# Temple of Sethraliss +12 — TIMED 28:00 · 4 deaths · Tyrannical, Fortified, affix 147 *2026-08-28 · 19:45 · `WoWCombatLog-082826_173447.txt`* > **AI-WRITTEN ANALYSIS** — The numbers come from the combat log. The verdict and notes were written by an AI reading them, and can be wrong in ways the numbers cannot. Written by Claude (Opus 5), via Claude Code. **Tip:** Divine Shield when Burrowquake lands; Burrow follows ## Burrow hits for more than any non-tank's entire health pool. It landed twice all key, both times on Hyporock, and the second was 1.5 seconds after a battle rez stood him back up inside it. Nobody lost this key. It was timed in 28:00 with four deaths, and two of those four were a single mechanic that no amount of healing could have answered. Merektha's Burrow hits for 775,082 before mitigation, every time. That number is worth holding next to the party's health pools: Doryhunky 1,669,173, Hyporock 739,840, Havoc DH 733,011, Outlaw Rogue 668,080, Arcane Mage 648,480. Burrow kills any of the four non-tanks outright from full health. Hyporock was at roughly 95% when the first one landed on him. So the two healer deaths are not a healing failure and not a "should have been topped up" failure. They are a positioning failure with no second chance attached. The proof that it is dodgeable sits in the same key: the first burrow phase at 11:25 caught nobody at all, with all five players standing between 4 and 10 yards of the boss. The second death is the one worth talking about, and it is only half his. He was battle-rezzed at 12:57. Burrowquake, the ground effect that marks the burrow phase, was re-applied to him alone at 13:12. He was dead again at 13:13, one and a half seconds after standing up, having got as far as pressing Avenging Wrath. Whoever pressed that rez put him back inside a live mechanic. **What to change** 1. Hyporock: treat Burrowquake landing as a four-second countdown on a hit you cannot survive. Divine Shield is the answer and it was off cooldown for both deaths. Blessing of Protection does nothing here, because Burrow is Nature damage. 2. Nobody offer or accept a battle rez while Burrowquake is still on the ground. The second death was 1.5 seconds after standing up. 3. Doryhunky: interrupts were the soft axis here, 27 stops against 34 predicted, with Lightning Bolt leaked 11 times of 59. 4. Do not act on the utility score for either player. Both denominators are dominated by buttons that should not be pressed on cooldown. ## The run | | | |---|---| | on the clock | **28:00** — 26:04 fighting | | deaths | **4** — timer agrees (4.03 implied) | | party damage taken | **247,488,070** — 158,281/s while fighting | | enemies killed | **210** | | interrupts | **85/107** — 79% | | dispels | **24/26** — 6,333,131 came through | ## Players | player | spec | role | ilvl | DPS | HPS | taken/s | deaths | grade | |---|---|---|---:|---:|---:|---:|---:|:---:| | Doryhunky | Veng DH | tank | 311 | 111,176 | 36,180 | 67,348 | 0 | **B-** 73 | | Hyporock | Holy Paladin | heal | 311 | 16,793 | 90,018 | 23,600 | 2 | **B** 78 | | Outlaw Rogue | Outlaw Rogue | dps | 305 | 195,063 | 4,460 | 19,620 | 0 | **B** 78 | | Havoc DH | Havoc DH | dps | 310 | 194,236 | 10,147 | 23,439 | 0 | **C+** 70 | | Arcane Mage | Arcane Mage | dps | 306 | 201,736 | 480 | 17,989 | 2 | **D+** 60 | ### Doryhunky — Veng DH — B- (73/100) **Where the points went** - **-9 interrupts (rank)** — scored 66: 27 stops vs 34 expected - **-5 mitigation** — scored 81: 97% uptime of the 100% this kit can hold, 100% of worst windows covered - **-5 damage (rank)** — scored 71: 111,176 dps, 56% of the average DPS
All axes - **survival− 100** [no charge] — 0 death(s) - **mitigation 81** [w 27%] — 97% uptime of the 100% this kit can hold, 100% of worst windows covered - **damage~ 71** [w 18%] — 111,176 dps, 56% of the average DPS - **mechanics− 100** [no charge] — 0% of intake was damage comparable players did not take - **interrupts~ 66** [w 27%] — 27 stops vs 34 expected (group stopped 79%, not scored here) - **activity 75** [w 15%] — 97% engaged while alive in combat - **utility 75** [w 12%] — 1 used of ~5 the kit offered - _never pressed: Imprison_ - _not counted as button presses (procs, ticks): Fracture, Soul Fragment_
### Hyporock — Holy Paladin — B (78/100) **Where the points went** - **-8 survival** — scored 50: 2 death(s) - **-3 response** — scored 86: put back 47% of the 67,629,860 that landed in the party's worst windows - **-3 activity** — scored 78: 91% engaged while alive in combat
All axes - **survival− 50** [−7.5 pts] — 2 death(s) - **throughput 96** [w 34%] — covered 88% of the 138,125,765 that needed an outside heal (49% of the party's intake was self-covered or overkill); 12,478,487 absorbed - **response 86** [w 25%] — put back 47% of the 67,629,860 that landed in the party's worst windows - **mechanics− 94** [−0.8 pts] — 2% of intake was damage comparable players did not take - **interrupts** — n/a. no interrupt on the bar - **dispel 84** [w 16%] — 24/25 of what one dispeller could reach (26 landed; the rest overlapped a debuff already being cleansed) - **activity 78** [w 15%] — 91% engaged while alive in combat - **utility 75** [w 10%] — 41 used of ~80 the kit offered - _overcapped Holy Power: 143 of 803 generated (18%)_ - _not counted as button presses (procs, ticks): Reclamation_
### Outlaw Rogue — Outlaw Rogue — B (78/100) **Where the points went** - **-10 damage (rank)** — scored 75: 195,063 dps, 98% of the other damage dealers' median - **-6 interrupts (rank)** — scored 82: 20 stops vs 17 expected - **-6 activity** — scored 78: 95% engaged while alive in combat
All axes - **survival− 100** [no charge] — 0 death(s) - **damage~ 75** [w 42%] — 195,063 dps, 98% of the other damage dealers' median - **mechanics− 100** [no charge] — 0% of intake was damage comparable players did not take - **interrupts~ 82** [w 32%] — 20 stops vs 17 expected (group stopped 79%, not scored here) - **activity 78** [w 26%] — 95% engaged while alive in combat - **utility** — n/a. no utility cooldown with a measured recharge yet - _overcapped Combo Points: 570 of 4,290 generated (13%)_ - _overcapped Energy: 370 of 6,368 generated (6%)_ - _never pressed: Evasion_ - _not counted as button presses (procs, ticks): Blade Flurry, Pistol Shot_
### Havoc DH — Havoc DH — C+ (70/100) **Where the points went** - **-11 damage (rank)** — scored 75: 194,236 dps, 97% of the other damage dealers' median - **-9 activity** — scored 67: 91% engaged while alive in combat - **-8 interrupts (rank)** — scored 76: 14 stops vs 17 expected
All axes - **survival− 100** [no charge] — 0 death(s) - **damage~ 75** [w 42%] — 194,236 dps, 97% of the other damage dealers' median - **mechanics− 86** [−2.8 pts] — 3% of intake was damage comparable players did not take - **interrupts~ 76** [w 32%] — 14 stops vs 17 expected (group stopped 79%, not scored here) - **activity 67** [w 26%] — 91% engaged while alive in combat - **utility 0** [w 0%] — 0 used of ~10 the kit offered - _overcapped Fury: 2,816 of 28,652 generated (10%)_ - _never pressed: Consume Magic, Sigil of Misery_ - _not counted as button presses (procs, ticks): Felblade, Shattered Souls_
### Arcane Mage — Arcane Mage — D+ (60/100) **Where the points went** - **-9 damage (rank)** — scored 78: 201,736 dps, 105% of the other damage dealers' median - **-9 survival** — scored 50: 2 death(s) - **-8 activity** — scored 69: 92% engaged while alive in combat
All axes - **survival− 50** [−9.0 pts] — 2 death(s) - **damage~ 78** [w 42%] — 201,736 dps, 105% of the other damage dealers' median - **mechanics− 67** [−6.5 pts] — 8% of intake was damage comparable players did not take - **interrupts~ 77** [w 32%] — 15 stops vs 17 expected (group stopped 79%, not scored here) - **activity 69** [w 26%] — 92% engaged while alive in combat - **utility 79** [w 0%] — 3 used of ~67 the kit offered - _never pressed: Ice Block_
*Axes marked ~ are ranked against the other players in this key, not measured on their own. Someone is always in the middle.* ## Deaths 4 deaths in 3 separate incidents. The cascade at 13:14 (2 deaths in 5s) was the big one. **What was landing while people died** — Counts only the seconds before each death, not the whole key. | spell | in | damage | killing blows | |---|---:|---:|---:| | Burrow | 2 of 4 | 1,406,821 | 2 | | Volley | 1 of 4 | 874,857 | 1 | | Scouring Sand | 1 of 4 | 565,879 | — | | Burrowquake | 1 of 4 | 495,175 | 1 | | Refraction | 1 of 4 | 91,300 | — | ### Death · 5:07 · Arcane Mage · dps **5:07 · Arcane Mage · dps** — killed by **Volley** for 874,857 from Sandswept Marksman took **1,440,736** · in **0.1 s** - **Volley** (killing blow) — 61%, 874,857 - Scouring Sand — 39%, 565,879 **no defensive pressed** · 1,599 healed · Hyporock ### Death · 12:52 · Hyporock · heal **12:52 · Hyporock · heal** — killed by **Burrow** for 735,053 from Merektha took **735,053** · in **one hit** - **Burrow** (killing blow) — 100%, 735,053 Shield of the Righteous · **no healing landed** ### Cascade · 13:14 · 2 deaths in 5s — the healer went first **13:14 · Hyporock · heal** — killed by **Burrow** for 671,768 from Merektha took **671,768** · in **one hit** - **Burrow** (killing blow) — 100%, 671,768 **no defensive pressed** · **no healing landed** **13:19 · Arcane Mage · dps** — killed by **Burrowquake** for 48,597 from Merektha took **586,475** · in **14.0 s** - **Burrowquake** (killing blow) — 84%, 495,175 - Refraction — 16%, 91,300 **no defensive pressed** · 388,811 healed · Havoc DH, Hyporock ## Where the time went | | seconds | | |---|---:|---| | combat | 1,564 | 26:04 | | routing | 52 | 0:52 | | death penalty | 60 | 1:00 | | lead in | 10 | 0:10 | Costliest pulls — *Cost is the time the pull lost: a wipe loses the whole pull, a pull that held loses only its deaths.* | pull | at | fighting | enemies | deaths | taken | cost | |---:|---|---|---:|---:|---:|---:| | 2 | 4:55 | 8:40 | 79 | 4 | 91,169,298 | 1:00 | ## What the group left up **Casts that could have been interrupted and were not** - Lightning Bolt: 11 of 59 got through - Venom Bolt: 11 of 26 got through **Debuffs that could have been dispelled and were not** - Cytotoxin: 2x, up for 20s in total, 6,333,131 damage **Damage one player took that similar players did not (check the raw hits before acting on this; a spell only one role can take looks the same)** - Arcane Mage: 2,094,141 (mostly Spark Step) - Havoc DH: 1,259,836 (mostly Siphon Energy) - Hyporock: 584,434 (mostly Spark Step) **Hit most of the party (not graded: everyone took it, so nobody is charged for it)** - Caustic Stomp: 5/5 players, 204 hits, 13,026,385 (5% of party intake) - Burrowquake: 5/5 players, 268 hits, 9,310,276 (4% of party intake) - Arrow Barrage: 4/5 players, 120 hits, 7,309,910 (3% of party intake) - Release Charge: 5/5 players, 26 hits, 6,791,657 (3% of party intake) - Cytotoxin: 4/5 players, 84 hits, 6,333,131 (3% of party intake) - Induction: 5/5 players, 25 hits, 4,912,870 (2% of party intake) ## Notes ### What Burrow actually does Reconstructed from this log, not from a guide. Merektha's burrow phase opens by stamping a debuff called Burrowquake on everybody standing in the area. At 11:25 that was all five players; at 12:47 all five again; at 13:12 it was Hyporock alone, because he was the only one who had just been resurrected into it. Burrowquake itself is minor. It ticked on him for four seconds while he sat between 91% and 99% health. Burrowquake is not the damage. It is the timer. Roughly four seconds after it lands, the boss erupts under one player, and that eruption is Burrow: ``` 11:25 Burrowquake on all 5 -> Burrow hit nobody 12:47 Burrowquake on all 5 -> Burrow hit Hyporock, 735,053, dead 13:12 Burrowquake on Hyporock -> Burrow hit Hyporock, 671,768, dead ``` Two Burrow hits in the entire key. Both on the same player. The base damage was identical both times, 775,082, so the second was smaller on the meter only because Arcane Mage's Protective Toadstools absorbed 41,234 of it. One detail decides which button is the answer: Burrow is Nature damage (school 8). Blessing of Protection only stops physical damage, so it does nothing here. Divine Shield does, and it was available for both deaths. The last time it was pressed in this key was 0:26, twelve minutes before the first death, against a measured 206-second recharge. ### Hyporock: strip out the two deaths and this was the best healing of the night - Throughput 100. What that means: of the 138,125,765 damage that actually needed an outside heal (the party's intake, minus what they self-healed, minus overkill), he covered 88%. The scale expects 73%. That is the top of the band and the only 100 he has posted on that axis in six keys. - Dispel 95: 24 of the 25 debuffs one dispeller could physically have reached. Cleanse recharges in 7.7 seconds, so "reachable" already allows for two debuffs landing at once. - Mechanics 94, the only sub-100 of his night. The single charge is Spark Step, which he took 5 times where the comparable player took 1, worth 584,434. Real, but it is 1.5% of what he took. He took 38,427,279 damage, more than any damage dealer in the group including the melee. That reads alarming and mostly is not. Blessing of Dawn deliberately redirects a share of the party's damage onto him, and a mechanics score of 94 says almost none of his intake was damage a comparable player avoided. Do not chase this number. ### Ignore the utility score. It is measuring Hammer of Justice. Hyporock's utility axis was his lowest in all four keys tonight (67 here, 58 in Blinding Vale, 75 and 78 in the two Nalorakk keys) and it should not be acted on. Here is what the denominator is made of. The axis divides utility used by "presses the kit offered", and it prices every button by how often its cooldown came back. Over roughly 27 minutes alive that gives: ``` Hammer of Justice 34.2s recharge ~48 presses "offered" Blinding Light 87.2s ~19 Blessing of Sacrifice 136.9s ~12 Aura Mastery 186.9s ~9 Blessing of Protection 304.1s ~5 ``` Crowd control is in the numerator at all only because Holy Paladin has no interrupt, which is a fair rule. But it means about 70% of the presses "offered" are a single-target stun and a short-range blind, priced as though they should be pressed on cooldown for a whole dungeon. They should not be. This is the same failure the project already caught and reverted once, with Blessing of Freedom: a situational button cannot be priced by availability. Worth noting alongside it, Lay on Hands is not in the externals list at all, so the three times he pressed it counted for nothing. Read the dispel line and the externals he actually used. Do not read the number. ### Doryhunky: a solid key, and his weakest interrupt line of the four A- (86.0), no deaths, mechanics full. The one axis with something in it is interrupts, at 69, his lowest of the night. He landed 21 kicks and 6 other stops, 27 in total, against the 34 that his position and uptime predict for a tank who is in every pack from the first second. The group stopped 85 of 107, or 79%, and leaked Lightning Bolt 11 times of 59 and Venom Bolt 11 of 26. Two caveats before anyone acts on the rest of his sheet. His mitigation score of 100 is not information: with the aura-parse bug fixed, every tank in this corpus now scores at or near 100 and the axis no longer separates anybody. And his utility score of 93 comes from one press out of about five offered, because for a Vengeance DH that axis prices exactly two buttons while all three sigils sit outside it as crowd control. Of his seven axes, three are not currently measuring what their names suggest. What is left and is real: survival, mechanics, activity and interrupts. He is clean on the first three. ### Doryhunky across all four keys: the damage axis is reading the roster, not him Grades on the night: A- 87.3, A- 86.6, B+ 83.7, A- 86.0. That is about as flat as this tool gets. The damage axis was 44, 23, 35 and 46 — mean 37, and by a wide margin the lowest thing on his sheet. It is 12% of his grade and it is the only reason his letters are not higher. Here is why it should not be read as a fault. The axis measures his damage as a share of the party average, on a band where 36% scores zero and 80% scores full marks. His shares tonight were 0.555, 0.461, 0.515, 0.564. That band was cut when this corpus was almost entirely him in +2 to +9 keys, where his own median share was 0.670, and it has not been re-cut since. What actually changed is the denominator, not him: ``` his dps share implied party avg 18-19 Aug, ilvl ~283 74,507 0.678 112,344 28 Aug, ilvl ~311 100,415 0.535 193,075 ``` His own output is up 35%. The damage dealers around him are up 72%. A tank's damage scales with his gear; a damage dealer's scales with gear and with bigger pulls and longer fights at higher keys. The share had to fall. The control that settles it, across all 46 archived keys, at +11 and above: ``` Doryhunky Veng DH median share 0.552 (n=14) A Prot Paladin Prot Paladin 0.554 A Fury Warrior Prot Warrior 0.554 A Blood DK Blood DK 0.761 A Blood DK Blood DK 0.760 ``` The two non-Death-Knight tanks land on his number to three decimals. The two Blood DKs are twenty points clear of all three. That is four tank-keys of control, so it is suggestive rather than settled, and each of those keys was a different group — but it points at spec and content, not at effort. What is real on his sheet is interrupts: 79, 83, 80, 69, mean 78. He is the single largest source of stops in every key of the night (22, 24, 33 and 21 kicks, plus crowd control), and the axis still reads slightly under the share his position and uptime predict, because a tank is in every pack from the first second. Temple of Sethraliss, at 69, is the one to look at. ### Hyporock across all four keys: strong night, one repeating item 87.9, 88.9, 87.0, 83.3. Two of those four are in his all-time top six of 46 archived keys, and the fourth is dragged down entirely by two deaths to a mechanic that kills outright. The healing itself improved through the night: ``` covered response Holy Power wasted Nalorakk +12 73% 44% 108 of 707 (15%) Nalorakk +13 75% 46% 77 of 737 (10%) Blinding Vale +13 76% 46% 132 of 1003 (13%) Sethraliss +12 88% 47% 143 of 803 (18%) ``` "Covered" is the share of damage that genuinely needed an outside heal which he put back; the scale expects 73%. "Response" is the share he put back inside the party's worst eight-second windows, on a band running 18% to 54%. Both went up in every key. His response figure is strikingly stable — across all 46 keys it has sat between 36% and 60% regardless of dungeon or level, which says his output rises by a consistent amount when damage lands. The one item that repeats, and it has been open since 20 August, is Divine Shield at death. Three deaths tonight, and the button was off cooldown for all three: ``` Blinding Vale 20:19 1,735,135 over 10.4s Divine Protection pressed, Divine Shield last used 8:26 Sethraliss 12:51 735,053 in one hit Divine Shield last used 0:26 Sethraliss 13:14 671,768 in one hit same ``` Fairness matters here, because the three are not the same shape. The two Sethraliss deaths were single hits larger than his health pool, with a four-second telegraph — surviving those needs the button pressed before the hit, not in reaction to it. The Blinding Vale death was collateral from a cone aimed at somebody else, though it did take 10.4 seconds and he pressed Divine Protection during it, so he knew he was in trouble and reached for the smaller of the two buttons. So: one clear reactive miss, and two that needed pre-emption off a telegraph. That is a real pattern and it is the single highest-value change available to him, but it is not three of the same mistake. Holy Power waste averaged 14% across the night and was worst in the last key, which is the ordinary shape of a tired fourth key. ## Method Each axis is scored 0 to 100 against what this run allowed, not against a fixed target. A player is only scored on things their spec can do. An axis that does not apply is left out and the other weights are scaled up. 100 means nothing was missed; a solid, even key usually lands near 75. Survival and mechanics carry no weight of their own: they start at 100 and take points off when something went wrong. *Generated 2026-08-29T20:20:48 · schema 5 · written analysis: Claude (Opus 5), via Claude Code*