A comparison of tanking proc relics

Gardimuer · March 6, 2016

I've always followed the theorycrafter recommendation that the Reactive Warding relic provides the most mitigation over the course of a fight. However, recently I had a dispute with my raid lead over which tanking relics provide the best mitigation, so I decided to test the RW relic against a Shield Amplification relic in a full clear of NiM DP and a partial NiM TFB.

Method:

I wore an Exarch Relic of Reactive Warding and an Exarch Relic of Shield Amplification while tanking each boss on my Assassin. I left out Bestia and the Dread Guard because the stacking debuff would skew the results.

The RW relic proc absorbs 2840 damage.

The SA relic grants 652 absorb rating, which gives 6.804595% absorb at my current stat budget. My shield chance with Dark Ward up is 58.65%, my defense chance is 35.83%

After each successful clear I used Starparse to do the following for each fight:

I recorded the number of times the RW relic proced. I then multiplied this number by 2840 to get the total damage absorbed. This number is the total mitigation provided by the RW relic.
I recorded the average damage of each attack type when completely unmitigated.
I recorded how many of each attack type was shielded during the uptime of the Shield Amplification proc.
I then multiplied separately for each attack type: (average unmitigated damage)*(6.804595%)*(number of shielded attacks during proc)

Results:

The Shield Amplification Relic mitigated the most damage in five out of seven fights. My tentative conclusion is that the SA relic is better for Assassins for any fight with very little internal/elemental damage. Even if the SA relic is not the best, it is at least not as far behind as I had previously thought based on the theorycrafting.

(Detailed results and calculations here: [spreadsheet])

Tyrans

DTPS - 2857
% internal/elemental - 10.3%
melee/ranged - 41.6%
Fight length - 6:24
RW total mitigated - 19880
SA total mitigated - 25703

Calphayus

DTPS - 2382
% internal/elemental - 14.5%
melee/ranged - 70.6%
Fight length - 8:38
RW total mitigated - 31240
SA total mitigated - 23475

Raptus

DTPS - 1421
% internal/elemental - 2.9%
melee/ranged -10.5%
Fight length - 7:33
RW total mitigated - 11360
SA total mitigated - 17637

Council

DTPS - 2636
% internal/elemental - 16.1%
melee/ranged - 44.8%
Fight length - 9:03
(I tanked Calphayus and Brontes.)
RW total mitigated - 34080
SA total mitigated - 30897

Writhing Horror

DTPS - 3288
% internal/elemental - 0.9%
melee/ranged - 41.5%
Fight length - 5:04
RW total mitigated - 14200
SA total mitigated - 23851

Operator IX

DTPS - 1926
% internal/elemental - 0%
melee/ranged - 69.8%
Fight length - 8:31
RW total mitigated - 25560
SA total mitigated - 35116

Kephess the Undying

DTPS - 2881
% internal/elemental - 1.9%
melee/ranged - 7.8%
Fight length - 4:31
RW total mitigated - 17040
SA total mitigated - 18271

Caveats:

In some fights the average unmitigated attack damage was difficult to determine. For example, in the Calphayus fight there are 3 different attacks named "Strike" which each hit for different amounts. Also, some attacks did not hit me often enough to determine a reliable average (Rising Slash and Driving Thrust on Raptus, Leap Slam and Thundering Blast on Council.)
I did not have perfect uptime on Dark Protection, the "Protected" healer buff, or the Survivor 2 piece set bonus. In some cases Overcharge Saber was active during the SA proc. I hope that taking the average unmitigated damage of each attack would correct for this.
In some cases the SA relic proc occurred during the uptime of Deflection, and I therefore shielded less attacks than I would have otherwise.
The effectiveness of the SA relic varies the most in fights with infrequent large hits (such as Raptus.) The RW relic would have won if I had shielded two less of the big hits in that fight . To be accurate, I would need to tank these bosses several times and record an average between fights.

Edited March 9, 2016 by Gardimuer
adding TFB numbers

LeeTone · March 8, 2016

Thanks for the data.

I'm actually using both RW and SA at the moment.

dipstik · March 8, 2016

great data. what i noticed was the average time between relic procs. the ideal durations between procs would be 40 seconds for the warding relic and 20 seconds for the absorb relic. the fights closest to this ideal are council and kephess. the 20.84 seconds interval between shield amp procs is just about perfect, whereas the warding proc interval has a 5 second lapse (most likely due to the 30% chance of proc, although (1-defense)*shield is close to 30% for m/r attacks, but 50% vs. 30% for f/t shieldable attacks). also, the ideal mitigated damage per second is 65 to 71 dps for the warding relic (40 to 44 second cd), while that of the amp varies on boss damage. you have a few fights where the warding relic gets into the 60's but for the most part it is not seeing the maximum uptime assumed.

with more data i would be comfortable setting the proc interval to 44 seconds or so for the warding relic ( i used to use 42, then switched to 40 recently). i have suggested using the warding+shield amp since 4.0 dropped for sins, but for a single tank relic build, perhaps amp may be better (for sins, since i suggest the on use relic for jugg and pt, after warding).

to get 3110 damage mitigated every 40 seconds would require dps*.068*12=3110, gives dps of 3811 (assuming 6.8% absorb increase and 12 seconds of shield amp in a 40 second interval). this differs from my methodology, in that i calc the overall change in squish for each relic (time weighted) and take the difference from base squish to calc how much dps is required from the boss to make the relic in question equal to the warding relic. the equation dps*absorb*12=ward_bubble can be used to estimate which relic may be best for the fight in question.

the amp relic will outperform the warding relic if the amp proc happens during a high damage portion of the fight. for high internal/elemental fights the warding relic will mitigate what the shield relic cant, so that is something to keep in mind. bosses that have hard shieldable hits have a higher chance of making the amp relic outperform the warding relic. for instance, for tyrans, a single shielded thundering blast during the amp proc will mitigate 2600 damage, about as much as the warding relic over 40 seconds.

Edited March 8, 2016 by dipstik

Gardimuer · March 9, 2016

to get 3110 damage mitigated every 40 seconds would require dps*.068*12=3110, gives dps of 3811 (assuming 6.8% absorb increase and 12 seconds of shield amp in a 40 second interval). this differs from my methodology, in that i calc the overall change in squish for each relic (time weighted) and take the difference from base squish to calc how much dps is required from the boss to make the relic in question equal to the warding relic. the equation dps*absorb*12=ward_bubble can be used to estimate which relic may be best for the fight in question.

the amp relic will outperform the warding relic if the amp proc happens during a high damage portion of the fight. for high internal/elemental fights the warding relic will mitigate what the shield relic cant, so that is something to keep in mind. bosses that have hard shieldable hits have a higher chance of making the amp relic outperform the warding relic. for instance, for tyrans, a single shielded thundering blast during the amp proc will mitigate 2600 damage, about as much as the warding relic over 40 seconds.

Speaking of DTPS, shouldn't we consider only the DTPS during parts of the fight where tanks are taking damage? In Tyrans there is downtime after tank swaps. In Calphayus there is downtime after swaps and while waiting for portals to open, etc. In Raptus I couldn't tell what might be downtime due to how seemingly random the damage spikes were, so I didn't bother trying to calculate that. In Council the only downtime is the end of phase two when Brontes does her beam and I get my crystal. In Writhing Horror there was downtime after a tank swap when the Jealous Male was dead. In Op9 there is downtime while I channeled at my panel and for a short time after Black Obtuse. In Kephess there is downtime after tank swaps and between phases.