Ok so I was talking about Sharpshooter with a buddy so I kind of want to share information regarding the skill as I understand it and hopefully create some intelligent discussion on it.
We all know how SS gives you the big numbers to show off which is great... however I wanted to analyze how good it really is or not.
We all know that at at high levels of critical chance (crit rate), it becomes less useful and can be replaced with other skills. But at what crit levels would we consider it useless?
Lets take a look at some analysis of it (as i understand it... i could be completely wrong so please fix any errors in logic/math)
Understanding Expected Critical Chance
Lets say for a given of 2 APS and a 5% crit rate.... you can expect a critical once every 10 seconds
Now at 15% crit and same APS.. you can expect a critical every 3.3 seconds
Now at 25% crit... 2 seconds
and at 40% crit... 1.3 seconds
Now from my understanding each second that you dont get a crit you get one "stack" of SS (with each stack giving 3% crit).
With 5% crit rate, that means you can expect up to 10 stacks or 30% crit increase.... Now it gets complicated as the increase in crit rate from the stacks is progressive and will shorten the time in between criticals and not sure how to do the math here so someone can help me. But if you think of 10 seconds as WORST case scenario every time then the math becomes easier i guess?
Now at very high levels of crit we can see that it will yield a very low rate of return. at the WORST case, you get 1 stacks before you crit again.
However, can we say in the very worst... SS at high levels of crit will yield a 3% crit rate increase?
Benefits DPS wise
Obviously how beneficial a 3% crit to you depends on your level of gear and how you stacked up crit damage etc.
Using the awesome DH DPS calculator from the other thread... It calculated that each additional point of critical chance = 1810 DPS for me (2100 dex, 270% crit dmg) ... so a 3% crit increase is roughly 6000 dps increase which is very significant. and this is worst case scenario.
Now my real stats are around 2.5 APS with 15% crit so it works out to be an expected crit once every 2.6 seconds. In our theoretical scenario where i crit exactly once every 2.6 seconds... that means i'll have around 2 stacks of SS before I crit again or works out to be 12000 dps increase?
Is this right?
This bonus is reset one second after you successfully hit
Curious about this statement as I never really read through the whole tooltip.
Lets assume you had full stacks and had 100% crit rate... and you attack a monster, obviously that first hit will be a critical, but you get to keep the 100% crit rate for 1s? Not sure how this works hopefully someone can explain.
Also curious if anyone knows whether crit is calculated when the projectile is created or when the projectile hits the monster? And if it hits multiple monsters (pierce) does it crit all monsters or will it have a seperate roll per monster?
Benefits of SS
- Front load your damage
- good if you had low crit chance
- becomes stronger with higher crit dmg
- Favors low APS
Negatives
- Doesnt work as well for high crit chance + APS
I created this thread as kind of a discussion into SS as it seems to me it's not too bad even with high crit (to a certain point). Would love to hear everyones thoughts and what you guys think.
remember that you count on expected crits using single, one hit attacks.
elemental arrow whit tentacles hit a few times and even a whole group.
hungering arrow hits 1+35% to pierce and hit again (and an addition 35% to pierce and so on) so you will be doing critical hits allot more then.
i personally don't like sharpshooter, it give a high burst damage in the start, but mid fight its just a useless passive doing nothing really.
I am really interested in the 'when is the calculation of a projectiles dmg/crit done' part. Because this question opens different possible follow-up questions...
i.e. if dmg is calculated on impact but crit chance isnt, will all the projectiles i shot before the first impact until one second after be crits? would that make max-ias builds or even dw 1-handed xbows inferior (more projectiles before the first impact and possibly more +crit dmg on 2x 1handed xbow)?
Or is everything calced on impact and only the impacts 1sec after the first hit will be crits? would this make multishot inferior to nt (at least for 4mob-champs and trash-mobs since you would be able to get in more definite crits in that 1sec window than with NT?
Hi,
I decided to make a small matlab program to calculate the average crit chance using sharpshooter. I could have made a new topic but lets put it in here. The code is in the spoiler below. Note that this is NOT taking into account travel time of the shot, which hugely impacts nether tentacle effectiveness as discussed above.
You choose a base crit chance and aps. Then the simulation runs for 1e5seconds (approx 27 hours). So basically it assumes a person shooting at a training dummy with an infinitely fast arrow for 27 hours. Every second the crit chance is increased by 3 %, and every time a crit occurs a timer starts resetting the crit to the initial value after one second. Every second the in situ crit chance is stored. In the end the average crit chance is calculated, along with the max crit chance obtained.
In the following I present some results with some choice base crit and aps values:
iniCrit=0.05, aps=1: Mean crit = 0.187, max=0.68
iniCrit=0.05, aps=2: Mean crit = 0.152, max=0.50
iniCrit=0.05, aps=4: Mean crit = 0.127, max=0.38
iniCrit=0.25, aps=1: Mean crit = 0.336, max=0.73
iniCrit=0.25, aps=2: Mean crit = 0.311, max=0.61
iniCrit=0.25, aps=4: Mean crit = 0.296, max=0.49
iniCrit=0.5, aps=1: Mean crit = 0.555, max=0.83
iniCrit=0.5, aps=2: Mean crit = 0.542, max=0.68
iniCrit=0.5, aps=4: Mean crit = 0.536, max=0.65
- From this data it is clear that low aps is better.
- When having no crit chance on items, sharpshooter will give you an average of about 15-18%, depending on attack speed.
- At 25% base crit chance, sharpshooter gives a mean increase of about 7 crit chance, and this number is of course reduced as the base is increased further.
Feel free to implement this code in any kind of dps spread sheet you might be using
function sharpshooter(iniCrit, aps)
% function sharpshooter(iniCrit, aps)
% Calculates sharpshooter average crit chance
% "Gain 3% crit hit chance every second.
% Bonus is reset 1 second after you critically hit"
% I assume the crit increase pr second is timed individually
% from when you actually crit.
%*****************************************************
if ~exist('iniCrit','var')
iniCrit = 0.05; % Initial crit percent chance
end
if ~exist('aps','var')
aps = 2; % Attacks pr second
end
maxTime = 1e5; % Time in seconds when simulation finish
currentCrit = zeros(1,maxTime); %Output array
dt = 0.001; %Simulation resolution in seconds
currentCrit(1) = iniCrit;
curCrit = iniCrit;
isCrit = 0;
shootTimer = 0; %Timer for next shot
resetTimer = 0; %Timer for reseting the crit
for i=1:dt:maxTime
shootTimer = shootTimer+dt;
if isCrit
resetTimer = resetTimer+dt;
if resetTimer >= 1
curCrit = iniCrit;
isCrit = 0;
end
end
if shootTimer >= 1/aps %Time to fire a shot!
if rand<=curCrit
%We have a crit
isCrit = 1;
if resetTimer>1
resetTimer = 0; %When this value =1 crit is reset
end
end
shootTimer = 0;
end
if floor(i)==i && i~=1 %Write output and increase crit every second
if curCrit<1
curCrit = min(1,(curCrit + 0.03));
end
currentCrit(i) = curCrit;
end
end
disp(['Mean: ' num2str(mean(currentCrit))]);
disp(['Max: ' num2str(max(currentCrit))]);
Hi,
I decided to make a small matlab program to calculate the average crit chance using sharpshooter. I could have made a new topic but lets put it in here. The code is in the spoiler below. Note that this is NOT taking into account travel time of the shot, which hugely impacts nether tentacle effectiveness as discussed above.
You choose a base crit chance and aps. Then the simulation runs for 1e5seconds (approx 27 hours). So basically it assumes a person shooting at a training dummy with an infinitely fast arrow for 27 hours. Every second the crit chance is increased by 3 %, and every time a crit occurs a timer starts resetting the crit to the initial value after one second. Every second the in situ crit chance is stored. In the end the average crit chance is calculated, along with the max crit chance obtained.
In the following I present some results with some choice base crit and aps values:
iniCrit=0.05, aps=1: Mean crit = 0.187, max=0.68
iniCrit=0.05, aps=2: Mean crit = 0.152, max=0.50
iniCrit=0.05, aps=4: Mean crit = 0.127, max=0.38
iniCrit=0.25, aps=1: Mean crit = 0.336, max=0.73
iniCrit=0.25, aps=2: Mean crit = 0.311, max=0.61
iniCrit=0.25, aps=4: Mean crit = 0.296, max=0.49
iniCrit=0.5, aps=1: Mean crit = 0.555, max=0.83
iniCrit=0.5, aps=2: Mean crit = 0.542, max=0.68
iniCrit=0.5, aps=4: Mean crit = 0.536, max=0.65
- From this data it is clear that low aps is better.
- When having no crit chance on items, sharpshooter will give you an average of about 15-18%, depending on attack speed.
- At 25% base crit chance, sharpshooter gives a mean increase of about 7 crit chance, and this number is of course reduced as the base is increased further.
Feel free to implement this code in any kind of dps spread sheet you might be using
that is awesome. I didnt get a chance to go through code I will do that on way back.
Was exactly what i was expecting from raw data (although my calculations are way off)
From data, looks like you still see a minimum benefit of 3% crit at high levels aps and crit
remember that you count on expected crits using single, one hit attacks.
elemental arrow whit tentacles hit a few times and even a whole group.
hungering arrow hits 1+35% to pierce and hit again (and an addition 35% to pierce and so on) so you will be doing critical hits allot more then.
i personally don't like sharpshooter, it give a high burst damage in the start, but mid fight its just a useless passive doing nothing really.
That makes the calculations way too difficult at that point And there are also too many unknowns in the way they code.
Again this is a perception which i'm trying to ultimately clear especially with what gear levels you have.
I am really interested in the 'when is the calculation of a projectiles dmg/crit done' part. Because this question opens different possible follow-up questions...
i.e. if dmg is calculated on impact but crit chance isnt, will all the projectiles i shot before the first impact until one second after be crits? would that make max-ias builds or even dw 1-handed xbows inferior (more projectiles before the first impact and possibly more +crit dmg on 2x 1handed xbow)?
Or is everything calced on impact and only the impacts 1sec after the first hit will be crits? would this make multishot inferior to nt (at least for 4mob-champs and trash-mobs since you would be able to get in more definite crits in that 1sec window than with NT?
Aaaah... too many questions...
Yes these are the questions I'm looking for... very interested if anyone has answer for them
I do like SS, its well for bursting some champions down. A situation where it shines is when you have to kite or avoid/dodge incoming projectiles. During that time your crit can build up just to burst another time. Or you just kite a bit, build up crit, stand still and chain-smoke screen + preparation and burst them down.
This works especially well on groups with reflect damage. Not to mention that is so satisfying to one-shota mob you really do hate (with 100% crit stacked up).
To explain the 1s rule a bit: (NT = Nether Tentacle rune)
Lets say you have 3 attacks/sec and the travel time for NT to your mob is 3 sec as well. So the 1s counter will start after 3 seconds (first hit) and at that time you will have casted a total of 9 NT. During the next second you will cast another 3 NT, which sum up to 12 arrows.
Depending on whether the mob is moving / its hitbox size, the arrows can hit up 12-36 times, sometimes even more often.
And all of those 12-36 hits will be guaranteed crits as the crit chance is calculated at the time the arrow is generated.
The crit rate will be bound to each arrow. After the last second with 100%, all following arrows will have your standard critrate.
Conclusion: That's the explanation why IAS + Crit DMG is so viable for Demon Hunters with SS
We all know how SS gives you the big numbers to show off which is great... however I wanted to analyze how good it really is or not.
We all know that at at high levels of critical chance (crit rate), it becomes less useful and can be replaced with other skills. But at what crit levels would we consider it useless?
Lets take a look at some analysis of it (as i understand it... i could be completely wrong so please fix any errors in logic/math)
Understanding Expected Critical Chance
Lets say for a given of 2 APS and a 5% crit rate.... you can expect a critical once every 10 seconds
Now at 15% crit and same APS.. you can expect a critical every 3.3 seconds
Now at 25% crit... 2 seconds
and at 40% crit... 1.3 seconds
Now from my understanding each second that you dont get a crit you get one "stack" of SS (with each stack giving 3% crit).
With 5% crit rate, that means you can expect up to 10 stacks or 30% crit increase.... Now it gets complicated as the increase in crit rate from the stacks is progressive and will shorten the time in between criticals and not sure how to do the math here so someone can help me. But if you think of 10 seconds as WORST case scenario every time then the math becomes easier i guess?
Now at very high levels of crit we can see that it will yield a very low rate of return. at the WORST case, you get 1 stacks before you crit again.
However, can we say in the very worst... SS at high levels of crit will yield a 3% crit rate increase?
Benefits DPS wise
Obviously how beneficial a 3% crit to you depends on your level of gear and how you stacked up crit damage etc.
Using the awesome DH DPS calculator from the other thread... It calculated that each additional point of critical chance = 1810 DPS for me (2100 dex, 270% crit dmg) ... so a 3% crit increase is roughly 6000 dps increase which is very significant. and this is worst case scenario.
Now my real stats are around 2.5 APS with 15% crit so it works out to be an expected crit once every 2.6 seconds. In our theoretical scenario where i crit exactly once every 2.6 seconds... that means i'll have around 2 stacks of SS before I crit again or works out to be 12000 dps increase?
Is this right?
This bonus is reset one second after you successfully hit
Curious about this statement as I never really read through the whole tooltip.
Lets assume you had full stacks and had 100% crit rate... and you attack a monster, obviously that first hit will be a critical, but you get to keep the 100% crit rate for 1s? Not sure how this works hopefully someone can explain.
Also curious if anyone knows whether crit is calculated when the projectile is created or when the projectile hits the monster? And if it hits multiple monsters (pierce) does it crit all monsters or will it have a seperate roll per monster?
Benefits of SS
- Front load your damage
- good if you had low crit chance
- becomes stronger with higher crit dmg
- Favors low APS
Negatives
- Doesnt work as well for high crit chance + APS
I created this thread as kind of a discussion into SS as it seems to me it's not too bad even with high crit (to a certain point). Would love to hear everyones thoughts and what you guys think.
elemental arrow whit tentacles hit a few times and even a whole group.
hungering arrow hits 1+35% to pierce and hit again (and an addition 35% to pierce and so on) so you will be doing critical hits allot more then.
i personally don't like sharpshooter, it give a high burst damage in the start, but mid fight its just a useless passive doing nothing really.
i.e. if dmg is calculated on impact but crit chance isnt, will all the projectiles i shot before the first impact until one second after be crits? would that make max-ias builds or even dw 1-handed xbows inferior (more projectiles before the first impact and possibly more +crit dmg on 2x 1handed xbow)?
Or is everything calced on impact and only the impacts 1sec after the first hit will be crits? would this make multishot inferior to nt (at least for 4mob-champs and trash-mobs since you would be able to get in more definite crits in that 1sec window than with NT?
Aaaah... too many questions...
I decided to make a small matlab program to calculate the average crit chance using sharpshooter. I could have made a new topic but lets put it in here. The code is in the spoiler below. Note that this is NOT taking into account travel time of the shot, which hugely impacts nether tentacle effectiveness as discussed above.
You choose a base crit chance and aps. Then the simulation runs for 1e5seconds (approx 27 hours). So basically it assumes a person shooting at a training dummy with an infinitely fast arrow for 27 hours. Every second the crit chance is increased by 3 %, and every time a crit occurs a timer starts resetting the crit to the initial value after one second. Every second the in situ crit chance is stored. In the end the average crit chance is calculated, along with the max crit chance obtained.
In the following I present some results with some choice base crit and aps values:
iniCrit=0.05, aps=1: Mean crit = 0.187, max=0.68
iniCrit=0.05, aps=2: Mean crit = 0.152, max=0.50
iniCrit=0.05, aps=4: Mean crit = 0.127, max=0.38
iniCrit=0.25, aps=1: Mean crit = 0.336, max=0.73
iniCrit=0.25, aps=2: Mean crit = 0.311, max=0.61
iniCrit=0.25, aps=4: Mean crit = 0.296, max=0.49
iniCrit=0.5, aps=1: Mean crit = 0.555, max=0.83
iniCrit=0.5, aps=2: Mean crit = 0.542, max=0.68
iniCrit=0.5, aps=4: Mean crit = 0.536, max=0.65
- From this data it is clear that low aps is better.
- When having no crit chance on items, sharpshooter will give you an average of about 15-18%, depending on attack speed.
- At 25% base crit chance, sharpshooter gives a mean increase of about 7 crit chance, and this number is of course reduced as the base is increased further.
Feel free to implement this code in any kind of dps spread sheet you might be using
that is awesome. I didnt get a chance to go through code I will do that on way back.
Was exactly what i was expecting from raw data (although my calculations are way off)
From data, looks like you still see a minimum benefit of 3% crit at high levels aps and crit
That makes the calculations way too difficult at that point And there are also too many unknowns in the way they code.
Again this is a perception which i'm trying to ultimately clear especially with what gear levels you have.
Yes these are the questions I'm looking for... very interested if anyone has answer for them
Agreed
This is not at all true. Please stop spreading misinformation.
Don't have time to dissect the OP, but your calculations are all assuming NT is hitting one mob one time, which pretty much never the case.
It's called a bug. Confirmed by Blizzard.