Diamond Flask - actual crit chances (graph)

"

Herpy_Derpleson wrote:

So if I'm at 95% chance to crit, Diamonds will effectively make that 99.75%
Seems balanced.

Yes it is.

this flask is a Carnage flask

tried it on my crap ranger with like 36%. anytime there is a tough/high hp mob - pop this baby and he WILL die. ofc crit curse is mandatory to spread out the carnage effect.

i think this flask is mandatory for all builds using and stacking crit. it is 'you are going to die' button. having two of these should allow a much lenient usage but impacts pots selection very much

For the OP: I recommend expanding out 1 - (1-p)^2 to 2p-p^2 for readability. It's a bit easier to compute mentally, and is easily compared to double your crit chance (2p).

"

lilbuddha wrote:

"

Herpy_Derpleson wrote:

So if I'm at 95% chance to crit, Diamonds will effectively make that 99.75%
Seems balanced.

80 charges / 80 used

What kind of uptime do you really expect to maintain here?

What kind of uptime do I need with 95% chance to crit and 600% muliplier before flask?
One use will be enough to drop any boss.

I'm reserving judgment until I've tried it out, but it still seems OP to me.

the flask is far from op its a one time only thing and takes ages to refill

my critchance increaes from 50 to 75 (recording to that graph)

as you see the flask has dimished returns so 50 is the sweetspot imho if you dont want to invest into critchance nodes

"

Herpy_Derpleson wrote:

So if I'm at 95% chance to crit, Diamonds will effectively make that 99.75%
Seems balanced.

What kind of uptime do I need with 95% chance to crit and 600% muliplier before flask?

You're not really getting much out of that. 95% and 100% aren't all that different from each other in terms of DPS, nor are 5% and 10%. There are diminishing returns on both ends of the spectrum. Your DPS should only be increasing by about 4% with the flask active. See the graph below for proof.

"

ventiman wrote:

50 is the sweetspot

The "sweetspot" depends on your crit multiplier. Here is a graph of DPS increase versus base crit chance.

DPS_Increase = ( 1 + ( 1 - Crit_Multi ) * ( 1 - ( 1 - Crit_Chance ) ^ 2 ) ) / ( 1 + ( 1 - Crit_Multi ) * Crit_Chance )



The sweet spot is around 50% for the base crit multiplier, but no 50% crit chance build is going to have 0% increased crit multiplier. One thing is clear though: Diamond Flasks benefit you more the higher your crit multiplier is.

I'm seeing about 32% increased DPS from Diamond Flasks on my Shadow, who has 48% crit chance and 455% crit multiplier. With my level 18 Critical Weakness curse (48% + 9% = 57% crit chance, 100% + 1.67 * 355% = 693% crit multiplier), the DPS increase stays about the same (32%).

It seems like Diamond Flasks are most called for on builds with low crit chance (base 15-20% and Critical Weakness) and high crit multiplier (base 500-550% and Critical Weakness). One of these builds could see a near-50% increase in DPS with the flask active.

"

sergeantminor wrote:

"

Herpy_Derpleson wrote:

So if I'm at 95% chance to crit, Diamonds will effectively make that 99.75%
Seems balanced.

What kind of uptime do I need with 95% chance to crit and 600% muliplier before flask?

You're not really getting much out of that. 95% and 100% aren't all that different from each other in terms of DPS, nor are 5% and 10%. There are diminishing returns on both ends of the spectrum. Your DPS should only be increasing by about 4% with the flask active. See the graph below for proof.

No, you don't understand. If I'm discharging, and I fail to crit, I need to rebuild charges.
If I only fail to crit 0.25% of the time, I'mma drop shit in that one flask use.

Your graphs mean nothing when you don't understand practical application.

"

Herpy_Derpleson wrote:

"

sergeantminor wrote:

"

Herpy_Derpleson wrote:

So if I'm at 95% chance to crit, Diamonds will effectively make that 99.75%
Seems balanced.

What kind of uptime do I need with 95% chance to crit and 600% muliplier before flask?

You're not really getting much out of that. 95% and 100% aren't all that different from each other in terms of DPS, nor are 5% and 10%. There are diminishing returns on both ends of the spectrum. Your DPS should only be increasing by about 4% with the flask active. See the graph below for proof.

No, you don't understand. If I'm discharging, and I fail to crit, I need to rebuild charges.
If I only fail to crit 0.25% of the time, I'mma drop shit in that one flask use.

Your graphs mean nothing when you don't understand practical application.

Calm down, you don't have to be insulting. I wasn't aware of your particular practical application. That doesn't mean I don't understand practical application. The graphs are useful for people to whom damage per second actually matters (the majority of people reading this thread). For Discharge builds, DPS is pretty irrelevant. In your case, since your build depends on 100% charge uptime, I can see the benefits in using a flask. Either way, your specific build isn't an argument for whether or not the flask is "balanced."

"

Herpy_Derpleson wrote:

"

sergeantminor wrote:

"

Herpy_Derpleson wrote:

So if I'm at 95% chance to crit, Diamonds will effectively make that 99.75%
Seems balanced.

What kind of uptime do I need with 95% chance to crit and 600% muliplier before flask?

You're not really getting much out of that. 95% and 100% aren't all that different from each other in terms of DPS, nor are 5% and 10%. There are diminishing returns on both ends of the spectrum. Your DPS should only be increasing by about 4% with the flask active. See the graph below for proof.

No, you don't understand. If I'm discharging, and I fail to crit, I need to rebuild charges.
If I only fail to crit 0.25% of the time, I'mma drop shit in that one flask use.

Your graphs mean nothing when you don't understand practical application.

So, you use a flask slot to have (almost) no chance of having to rebuild charges, instead of 5%. That IS balanced, it doesn't do that much for your char, because the practical application is actually limited as fuck.

Why would you graph this instead of chart it? presentation of data is 90% of the research.