Celtic Heroes

This study is towards determining how resists affect damage.

For the old timers here, in 2013 Papi determined the following formula (http://www.celtic-heroes.com/forum/view ... =4&t=24804):

% damage taken = raw damage / (raw damage + resist / 2)

Or, to get your average damage:

average damage = raw damage * .75 * (raw damage / (raw damage + resist / 2))

This formula was made by testing in PvP and fit the data quite nicely. From it you could in theory calculate the resistance of a boss by recording your average damage and solving for resist. However, I think this formula needs to be reworked, as I will show below. In case you are unaware, PvE ("Player vs. Everything") means fighting mobs in the game as opposed to PvP.

Let's start with the following question:

Question: If your double your Raw Damage, what happens to your Actual Damage?

There are a few possible results:

1. Increasing Raw Damage by 2x will increase Actual Damage by less than 2x.

That would indicate the modifier is a formula like resist / (resist + raw damage)

2. Increasing Raw Damage by 2x will increase Actual Damage by more than 2x.

That would indicate the modifier is a formula like raw damage / (raw damage + resist), such as Papi's formula.

3. Increasing Raw Damage by 2x will increase Actual Damage by exactly 2x.

That would indicate the modifier is a formula that does not use damage, such as 1 / resist

So I tested this, and the TL;DR results are that the third option held: Increasing Raw Damage by 2x also increased Actual Damage by 2x.

Study:

I tested this on a Deathclasp Extractor using Lightning Strike (LS) at two levels:

High LS: 8411 raw damage
Low LS: 4133 raw damage

The high LS is 2.03x the raw damage of the low LS, which is close enough to doubling the raw damage. Since LS is pure magic damage, this would only consider the magic resistance of the mob.

As a side note, I am level 230 and the mob is level 220, however it seems your damage is not affected when you are at or above a mob's level.

I hit the Deathclasp Extractor 100 times with each LS level, and got the following averages:

High LS: 3705.93 average hit, 527.5 standard deviation
Low LS: 1869.09 average hit, 263.27 standard deviation

The High LS hit on average 1.98x the Low LS damage. I would say that is close enough to 2x to say that doubling the raw LS damage doubled the actual LS damage.

Unfortunately, that means Papi's equation does not hold, as I will show below.

Assuming Papi's equation is correct, we can solve for the magic resistance of that Deathclasp Extractor:

average damage = raw damage * .75 * raw damage / (raw damage + resist / 2)

With an average High LS damage of 3705.93 and a raw damage of 8411, we calculate the magic resist to be 11812.48. That is impossibly high, going against what is seen in the game with how lures affect damage. However, assuming it is correct, we can plug in the Low LS raw damage (4133) to see what the expected average hit would be:

raw damage * .75 * raw damage / (raw damage + resist / 2) = 1276.12

1276 vs. 1869 is quite a difference, enough to say that either the mechanics have changed sometime in the past 8 years, or PvP damage mechanics were never mirrored in PvE. While I know many here have lost interest in CH due to the lack of communication from VR, if anyone is interested in doing what I did to collect data in PvE we can determine a new equation.

I did this testing about a year ago, I have several thousand data points and found a clear curve. I did the testing to find the resists and deduce why suddenly druids were massively out performing melee damage characters. All of my numbers were found in PvP and then we used raid bosses to verify that the damage changed in-line with the PvP results.

I can share the methodology for testing that the resistance curve is the same on Mobs / Bosses?

The boss' resistance does not appear to be linear in your test, 11.8k resistance for higher raw dmg and 5.4k resistance for lower.

If you can include lures into the mix to see more variations, it will undoubtedly shine more light into this puzzle.

I don't know how to see Robert's data I probably wouldn't have much time to look at it either but I wonder if the formula could be different for physical versus elemental damage. I would love to see what you guys come up with, and any conclusions. I'm down to help a little if I can.

The boss' resistance does not appear to be linear in your test, 11.8k resistance for higher raw dmg and 5.4k resistance for lower.

Exactly. That is due to Papi's equation (which is non-linear), and that discrepancy indicates that Papi's equation is no longer correct in PvE. The correct equation would result in the same resistance for both trials.

If you can include lures into the mix to see more variations, it will undoubtedly shine more light into this puzzle.

Yep. The two variables in the equation are damage and resist, so we need trials fixing each variable to see the effect. The trial I did fixed resist and changed damage, the next trials should fix damage and change resist (via lures). I believe there are a fair few trials with this sort of data.

I did this testing about a year ago, I have several thousand data points and found a clear curve ... All of my numbers were found in PvP and then we used raid bosses to verify that the damage changed in-line with the PvP results.

I can share the methodology for testing that the resistance curve is the same on Mobs / Bosses?

Sure, I would love to read it.

...I wonder if the formula could be different for physical versus elemental damage. I would love to see what you guys come up with, and any conclusions.

I highly doubt that the formula would change based on the damage type, mainly because of how complex that would be. Instead it is likely that elemental and physical resists vary widely on mobs as part of class balancing. If you are interested in collecting data, try what I did except instead of changing your damage, change the mob's resists (by using a lure of some sort). Also, be sure to only use one damage type.

This is data myself and some clannies collected.



Things of note:
At 0 resists, the minimum hit is 60% of your max, however the average hit is 75%, not 80%.
Maximum resistance is approximately 92%, when a boss' resists become "plateaued"
The first 1000 protections equals 60% damage mitigation

These results were validated on BT, Gelebron and Dhiothu. The expected change in damage occured when we used lures to reduce the bosses resistance and matched the curve of the data obtained in PvP. All damage types had the same resistance curve.

Do you have any tested resistances data from bosses?

Do you have any tested resistances data from bosses?

Yes, I'd have to look through all my old sheets and chats but I believe BT has about 3800 Magic resists, Gele has 5800 physical resists.

Do you have any tested resistances data from bosses?

Yes, I'd have to look through all my old sheets and chats but I believe BT has about 3800 Magic resists, Gele has 5800 physical resists.

This is gold!
Please do so when convenient, thank you in advance.

I have a follow up question to resistances.

Is it better to have one type of elemental damage gear and jewels?

If I have both fire and cold will my damage output be less than if I focus on 1 type?