You already know
a lot that is wrong.

This is not an abstract question for us. Six Plane of Sky reward tooltips on this site carry percentage haste — five of them the identical +41%, which is a copied constant rather than five readings — they are marked suspect in place rather than deleted.

The evidence

• T3EQL Tools Slow, haste & crowd control, read 14 Aug 2026States flat attack-speed values for Legends, with a worked example.
• T5eqlwiki Haste Guide, read 14 Aug 2026Classic percentage formula, and the surrounding content describes classic-era raid targets. Last edited 4 Aug 2026, which is after launch and does not by itself make the mechanic current.

What would settle it. One screenshot of a Legends haste item tooltip. If it reads a bare number rather than a percentage, EQL Tools is right and every percentage figure on this site is a classic import.

This is the shape of mistake this register exists for, and an auditor made it: reading the classic sentence inside the structured NPC page and grading our survey against it. The survey had read the item record and was right. A tier 5 sentence inside a tier 2 container is the most dangerous object in this ecosystem, because it wears the wrong clothes.

The evidence

• ReportFirst-hand report, 14 Aug 2026The boots seen dropping for other players, confirmed by guildmates. First-hand and unparsed, so it carries no tier: the scale ranks readings of documents, and this is neither a document nor a log.
• T2eqlwiki Journeyman’s Boots item recordNames Najena and Drelzna as the source.
• T5eqlwiki Drelzna NPC pageCarries the Project 1999 sentence about the 1999 quest change, verbatim.

What would settle it. One combat log line showing the loot. No parsed session here holds a Drelzna kill, so this stays a report until a parse replaces it.

Asked to be more certain because the answer was worth dozens of hours, the same tool returned more sources and more specific detail. Confidence rose; evidence did not. That is the failure mode every player using an AI assistant will meet, and it is worth more than this one spell.

The evidence

• T2eqlwiki Buff Lines, read 10 Aug 2026Lists Cantata of Replenishment at +13, Cassindra’s Chorus of Clarity at +7 (+14 pulsed) and Clarity II at +11 in its mana regeneration tables. Slots run 1–8 plus a Layer 2. Paragon of Spirit does not appear.
• —eqlwiki Paragon of Spirit, read 10 Aug 2026No such page. HTTP 404.

What would settle it. Cast Paragon of Spirit, then have a bard sing Cantata of Replenishment and an enchanter cast Clarity on the same target. A screenshot of the buff bar showing all three at once settles it outright. The combat log adds the timestamps and any overwrite message — if one displaces another, the log shows the displaced spell wearing off within a second of the new cast rather than at its natural duration.

Raised 10 August 2026, after an AI assistant gave two confident and differently-worded answers.

eqlwiki’s Kedge Keep page carries {{Classic Era}}, discusses breathing at length, and says nothing whatever about weapon types underwater. Its search returns zero results for “underwater” on a wiki whose own Kedge Keep page uses the word repeatedly, so the search index is unreliable and an empty result there proves nothing.

The evidence

• ReportA player report, 9 August 2026First-hand, in Kedge Keep, using a sword and landing hits. Unconfirmed: a report of play, not a parsed log.
• T5eqlwiki Kedge Keep, read 9 Aug 2026Classic Era import. Silent on weapon types underwater. Neither confirms nor denies.

What would settle it. One combat log line showing a slashing or blunt weapon landing normal damage on a mob while submerged. If a penalty exists it should show as a visible drop in average damage against the same mob type fought on land.

Credited on the credits page.

So the two halves separate. Doing the quest and being able to hold the reward are different permissions, and only the second is still a class gate — which a trio carrying Paladin would clear. This is one epic, on one character. It is a reason to test the other fifteen, not evidence about them: we have no report on any other epic and do not assume this generalises.

The evidence

• ReportA player report, 10 August 2026First-hand: “can now confirm can do all parts of the fiery avenger including final turn in as a non paladin”. Reported with an in-game screenshot of the completed item. A report of play, not a parsed log.
• ReportItem inspection, 10 August 2026The reward reads Class: PAL, Race: ALL, Lore Equipped, No Trade. The class restriction on the item itself has not changed.

What would settle it. The same run on a second epic with a different class gate would show whether this is the rule or one quest’s wiring. A log of the final turn-in dialogue would raise this from a report to a measurement.

Credited on the credits page. The same player raised the underwater weapon question above.

That reframes a raid week. A first clear at D0 costs nothing at D1 and above, so there is no reason to open at a difficulty you are unsure of. The per-difficulty part is a player report, not a patch note, and the notes neither state it nor contradict it.

The evidence

• T1Patch notes, 23 June 2026“you will only receive 1 piece of loot per named raid creature that has its own lockout. The raid bonus loot lockouts reset every Tuesday at 8AM PST.”
• T1Patch notes, 28 July 2026“Killing a raid boss while you have a loot lockout will now give one guaranteed drop from that boss’s unique treasure tables.”
• ReportAnnalise (AnnaWulf), 10 August 2026“weekly lock out per difficulty and then daily lock out for 1 loot item till a tues reset”. She also reports a D4 Lord Nagafen taking a long time even with a group, and D1 hit-point pools being a drag solo.

What would settle it. Kill one boss at D0, then open the same boss at D1 and check whether the full loot lockout applies or only the daily one.

Per-difficulty behaviour reported by a player, 10 August 2026. Credited on the credits page.

This invalidates a large amount of inherited advice in both directions, and it is the single biggest reason a classic guide can be confidently wrong about Legends. What level the active trio runs at is a separate question with a different amount of evidence behind it, and it has its own entry below.

The evidence

• T1Official documentation and the level 11 lockPublished behaviour, not inferred. The deity, race and primary class unlock tokens are priced in the Producer’s Letter of 8 July 2026, which only makes sense because those choices lock. That covers the structure — three classes, two at creation, a third at 10, locking at 11 — and nothing beyond it.

What would settle it. Settled.

This sat inside the multiclass entry above until 31 August 2026, asserted at T1 and marked Settled on evidence that spoke only to the level–11 lock: three claims, one source, one badge. That is the exact failure this page exists to refuse, on the page that refuses it. It is carried as open rather than deleted, because the claim may well be right and removing it would lose the question along with the error.

The evidence

• —No source readThe Producer’s Letter of 8 July 2026 is cited above for the multiclass locks. It does not address what level the trio runs at, and no other source has been read for it.

What would settle it. One read of the Producer’s Letter of 8 July 2026 settles it either way, and it is the cheapest evidence available.

And it starts earlier than published. The published claim is that named mobs run multiclass from D2. In Castle Mistmoore at D1, two ordinary trash types backstab while the same types cast Root, Screaming Terror and Shadow Vortex. Backstab is a rogue ability and a spell list is not, so that mob type carries two kits at D1, on trash.

The evidence

• Tier MMeasured in play, Castle Mistmoore at AwakenedBoth trash types backstab, and both are logged casting. Melee reaches 39 and backstab 168, which is why no single average is printed for a mob that backstabs.
• T1Zone line on entryYou have entered The Castle of Mistmoore 1 (Awakened). The number and the name agree, and the modal loot drop of +1 agrees independently.

What would settle it. Settled for D1 and D2, and for the kits: bosses measured at D3 and D4 carry every spell each cast, Cazic-Thule and Innoruuk among them. What is still pinned by nobody is D3 and D4 hit points — damage to kill bounds them from above and no further.

We carried this as open while the patch note said it plainly the whole time. The lesson is about where we looked, not about the answer.

The evidence

• T1Developer patch notes“Removed placeholders from and lowered maximum respawn times in several dungeons: The Hole, Nagafen’s Lair, Lower Guk, Lair of the Splitpaw, The Warrens, Castle Mistmoore.” That is the whole list. This register printed the same sentence with Upper Guk, Crushbone, Befallen, Blackburrow, Najena appended inside the quotation marks, badged T1, until 18 August 2026 — when the note was fetched and stored verbatim at sources/raw/2026-07-28-eql-update-notes.txt. Five zone names had been added to a developer quotation.
• ReportFirst-hand playConfirmed in play across these zones before this site existed, and again at one camp across consecutive cycles with no placeholder seen.
• T5eqlwiki individual mob pages, againstDrelzna: “Her placeholder is a necromancer that spawns in front of the chair.” Classic text nobody has revisited. It is wrong.

What would settle it. Settled by the patch note for the six it names, and on weaker sources for three more. What it does not settle is the second half of its own sentence — maximum respawn times were lowered and no figures were published, so every respawn on this site is a pre-patch ceiling rather than a current value.

The evidence

• T5eqlwiki Per-Level Hunting GuideProject 1999 import by origin, with its own internal citations to P1999 forums.

What would settle it. Settled.