Curse

evenfall · Jun 13, 2009, 06:16 PM // 18:16

Happen to saw that izzy had posted the formula for base critical hit chance.

An interesting revelation is that there are 4 variables, one of which was never mentioned before AFAIK.

Critical Hit Chance=IF(((0.05*2^((((8*LevelA)+(4*WeaponSkill)+6 *MIN(WeaponSkill,(LevelA+4)/2)-(15*LevelD)-100))/40))*(1-(WeaponSkill*0.01)))+WeaponSkill*0.01-WeaponCritChance>0,((0.05*2^((((8*LevelA)+(4*Weapo nSkill)+6*MIN(WeaponSkill,(LevelA+4)/2)-(15*LevelD)-100))/40))*(1-(WeaponSkill*0.01)))+WeaponSkill*0.01-WeaponCritChance,0)

LevelA - Level of attacker
LevelD - Level of defender
WeaponSkill - Weapon mastery attribute level
WeaponCritChance -

Does this means that different weapons actually have different base critical hit chance?
Anyone has further experience or info?

RadaArashi · Jun 13, 2009, 06:59 PM // 18:59

Wait nvm that'd be WeaponSkill I just described.

Ph33rtehGD · Jun 13, 2009, 07:27 PM // 19:27

Well, I haven't really looked at it too deeply, but here's what I think from a programmer's standpoint. if you make something a variable like WeaponCritChance is in that equation, that would mean there was at least an intention that it could change depending on certain circumstances. The most logical choice would be the type of weapon you have. However, I'm not speaking from any experience, I'm just logically trying to reason things out.

Trenixz · Jun 13, 2009, 09:32 PM // 21:32

I think wands and staffs have a low crit chance, so it could be that that's where the change is.

Chthon · Jun 13, 2009, 11:42 PM // 23:42

Quote:

Originally Posted by evenfall

Happen to saw that izzy had posted the formula for base critical hit chance.

1. Where did you find this formula?

2. If you set both levels to 20 and set WeaponCritChance around -0.06, you can get results at least in the neighborhood of the wiki's empirical results.

3. I have some doubts about whether this formula is correct because
a. WeaponCritChance is negatively correlated to the final output and needs to be a negative number to get "sane" results. That strikes me as odd.
b. WeaponCritChance has a HUGE impact on the final result. Much bigger than any other variable. Either this variable is a remnant of early game design that got locked down as a constant, totally misdescribed (maybe it's actually related to crit bonuses from assassin skills?), or the formula isn't accurate.

4. Of course, it's possible I misparsed the equation. Could some other people plug in numbers and see if they agree?

[edit:

Quote:

Originally Posted by Trenixz

I think wands and staffs have a low crit chance, so it could be that that's where the change is.

That could be it. Perhaps WeaponCritChance is one of two constants. Of course, that still doesn't explain why i needs to be negative....]

Ensign · Jun 14, 2009, 12:19 AM // 00:19

That can't possibly be right as written; the exponential part reduces to <1% for almost all situations, and as written gives a crit% of 12.55% at 12 spec, which is empirically too low.

For level 20 characters fighting whatever the equation should reduce to:

CRIT% = 5*2^((6*MIN(WM,12)+4*WM-120)/40)+WM

Or something very close to that, with WM being your weapon mastery; 17% at 12 spec and ~22-23% at 16 spec. The exponential part is a pain to pinpoint since the 1% per level is the larger share and the exponential only gets big when the level differences get large, and that introduces another variable. I've just assumed that it follows the same sort of scaling as all the other equations, with the exponential being 0 when everything matches.

I presume that WeaponCritChance is legacy code.

evenfall · Jun 14, 2009, 03:38 AM // 03:38

It should be a copy and paste from excel, so if there is any mistake, it probably means izzy himself got it wrong.

The post is from here.

Another thing to note is that WeaponCritChance may not be a constant.

I saw the old empirical test data from the unofficial wiki, but no mention of the weapon used was made.

FoxBat · Jun 14, 2009, 05:34 AM // 05:34

Reducing the exponential part of the OP's forumla to something comprable to what Ensign posted:

(6*MIN(WM,12)+ 4*WM - 240)/40

-240 and -120 just seem too closely related to be coincidence. Maybe this was an older version they were using.

One thing the listed formula DOES do is offer incredible boosts when the level difference is greater than 10, maybe that's all the old version was intended to do. That would also explain Izzy's "recollection" that the formula was basically weapon mastery for even-leveled chars.

Chthon · Dec 26, 2009, 11:08 PM // 23:08

Thread rez here.

Think I figured it out. If we assume 1 typo in Izzy's formula and that WeaponCritChance is simply a zero for martial weapons, we can get sane values using this:

Quote:

BaseCritChance = (0.01*Mastery) + ((1 - (0.01*Mastery)) * 0.5 * 2^(((8*AttackerLvl) + (4*Mastery) + (6 * Min{Mastery, ((AttackerLvl + 4)/2)}) - (15* DefenderLvl) - 100) / 40))

Full post: here.