Quick Answer: The launch problem list has three tiers. The heavyweight is the Fatal Error crash cluster. Work the support-suggested checklist first — update your driver, disable NVIDIA Smooth Motion, disable Intel AI Boost, turn off Mandatory ASLR in Windows Exploit Protection — but know that at least four players report it doing nothing, and support says AMD solutions are still being worked on. Crashes are expensive here: a solo player who crashes mid-raid re-enters dead with gear gone, while trios can reconnect. Tier two is a KERNEL_SECURITY_CHECK_FAILURE blue screen tied to GPLDriver.sys, with a two-player-verified workaround that disables real Windows security features, plus a Steam admin-rights fix for detection errors. Tier three is the server story: rubberbanding and latency spikes reported worsening around August 4 — officially under investigation, not fixed — plus no Oceania servers and a stuck-spectator softlock. The August 6 client update cleared a batch of crashes and softlocks but named neither the Fatal Error cluster nor the blue screen, and on August 7 the developers stated plainly that a one-shot fix for every lag issue may not be possible — expect gradual improvement across updates, not a single patch.
The fights you least want to lose to a Fatal Error. For solo players, a crash here costs everything carried. (Screenshot: Bellring Games)
Version note: Current for the August 7 server-side update, checked August 10, 2026 — no client patch has shipped since, and this is the week the developers named for the next one. Sourced from the developers' day-one known-issues list, the launch patch notes, the August 6 patch notes, the August 7 server update notice and player threads, all linked inline.
Fatal Error Crashes: The Big One
The game dies to a "Fatal Error" dialog at launch, in the camp, or mid-raid. It's the most-reported technical problem of the launch window (main thread, August 4 thread), and reports come from Steam and Game Pass installs alike.
Work the checklist in this order. Two forum accounts post it repeatedly with official-sounding templates but carry no developer badge, so treat it as support-suggested rather than developer-confirmed:
- Update your GPU driver.
- Disable NVIDIA Smooth Motion (NVIDIA app or control panel).
- Disable Intel AI Boost in Device Manager, on Intel Core Ultra systems.
- Turn off Mandatory ASLR: Windows Security → App & browser control → Exploit protection settings → Mandatory ASLR off. Support calls this step "most effective for Intel/NVIDIA users."
Then set your expectations. At least four players across the two threads report the checklist, ASLR step included, changing nothing: "Still happens over and over," one says the game now "spam crashes and relaunches on its own," and a Game Pass user reports no effect. Support has said outright that "for players with AMD, we're still working hard to find better solutions" — so on an AMD GPU, the checklist isn't aimed at you and the real fix is the pending patch.
Three fragments worth trying while you wait, each a single unverified report: allow the game through the firewall, force DirectX 11, or check whether your crashes are camp-only (one other player confirmed that variant). File tickets through the official Discord — the support email reportedly returns auto-replies.
Protect your loot in the meantime, because a mid-raid crash is a death if you're solo. Trio players report reconnecting into the match; solo players re-enter to the death screen with everything carried gone (thread). Treat your safe pocket as crash insurance — full retention rules in Death & Loot Rules.
Blue Screen: KERNEL_SECURITY_CHECK_FAILURE (GPLDriver.sys)
Three reports, one solid workaround (thread). The symptom is a full-system BSOD with stop code 0x139 KERNEL_SECURITY_CHECK_FAILURE, and debugger analysis points at GPLDriver.sys — a shadow-stack violation, triggered around shader pre-compilation or the main menu.
The fix, confirmed by two players independently: Windows Security → Device Security → Core Isolation → disable Memory Integrity / kernel-mode Hardware-enforced Stack Protection, then reboot.
Read the caveat before you do it. Those are genuine Windows security mitigations, not game settings. Disabling them trades system hardening for a working game. That's your call, and re-enabling them once a proper fix ships is the responsible end of the trade. A third player with the same symptoms reports every conventional fix failing, so if this is your crash signature, Core Isolation is the only route with confirmations. No developer response in the thread.
"Steam Detection Error" / Crashes Tied to Admin Rights
If the game was ever run at a different privilege level than Steam, it can throw Steam detection errors or crash at launch. The fix from a community fix thread (steps in English and Spanish): fully exit Steam, run Steam as administrator, then launch the game from the library so both share the same privilege level. Single thread, but the poster describes multiple players with the same signature, and it costs nothing to try.
Stuck in Spectator After Dying
Roughly six reports across two threads (one, two), persisting since July 30. After a death — one player says every death to a player kill — spectator mode won't release. Esc does nothing and there's no return-to-camp option.
Two player-found exits: wait it out, since it eventually self-resolves for one reporter, or Alt+F4. Quitting from spectator costs you nothing the death didn't already take. No developer reply on either thread.
Servers, Lag & Rubberbanding
Netcode complaints are the launch's background hum, and they spiked in the days before this page was written. All community-reported:
- Rubberbanding and latency are worsening. A player who'd had 18-22ms since launch reports 38-44ms with rubberbanding starting August 3-4 (one corroborating reply, thread), and a separate rubberbanding thread is blunter. A well-upvoted Reddit thread claims one server apiece for the US and EU — a community claim, not verifiable against any official server list beyond the five named regions.
- A "latency is too high" warning can contradict a displayed 30-50ms ping. Single report, but a useful symptom to search if it hits you.
- No Oceania servers — that part is official. The OCE vote runs through Discord. High OCE pings are working as configured.
- Matchmaking Error 1005102 ("from message:1600") has at least two reports in English and Russian threads, and no known fix.
- Try the Performance graphics preset if you get frame stutter. It noticeably reduced hitching for one player, and it's the cheapest experiment on this page.
Watch List (Single Reports, Not Yet Patterns)
- Falling through the ground at raid start, then getting kicked
- Silent disconnects with no error dialog
- Keybinds not persisting between sessions — tracked with the rest of the input issues in the Controller & Console Guide
- The whole-team silence bug and the action-lock bug, where you can't attack, jump or swap weapons — the workarounds are die-and-revive or hide-quit-rejoin, covered with the other combat bugs in PvP Class Counters
- A cosmetic texture bug on the Celestial Mistveil skin
What the August 6 Update Fixed (and What It Left Alone)
The promised client update landed on August 6 at 09:00 UTC with full patch notes. Update your client through your app store — it was a no-downtime server update plus a client patch.
Crash fixes named in the notes:
- Gem affixes occasionally crashing the game during matches.
- The PS/Xbox friends list failing to retrieve friend information and crashing the game.
Softlock and match-flow fixes:
- The Soul Tree model trapping players who got stuck in it.
- Choosing to leave a match mid-battle triggering settlement normally but failing to actually exit.
- Hallowgrove's Gyldenmist dealing no damage under specific circumstances.
Loot and UI fixes: equipment looted from defeated rival Gyldhunters no longer displays as bound (details in Death & Loot Rules), stackable items stack correctly, "Allowed Combat Bags" displays correctly for Brandrgarde in Cataclysm, weekly task refresh times in Contracts and the Fate Chart are correct, plus fixes to the Contract page, avatar frames after a body-type switch, controller loot-drag sound spam, missing audio on some Cursed Moonwane skills, HDR text showing English under Chinese, and background music dying after a disconnect.
What it does not touch — check here first if you're still broken:
- The Fatal Error crash cluster is not in the notes. No item names it, so the checklist above is still the whole answer, including for AMD users support said it was still working on.
- The GPLDriver.sys blue screen is not in the notes.
- Stuck-in-spectator is not named, though the mid-battle exit fix is adjacent enough to re-test.
The August 7 Server Update and the Lag Statement
A day later the developers pushed a server-side update — no client download — fixing one thing: cancelling a Polearm and Shield charge abnormally triggering the Focused affix's movement speed bonus. The same notice carries the most concrete statement yet on the problems this page tracks.
On lag, set your expectations by their own words. The investigation is ongoing, in-stage environments are complex enough that lag has multiple causes, and — the part worth reading twice — "a one-shot fix for every lag issue may not be possible," with the issues expected to "taper off with each successive update." That is an official prediction of gradual improvement, not a fix date. If your play is currently unenjoyable because of lag, plan around weeks, not one patch.
Three things they committed to:
- A Block feature ships in next week's update. Nearly 30 advertising accounts have been banned from chat so far; blocking will hide a blocked account's messages from you.
- Report cheating and illegal teaming in-game. They are adding feedback emails triggered by successful reports, so you learn what happened to the account you reported.
- Next week's client update launches Soul Harvest and opens Season Tasks, plus another batch of the most-reported fixes. Contents to be announced when confirmed.
New Since the August 6 Update: Freezes
Reports of the game freezing after the August 6 client update have roughly doubled, and a second thread opened on August 9 collected three more within two hours — "it is gradually appearing for more and more people", as one of them puts it. Between the two threads the pattern now has some shape: one player specifies an "Application hang detected" error with a freeze at character select, another hits it repeatedly on the skills menu, and a third describes the game process surviving the freeze and refusing to die even from the command line, forcing a full reboot.
Assume a hard reboot is the cost, and save your session plans accordingly — if the process cannot be killed, anything you were mid-raid with is gone. The developers still have not named this, and it is not universal (one reporter's friends run the same build cleanly), so it is a watch item rather than a diagnosis. If you were stable before August 6 and are not now, nothing about your hardware changed.
Separately, a player analysing stutter reports it behaving as a 3-5 minute warm-up cost at the start of a session rather than a constant drag, and another says a GPU driver update cleared it for them and everyone they know. Neither is confirmed, but a driver update costs nothing and the warm-up pattern is worth checking against your own first few minutes before you start changing settings.
"Ran Out of Memory" Crashes and the Paging File
One detailed report on August 9 posts the actual exception rather than the generic dialog, which makes it worth acting on even as a single case:
[File:...\Runtime\Core\Private\GenericPlatform\GenericPlatformMemory.cpp] [Line:269]
Ran out of memory allocating 2097152 (2.0 MiB) bytes.
Last error msg: The paging file is too small for this operation to complete.
That second line is Windows, not the game. Check that your paging file is system-managed rather than disabled or pinned to a small fixed size before you assume the crash is a driver problem — the reporter is on a Ryzen 7 5800X with 32 GB of RAM, which is exactly the configuration where people turn the page file off because the RAM "should be enough". Their crashes came during high-density extraction sequences.
The expensive part is what happens after. The reporter describes the crash registering as a mid-match death rather than a disconnect, costing a fully socketed purple BlackArrow set. That matches the pattern already documented for solo crashes in Death & Loot Rules — the server does not treat a client dying as a safe exit.
If you are filing a gear-loss claim, confirm the channel is alive first. The same player reports the official Discord ticket system rejecting new submissions on August 9 and mail to the published support address timing out at the recipient server, with the developers' automated replies still pointing at Discord. They also report the support address being changed later that day. None of that is confirmed by an official announcement, so the practical rule is to keep a screenshot of your own send failure rather than assume a sent email means a filed claim.