Your question is why we are asking the question. If it is 15% of the base, that means 16 adds of 15% is X. However, if it is 15% of 15% of 15%, it is Y. Also, if it is 15% of 15%X15% that is Z. X and Z are not that bad and should exceed 100% without much of a problem. Even if they scale to difficulty, it should still get above 100%. Unfortunately, Y means even smaller amounts of increase that mean increasing the modifier much beyond two or three steps is useless.
That’s not how success over multiple attempts is calculated, unless Blizz is doing something super weird with multi hit abilities.
Ignoring any other multipliers, if you have 5 shots and each one has a flat 20% chance to trigger, then you would have a 67.232% chance of getting a Lucky Hit.
Here’s the formula: 1 - (1 - 0.2)⁵ = 0.67232
Basically, just calculating the chance for failure then subtract it from 100% to get the chance for success.
2 Likes
The people that have tested it say the only things that aren’t calculated per hit are conjurations and channelled skills.
The deeper I drive into this lucky hit nonsense the more that makes sense though.
Like say they used the per hit lucky hit chance on hydra and the way increased lucky hit chance is assumed to work.
Right now hydra is 77% and hits about 36 times per cast so could assume it would be ~2% lucky hit of it was labeled per hit.
A 10% lucky hit is about a 0.2% chance to happen per shot and “up to” traits like firebolt enchant are adding about 2% of the listed maximum value.
Now if we add +20% lucky hit chance from a piece of gear.
The over duration becomes 97% chance over ~36 hits. Which kind of turns into ~0.25% per shot and 2.5% of an “up to” trait
If it was labeled as 2% and counted per hit that same boost would take it to 22% and translate that same 10% proc into 2.2% chance per shot and 22% of an “up to” trait
Thats a pretty massive power swing.
Would hope they find a way to add the compounded values somewhere. I imagine the average player that hasn’t spent the time to untangle it will be thinking the game is broken when their 20% chance to do something cool skills are only happening every 5 minutes.
You don’t even know there is a lucky hit component on the skills without advanced tooltips on.
1 Like
Pretty sure people will not be surprised with the “hey my Blizzard should execute every 5th shot, what is this BS”, i.e. won’t expect every skill to behave the same, BUT yes, Blizz should probably have a spot (probably right above the skillbar with a small/hoverable (i) or sth) where it says "this skill has lucky hit of (calculated % number) to (lucky hit equipped))
I would give the developers more credit than that. It’s going to be noticeable if you spec into it.
The fact that Character Planning and True Numbers are not anywhere to be found in the game gives me little confidence. Diablo 4 heavily relies on making consistent investments to build your character. The end-path is not clear and then on top of that reapeccing your entire Paragon Board and regraring isn’t going to be really easy. There isn’t any sandbox to test out builds prior to making investments.
You would think that as a QoL tool this would be inherently in the game to help guide players. Instead, everyone is going to be heading over to maxroll, watching long-winded egotistical streamers on YouTube just to pause it on the build.
The best suggestion I could make would be a dynamic tooltip for each skill on your bar, that does the lucky hit math and shows you the proc rate for each lucky hit effect you have equipped at the time. It would be messy if you had lots, but its the only think I can think of that would accomplish explaining this in the game in real time.
As I understand lucky hit change means chance to execute some passive effect on skill,
for example you have “25% chance to detonate enemy on kill” passive on the skill, AND lucky hit chance is 20% it means 0.2*0.25 * 100 which means 5% change to “DETONATE ENEMY ON KILL” which honestly sucks hard…
Imagine even if you have 100% of lucky hit chance you will still “detonate on kills” 25% of a time. If this is true they either need to REMOVE it from the game or REFACTOR it entirely
Depends on the skill itself, AoE skills in general SHOULD have low proc % chance, the single-target ones on the other hand have a better LH% (some of them probably even 100%)
1 Like
I visualise it in my head as throwing two dices. First you roll if you get to throw the second one, and the second throw determines if the ability in question is taking effect.
They just need to remove the lucky hit mechanic. It’s terrible and doesn’t fit in the game. All lucky-hit builds are going to be complete garbage, or op. They won’t have good middle ground. Procs are just a dumb mechanic for an ARPG. Players want consistent predictable damage that they can control.
Grim Dawn is full of procs. Noone is crying over those.
Bad example… I guess depends on WHAT the proc itself is, GD is simply terrible in that regard, having a 3% chance to proc a “Death Beam” (or whatever the name was) and wiping out a 30% of HP of an elite while literally struggling with the rest… Another proc was that “blade barrier” or whatever that does a whirling blade AoE around the player, that’s like the 2nd highest DPS thing in game (right after the Death Beam)
Really really bad example/design… Yes, procs and LHs can enrich gameplay, but should be nowhere near as efficient/dominant as equipped skills
And it gets even worse, got curious to what a “Death Beam” build would look and specced exactly into that Elementalist/Warlock build that utilises the skill - sure, does about 30-50% more damage than that one that is procced, BUT costs a TON… i.e. your skill causes almost immediate mana depletion and by the time you can do your next cast the “free” one will proc like 3 times in between
In other words - the free one that procs does more effective DPS than the equipped skill of procs of the same skill… That should simply NEVER be the case…
IDK about others but really grown to hate GD, and those are the main reasons behind = that “damage salad” of damage-type conversions and procs for days, really bad example overall I think