New Critical chance final?

[kowocki]

Originally Posted by mind-trick

actually mages were initially spell casters and cast speed is just terribly slow without cast speed gear after the cs fix.

Simple solution-> do arcana devotion passive and swap its position at the spell tree with that intelligence passive.

With lvling the character player unlocks next lvls of this spell boosting its casting speed gradually-> mimicking characters proficiency in arcana of magic.

lvl1:+20% cs
lvl2: +40% cs
lvl3: +60% cs
lvl 4: +80% cs
lvl5: +100% cs

[-Mongoose-]

Here's some data that I collected that people may find interesting:

TL;DR: Long term & short term behavior of critical chance in 1.12 appears to match theoretical calculations.

Test Setup:
Level 60 hunter attacking fort door with a range 35 bow at max distance.
Unbuffed
Attack: 334-417
Critical Attack: 444-544
Critical Chance: 5.81%
Attack the door, recording the damage and if it was a crit. Did so until my bow was broken (wasn't at full durability to begin with).

Data Collection:
Number of Hits: 1043
Number of Critical Hits: 55

Data Analysis:
Sample Probability of Critical Hits: 0.0527
95% Confidence Interval: (0.0392, 0.0663)
95% confidence interval defined as p +- z*sqrt(1/n*p*(1-p)) representing a binomial confidence interval.

The sample probability falls within the bounds of what we would expect given the number of hits recorded.

Attached are several graphs, showing a histogram of damage and the damage vs time. If you look at the damage over time, you can see that critical hits are fairly well spaced out.

Looking at short term behavior, I wasn't sure how to analyze this, if anyone has better suggestions, I'm open to ideas.

I've taken several metrics. If you look at the file attached labeled moving average, I've plotted the sample probability of critical hits over a sequence of 100 hits. The first 100 hits and last 100 are not accurate reflections of the moving average since 100 points are not available to be averaged. If you look In the middle 800, the average still remains around 0.05, which is what we want to see.

Back to back critical hits were never observed. Remember this is an unbuffed archer, it's probably possible with the critical chance buffs, but I haven't tested those.

The last metric I looked at was examining how well crits were appearing over sequences of the expected time between crits. Theoretical crit probability is 0.0581; the expected value of time between crits is 1 / p = 17.2.

If you look at the file labeled sequences_of_expected_length_single_crit.png, the blue shows times when there was exactly 1 crit during the recent 17 hits. The closer this graph is to being completely blue, the closer the system is to having short term behavior match long term theoretical expectations during a short term sequence.

[LawZ]

Originally Posted by -Mongoose-

...

I love you and your Matlab (?) figures.

The analysis is sound.

Best,

[Proviva]

Originally Posted by -Mongoose-

Here's some data that I collected that people may find interesting:

This is relevant to my interests


Even tho, sometimes game just says "fuck you math"

[mind-trick]

Originally Posted by Takeyo

Mages were also capable of very respectable dps at one point in time. Not everyone wants to play mage only to support, gimp people to death, or compete without using good gear against players with good gear (which really doesn't work as well as you might think unless you are quite skilled). Many players would actually like to be able to have the chance to play the magical dps role that warlocks are supposed to be able to fill quite comfortably.

We don't want mages to be locked into one particular set of play-styles. We just want mages to be balanced such that if they wish to deal more damage (or more healing in the case of healers), they must sacrifice some control, and vice versa. The damage (or healing) gained by sacrificing control should be a viable option when executed intelligently.

To say that low damage output is a good thing on mages assumes that mages cannot fill the damage role, and were never intended to be able to fill the damage role. That is a huge misconception, and if the developers would take a look at staff mastery and mages' other attack-focused trees in terms of viability with attack in mind and restore mages' full range of versatility without lowering the skill cap, all while maintaining the interclass balance across the game, that would be a very good thing.

As Adrián says, such a task is easier to talk about than to execute, and balance is tough to hit on mages due to the high skill cap. All that aside, drastic changes do not ruin the game if they are balanced changes that foster dynamic gameplay. In that case, drastic changes are progress. Personally, I'm ready for some drastic changes over here if that's what it takes to bring mages up to date, because playing a class that is severely outdated just feels broken and wrong, even if it is effective under the right circumstances.

yes, its completely fine to add a new dimension to mages if you maintain the old one. i never really tried staff mastery though so i dont know if SM is worthy enough to call it a 'new dimension' for mages. but i believe it is quite effective with the right gear and you have to sacrifice a lot of skills points for it.

[Ygarl]

Originally Posted by Proviva

This is relevant to my interests


Even tho, sometimes game just says "fuck you math"

This is the wonder of a good random number generator:

If you flip a coin 10000 times, and get heads every single time - what's the chance of heads again?


1 in 2.
Sometimes - by its very definition - you can flip those coins and get 20 heads in a row. That's statistics for you.
However, theoretically if you flip <infinity> times, statistics states that half of those infinite coin tosses (in a perfect, sealed system) will be heads.
Regnum however - with all the continually changing variables, and a non-totally perfect pseudo-random generator - will roll you up some bullseyes a dozen times in a row.

C'est la vie...

(still sucks when it happens to YOU though!)