Some skills are ineffective due to how rounding works

[microcthulhu]

Example:
10% of critical skill 9 is zero. (one handed sword)
20% of weapon damage 4 is zero (w&s spec)
Sometimes rounding is not pure zero, but still making stat increase negligible.
Suggestion: remove all secondary stat increases less or equal 20% from all skills and add them to main stat increase.

lvl 48
HP 155, AP 12/3, AC 264% AD 29-43 BC 108%/6
WA 2, HH 1, CS 2, cleave 1, IF 3, Regen 1, DboS 1, One-handed sword 1, Shield 1, Light Armor 1, Heavy armor 1, Weapon&Shield 2,

[rijackson741]

Yes, I have also noticed this. It's especially an issue for CS, because CS is converted to ECC in a very non-linear way, and then ECC is also rounded down. So even if you had CS10, so 10% of that is CS1, that may not change your ECC at all.

I don't understand your suggestion though. What do you mean by "add them to the main stat increase"?

[microcthulhu]

I don't understand your suggestion though. What do you mean by "add them to the main stat increase"?

Usually skill increase more then one stat, one stat for 30% another for 10% for example. I suggest move minor increase to major - one stat for 40% instead.
If only one stat increased and still rounding is too detrimental - merge two skill step into one - instead sikll +x% at level 10 and +x% at level 20 make it +2x% at level 20
Changing how rounding works would be too expensive in coding/debug time probably - replacing all integers to float.

lvl 48
HP 155, AP 12/3, AC 264% AD 29-43 BC 108%/6
WA 2, HH 1, CS 2, cleave 1, IF 3, Regen 1, DboS 1, One-handed sword 1, Shield 1, Light Armor 1, Heavy armor 1, Weapon&Shield 2

[rijackson741]

OK, I understand what you mean, although I don't agree that it's the ideal solution. When skills boost AC or BC I don't think the rounding is a big issue. Although the rounding of AD is something of an issue now, it's likely that it will be much less of an issue in the future. As for CS, the problem is not just rounding, but also the way ECC is calculated. The highest CS available from a weapon is 20, so a 10% boost is only an additional CS2. The most that can get you is an increase in ECC (which is what actually matters) of 1%. But if you have a weapon with CS 20 you already have an ECC of at least 15%, so it's just boosted to 16% . And that's a best case scenario! I think what is needed is that the boost to CS is much higher, maybe 50% instead of 10%.

[microcthulhu]

The CS boosting is difficult question IMHO. CS probability should have diminishing return, otherwise it's just would be just too much unrealistic (unlike CS multiplier). So question here is purely what should be asymptotical capping of CS - 17% or 20% or 25%. Existing capping doesn't look too small to me, IMHO if damage is not enough multiplier should be raised instead. May be solution is to live CS chance as is but add items which increase multiplier linearly. That create opportunities for interesting equipment strategies.

[rijackson741]

We need to distinguish between Critical Skill, CS, and Effective Critical Chance, ECC. CS is the level of skill you get, and that is converted to ECC, which is the actual chance of a critical hit. The conversion is very non-linear, and although there is no cap on ECC it is effectively impossible to get it to very high levels (you would need a CS of 522 to get an ECC of 100%!). I posted a link to a spreadsheet that shows the conversion in this thread: http://andorstrail.com/viewtopic.php?f=3&t=4864