/app_forum/index.php?page=topic/view/25037/3 0