Onimusha: Way of the Sword Performance, Settings & Fixes

The launch crash wave — and what actually works

Within two hours of the PC unlock, Steam reviews opened at “Mixed” (58%) on the back of players whose game was crashing on startup — often before the title screen: RE Engine crash dialogs at boot, shader-compilation hangs, crashes right after creating a save, crashes when toggling frame generation — most at zero hours played, many noting the demo ran fine. The wave hit the first-unlock time zones hardest (the global PC unlock lands at midday in China), which is why Simplified Chinese reviews carry most of the damage. Capcom has not acknowledged the issue — the official launch sticky remains a generic troubleshooting checklist with no known-issues list — so what follows is the community’s verified fix ladder, strongest first:

1. Cap the frame rate at 60 in the in-game settings — the most-confirmed fix (multiple independent reports, including on an RTX 5080). Several players report they could uncap again once past the crash point.
2. Leave frame generation off. Toggling DLSS frame generation is the most consistently reported crash trigger (multiple independent reports across RTX 40/50 cards).
3. Delete the install folder and reinstall — several players traced their crashes to a corrupted preload; uninstall, delete the leftover OnimushaWotS folder, reinstall. ⚠ worked for several, failed for at least one
4. Reset graphics settings to default, run as administrator, and check drivers meet Capcom’s floor (NVIDIA 596.49+ / AMD 26.5.1+), then the official checklist basics: verify files, delete shader cache, antivirus allowlist.
5. The REFramework method — dropping the dinput8.dll from REFramework’s nightly build into the game folder has a 20+-reply thread of confirmations behind it and is marked as the thread’s answer. The poster blames DRM interaction ⚠ community attribution, no technical proof — but the fix works regardless of why. One caution added September 10: a player reports REFramework itself causing an instant crash at boot ⚠ single report. It is a mod on a nightly build, so if your crashes began after installing it, remove the dll first. The other direction got a third data point on September 17: a player whose Family-Sharing copy never reached the main menu — the window resized itself, then a Capcom crash report — reported it fixed by installing REFramework after driver updates and everything else he could find had failed ⚠ single report; whether the shared-library copy matters is untested.
6. Two narrow symptoms with narrow fixes: a crash during the opening cutscene often clears on a simple restart, and the “scenery fails to load after skipping a cutscene” bug has a reported fix in running the exe in Windows 7 compatibility mode ⚠ single detailed report, reversible.

On causes, honesty over certainty: the frame-rate/frame-generation link is the strongest pattern; some players blame Denuvo but no technical evidence supports that; and the crash is not universal — plenty of machines, including an RTX 5090 running uncapped, launched fine.

Status at day thirty-nine (October 12): Steam sits at Very Positive, 90.96% (26,291 positive to 2,612 negative), still above 90%. Simplified Chinese is at 86.89% against English’s 93.93%, a gap of 7.0 points, unchanged from October 10 and down from 7.3 on September 28. By the standard we set on September 13 that is still a drift rather than a turn. Japanese is at 89.67%, on a pool of about 690 reviews. The part worth knowing stands: no patch did any of this. The build Steam serves is unchanged since September 2, forty days now. Capcom has said nothing technical. The official troubleshooting thread is at 173 replies; the newest, on October 11, is a crash during fast travel that the player says forces Windows to reboot, and the one before it, on October 9, a crash at Kiyomizu-dera after which the reloaded autosave leaves the area entrance blocked — both in the ledger below. The Japanese pinned thread sits at seven; the newest there, on October 2, is a player who tried every listed fix, never reached the title screen and asked for a refund. The company’s only Steam post since launch is still the million-copies celebration. We did not read the board between September 29 and October 7, or on October 11; each read covers everything since the one before. We have now read every comment on the board — 14,671 across the 1,353 topics we have seen, as of October 12, including 50 that were new since October 10 — and the developer badge appears on zero. We count topics we have seen rather than topics Steam lists, because the list only shows topics with activity in the last 30 days, so old threads drop off it. It does sit on the two pinned opening posts — that is just Capcom’s account having started those threads — and an earlier version of this page folded those into the count, which made the silence look less complete than it is. Scoped to replies, the number is zero.

September 15 — a correction to what we wrote on September 14. On September 14 this page said four players had converged on capping the frame rate below what the GPU can sustain, and presented that as the fix. That was an undercount, and only of one side. Having now read every thread in full, capping is not a September discovery — a player credited it on launch day — and the tally is mixed: at least 14 players say it stopped their crashes, and at least 10 capped and crashed anyway, three of them reporting so on September 14 alone and a tenth on September 15; the fourteenth supporter, on September 16, capped a 9950X3D and RTX 5090 at 140 ⚠ a sample drawn from 141 matching posts; counts are minimums. So a cap is worth trying, but it is not a reliable fix, and we should not have said it was. What the successful fixes do share is reducing load on the GPU: a lower cap, ray tracing off, G-SYNC off, or running at a lower resolution. Treat those as a checklist rather than a single cure. Two single reports from September 15 belong on the same list, unconfirmed: one player’s crash loop in the village stopped when he turned HDR off in the display options, and three players tie trouble to the card’s VRAM ceiling at maximum settings — two of them crashes, over 12 GB in one account with ray tracing and DLAA, and on September 16 a third whose stutter and steadily climbing memory use stopped when he dropped textures from High to Medium ⚠ single reports, nothing measured. The list also has its failures on record: on September 16 a laptop player with an RTX 4080 (12 GB) reported the crashes returning after anywhere from ten minutes to two hours despite, in his words, every trick in the book — several drivers, no overlays or background apps, a cleared shader cache.

The cleanest driver test on the board is older than anything above and changes nothing but the driver: “616.64 and 616.92 both make this game crash constantly for me, the stable driver for me is 616.56. I change no settings in-between the driver changes.” A second player separately fixed his crashes by rolling back to 610.88, and a third found relief by turning off an NVIDIA filter. Two players report 616.92 running smoothly — the second, on an RTX 5070, on September 16 — so it is not universal, and on September 20 a third player, whose stutter arrived with 616.92 in nearly every game he owns, went back to 581.80 for good. A September 15 report cuts both ways: a player who had run clean since launch updated to 616.92 on reaching Okuni’s Kabuki Dance and has crashed on every load since — which fits the driver reading and the location reading below equally well, so it counts cleanly for neither ⚠ single-player tests; no driver newer than 616.92 mentioned anywhere, re-checked September 21.

A newer pattern worth watching: several players report playing cleanly for five to seventeen hours and then crashing repeatedly from a particular point onward. At least four place it at the sandbar island segment of Okuni’s Kabuki Dance — the same stretch that produced one of the softlocks below — and two more at Arashiyama. Two more arrived on September 15: one reached the same Okuni mission before the crashes began but updated his driver at the same moment (the report above), and one — five or six clean hours on a 9800X3D and RTX 5080, then a loop that a reinstall only postponed — names no location. September 17 brought two more that fit the shape: four clean hours, then constant crashes from the moment the Strength in Refuge quest began, with NPC dialogue cutting out after its first words; and a player who finished the game and went back to the arena of the first Benkei fight — a spot he had revisited without trouble before — where it now freezes every time ⚠ two single reports. Nothing has been patched in that time, so this is not a regression; it may be that those areas are simply heavier ⚠ a cluster of reports, no controlled test; the location link is a hypothesis.

What follows is the earlier reading, kept for the record. The second crash cluster — repeatable mid-game crashes tied to specific areas — is now a 152-reply thread, and the frame-rate cap that looked like a consensus 24 hours ago picked up its first counter-examples while gaining no new confirmations. One player who needed a 90 cap reports being stable uncapped since updating drivers; another found the opposite of the advice — that a driver-level 60 fps cap was causing his crashes, and removing it stopped them; a third capped to 140 and says it “doesn’t fix it but it crash less now”. A new confounder arrived at the same time: NVIDIA shipped driver 616.92 WHQL on September 9, and the reports about it directly contradict each other — one player credits it with fixing uncapped play, another says his crashes started with it, a third rolled back to a July driver, a fourth updated, got it working, and refunded anyway. The driver notes do not mention this game ⚠ contradictory single-player reports on both sides; nothing measured. A day later the tally is four players saying 616.92 broke the game for them against two saying it is fine, and two different rollback targets have worked for different people: 616.56 and 610.88. The frame-cap advice also took another counter-example — a handheld player whose game “crashes every time regardless” of whether it is capped at 60 or uncapped — and still no new confirmations. So the honest position is weaker than it was two days ago: capping the frame rate helps many people and is still the first thing to try, but it is not the answer, and if you are on a fresh driver your results may not match anyone else’s. What does still hold: this is not an RTX 5000-only problem — RTX 4080 and RX 9070 XT owners report the same crashes.

The stutter fix that reproduced: keep the game off your Intel P-cores

The traversal stutter in the Oni Refuge and the Kyoto hub has one fix that reproduced across at least eight players on 12th-, 13th- and 14th-generation Intel chips: stop the game from running on the performance cores — Hyper-Threading off in the BIOS, or the process pinned to the E-cores only (Process Lasso makes it stick). A 13700K owner tested it across three generations with a 10th-gen negative control; two confirmers call the improvement partial; the cost is peak CPU throughput; a 9800X3D owner says it does not apply to AMD, and nothing else has reproduced for AMD. Stutter that lines up with autosaves is a separate bug with four single-player fixes. Every confirmation with its date, the how-to in order of cheapness, and what did not work are on the stuttering page ⚠ player testing, not instrumented; Capcom has not commented.

Crashing on an AMD card: the first settings that help

Almost every fix on this page came from NVIDIA owners. The first detailed AMD report arrived on September 10, from a player on an RX 9060 XT under SteamOS: a crash at an early cutscene went away once ray tracing was turned off, and white-square artifacts and freezing cleared with FSR 4 on Performance and ray tracing on Medium. A second RX 9060 XT owner, on Windows 11, reports the same first-cutscene crash ⚠ two reports on the same card, one fix. If you are on Radeon and crashing early, ray tracing is the first setting to drop.

The single-user fixes keep coming too, listed here without endorsement: enabling the game’s own V-Sync for a LiveKernelEvent 141 crash on load, switching ray tracing to screen-space reflections, and dropping the process priority below normal with Process Lasso, which two players report independently ⚠ one report each, except the priority trick. Boot crashes have a fix ladder; the area crashes now have one reproduced lead and a patch that doesn’t exist yet. The first official patch lands on the patch ledger the day it exists.

The softlock pattern: reload while a quest script is broken, and you can be locked out for good

Nobody else is covering this. We checked rather than assumed: Maxroll’s own Shadows Fall and Mount Oe walkthrough pages contain zero occurrences of “softlock”, “bug”, “barrier”, “crash” or “warning”, and IGN’s wiki does not mention one either.

We first called this a crash bug. That was too narrow, and the correction changes the advice. The forum is far larger than it looks from the front page — about 1,160 active topics across seventy-eight pages on September 17 — and because it sorts by most recent reply, a report nobody answers sinks out of sight permanently. Reading every thread in full, all 11,690 comments, turned up nine complete cases. Only one was posted in the last few days; the rest had been sitting unanswered since September 5-8. That is a statement about how deep we had to dig, not about a sudden outbreak. Side by side, crashing is not the common factor:

The shape they share: a rift or event script fails to finish, the game state gets reset — by a crash, a death, saving and quitting, or even just fast-travelling away and coming back — and the player returns outside the arena, behind a barrier, or in a room with no way out. The case that settles the cause is the second one: there was no crash at all. He simply could not find the rift, quit on purpose, and came back locked out. That is why we are no longer calling this a crash bug ⚠ nine independent reports, each a single account, none acknowledged.

A fingerprint you can check yourself. In the best-documented case, the player noticed that “Save Game” was greyed out in the pause menu while “Load Game” still worked — which he read as the engine still believing a quest was running. That is exactly what a stalled script would look like from the outside. In the same case, Steam Cloud then overwrote his good autosave with the broken state, which is how a recoverable situation becomes permanent.

What to actually do. Reloading the autosave does not rescue you — two of these players tried and got the same broken state every time, and a third worked through five recovery attempts (travelling away again, waiting elsewhere, reloading, restarting the game, dying on purpose) and every one failed. As of September 21 there is no known way back from a locked-out save. So:

• Keep a manual save from before each rift or scripted encounter, and rotate at least two slots. It is the only lever that has worked for anyone.
• If a rift fails to spawn or enemies stop arriving, do not save, do not quit, and do not fast-travel away. Correction, September 15: this line used to say “leave the area and come back first”. That was wrong — one player did exactly that and it is what locked him out. Load your manual save from before the rift instead.
• If it has already happened, close the game before Steam Cloud syncs and restore an older local save if you have one.

If you have no manual save, the honest answer is that restarting is a coin toss. On September 15 one player described three stalls in a single playthrough — the rooftop chase in What Lurks in the Shadows, the final rift in The Roots of All Evil, and a third he could not name — each of which cleared when he restarted the game, and another player says a single restart cleared a malice wall that would not break. Set those against the three players above whom a quit or restart locked out for good. A stall that clears on restart and a stall that becomes permanent look identical while you are standing in them, which is why the manual save is the advice and restarting without one is a gamble rather than a fix ⚠ two accounts of restarts clearing stalls against three lockouts; nothing reproduced.

Which quests to be careful in. On September 14 we named Shadows Fall as the most-affected mission; with the whole board read, The Light of Vengeance now leads with three lockouts, and two more players report its rift never appearing without yet having been locked out. The stall happens after the second conversation with the father, when the rift should open at the shrine. A fourth player wrote “I’m having the same problem” in the original thread on September 15 without describing his trigger, and was pointed at a Nexus mod that teleports the player anywhere on the map as a possible way out — he thanked the poster but has not said whether it worked, so it is a lead, not a recovery ⚠ one undetailed report; the mod suggestion is untested. Shadows Fall and Okuni’s Kabuki Dance have two each. For Shadows Fall, the two reports may well describe the same rift rather than one at each, so save before both ⚠ whether the two Shadows Fall reports are the same rift is unconfirmed.

The first workaround in this family — and read the scope before you pin hopes on it. A player stuck on the Mount Oe gauntlet, unable to get the Oni Gauntlet to register anything scannable, reported getting through it: “I got past it, removed mods, traveled back to town then back to that spot.” That is genuinely useful, and it also gives that particular complaint its second reporter, so it is no longer a possible one-off ⚠ one player’s fix, not independently repeated. But it is a fix for the gauntlet blocker, not for the rift lockouts above. It will not recover a save where “Save Game” has gone grey and you are standing outside a sealed arena. We are separating the two because conflating them would give people in the worse situation false hope.

We also suggested that the words “removed mods”might point at mods causing some of these blockers. The evidence has not backed that up: that remains the only report of a blocker clearing after mods came out, nobody has repeated it, and two other players say removing mods did not help. Removing mods is still a free thing to try when you are stuck, but there is no evidence that mods are behind the rift lockouts.

A different trap, from September 30: do not fast-travel away from a boss door if a cutscene is what carries you into the arena. A player lost a late-story boss fight, used the Spirit Mirror at the boss’s door to go back and upgrade gear, and then could not return. The arena, he says, is only reachable through a scripted scene in which another character carries the protagonist there, and after the trip away that scene no longer plays. The mirror at the door cannot be travelled to, and the route marker does nothing. He posted it twice, once in Capcom’s official thread, and had no answer by October 8 ⚠ single report, no replies that help. We are not adding it to the nine lockouts above, because no quest script stalled first; it is a one-way transition that did not replay. The advice is cheap anyway: upgrade before you walk through a boss door, not after losing the fight.

Mission-by-mission flags are on the mission list.

Save files that will not load (September 10)

One new report worth flagging separately because it is about save integrity rather than performance: a player who crashed during the final boss says the game now “refuses to load any save file that is in the end game castle area”, while a brand-new game starts fine — so the install works and the saves do not. One other player reports the same shape of problem, and on September 15 a third player, who crashed during the end credits after beating the final boss, said he has not been able to load the game since ⚠ three reports, unacknowledged, no workaround. If you are approaching the end of the story, this is a reason to keep a manual save from before the castle rather than relying on autosaves alone. Alongside it, day seven brought fifteen new bug threads and several new hard-stop locations: a post-fight freeze after a Kyoto boss, the town’s left side, the loading screen after the Yoshitsune cutscene (reported twice, both on Linux/Proton; on October 1 the second player got through by turning DLSS off, after removing the /WineDetectionEnabled:False launch option and turning ray tracing off had not helped him, while the first had needed only that launch option removed), a Nue boss failing to spawn in Nijo Castle leaving the quest stuck, and a Burai cutscene that never fires in “A Tempest at the Temple” ⚠ single reports each.

Stuck after the boat to the sandbar island (Okuni’s Kabuki Dance)

A new softlock class surfaced on September 8, from three players independently. In the mission “Okuni’s Kabuki Dance” you take a boat to a small island; die there without having activated a Spirit Mirror, or crash right after arriving, and on reload the boat is gone — no way back to the island, the mission cannot continue, and one player’s last manual save was two days old. One has filed a bug report with Capcom. A fourth player could not reproduce it on a Carnage run and asked whether the others use a teleport-anywhere mod, which would explain it; the reporters have not answered ⚠ three reports in one day, unacknowledged; mod involvement unresolved. Until it is understood: make a manual save before boarding, and activate the mirror on the island before you fight anything. The softlock has not moved since: both threads are unchanged, neither original reporter has answered the question about mods, and nobody has reproduced it on a clean install — so treat it as a real risk with an unproven cause. On October 11 someone else in the second thread suggested using a teleport-anywhere mod to get back to the island; nobody has said it worked, and it is the same kind of mod the fourth player suspected of causing the problem, so we do not recommend it ⚠ one untested suggestion. A different Okuni failure did turn up on September 10: a player reports every rift sealed at 1 of 2 quest progress with the stage cutscene refusing to trigger. On September 23 a second player, writing in Russian, described the same shape: at the shallows where the cutscene with Okuni should start, nothing plays, the “close the rifts” objective appears instead, and it cannot be done because the Kamo River shallows Spirit Mirror stays shut without that cutscene ⚠ two single reports, two weeks apart, no trigger named, no fix. Neither says what came before, so this variant is not counted in the nine lockouts below. On September 27 the Russian player reposted the problem in English in Capcom’s official troubleshooting thread, adding that they have no backup save and no way back to before the mission. A third report from the same mission arrived on September 24, in Chinese, on Onimusha difficulty, and it has a different shape: the boss fight and Okuni’s stage-dance cutscene both played, and afterwards the rift-sealing objective stuck at 0 of 2, so the main mission could not be completed ⚠ single report, no replies, no trigger named. Three players have now hit a stuck objective in this one mission, by at least two different routes. By October 8 none of the three had a reply that helped, and Capcom had not answered the repost. The Russian player called the mission “THE PERFORMANCE BEGINS” in English; no English mission list we use has that name, and the mission is “Okuni’s Kabuki Dance” in all of them, so it is most likely the Russian title translated back. The advice stands: make a manual save when the mission starts, so a stuck objective costs you one mission, not a session.

A separate, older thread — 80 replies by September 21 — pins the game’s traversal stutter to the Oni Refuge: one player re-downloaded the demo, reached the same spot at 170–200 fps, then hit constant stutter there in the release build ⚠ one player’s demo-versus-release comparison, and a 5060 Ti owner reports the same area. That is the thread the P-core fix came out of; the whole story now lives on the stuttering page.

PC: what the release build actually does

Context for those numbers: Windows 11 only, DirectX 12, 16 GB RAM and a 50 GB SSD at every tier, with DLSS and FSR (both with frame generation) supported — full spec table on the requirements page. A demo-era quirk to know: some players found DLSS options missing entirely; the community traced it to a driver/bundled-library mismatch ⚠ modder diagnosis, corroborated by an RTX 5060 Ti report in the official thread. Launch reviews all used DLSS without trouble, so treat missing options as a driver-update prompt, not a missing feature. Capcom’s launch-day troubleshooting post makes the floor explicit: NVIDIA driver 596.49+ or AMD 26.5.1+ — update before you blame the port.

Read your benchmark score like Capcom does

The free official benchmark (Steam/Epic, ~18 GB) is the honest way to test before buying, and Capcom publishes the exact score bands — reproduced verbatim here because no outlet has:

Two official footnotes worth knowing: the results screen reports average and 1% low fps, and enabling frame interpolation lowers the score slightly (the tool scores raw frames) while raising the frame rate you’ll actually see. For scale: a Steam Deck on the lowest preset with FSR scores about 46 fps average with 1% lows around 20.

The washed-out image and the HDR problem

The demo’s most-reported visual complaint was a grey, low-contrast image. Three things are true at once. First, part of it is deliberate — heavy volumetric fog is the game’s look, and removing it via mods restores contrast at the cost of atmosphere. Second, the practical fix: in the calibration screens, push the second brightness slider from its ~4 default to about 9 — the widely-confirmed community adjustment. Third, the release-build verdict is genuinely split: Prima Games’ review says color calibration was “touched up” since the demo and the washed-out look is “mostly eliminated,” while GLITCHED’s review still calls some scenes “horribly washed out” with scene-to-scene inconsistency ⚠ CONFLICT: one single-source review each way; most launch reviews don’t mention it. The honest read: improved but scene-dependent — keep the slider fix in your back pocket.

HDR is a harder verdict. Two independent demo-era measurements agree the implementation was poor — peak brightness around 670-680 nits despite a calibration screen implying 2,100, raised black levels, and aggressive contrast processing; one analyst’s summary was “HDR as bad as it gets.” No one has re-measured the release build yet ⚠ demo measurements; release status unknown. Until then, the pragmatic answer many players landed on: play in SDR with the slider fix. By September 19 the board’s own answer had moved on to mods: four accounts across two threads recommend the RenoDX HDR add-on (it sits on ReShade) for contrast and shadows, and one suggests NVIDIA’s driver-level RTX HDR instead — player recommendations, mod territory, and none of it measured ⚠ player-sourced, September 18–19; no release-build measurement.

The known-issues ledger

Every “unknown” row is exactly what we verify as launch reports come in — status changes land here and in the patch ledger as they happen.

Consoles in one paragraph each

PS5 / PS5 Pro: Push Square calls the Pro build “immaculate” — 30/60 caps or an uncap to 120 fps, with ray tracing toggleable at no perceptible frame cost; the base PS5’s performance mode reads slightly soft but never tempts you back to 30. The Pro’s gains over base are honest-modest — demo-era analysis found the two hard to tell apart. Switch 2: very smooth in the unlocked 30-40 mode, with the one caveat of visibly lower handheld resolution — the highest-review-scored version overall (full platform comparison). Series S: holds its 1080p/60 performance target per early accounts ⚠ single outlet. No platform has produced crash reports as of launch morning.

What engine is Onimusha: Way of the Sword? (Not Unreal)

It runs on Capcom’s own RE Engine, not Unreal Engine 5 — a search people make often enough that it is worth stating plainly. The confusion is understandable: RE Engine games and UE5 games land in the same visual neighbourhood, and most 2026 third-party releases of this size are Unreal. This one is not, and two separate developer interviews put it beyond doubt.

Director Satoru Nihei ties the whole project’s green light to the engine maturing: “It wasn’t until around 2020 that the timing finally aligned for us to gather our core members. Coupled with the fact that Capcom’s RE ENGINE had reached a high level of maturity, we were finally able to kick the project into full gear.” Art director Hiroyuki Nara describes a pipeline that ends inside it — Maya, ZBrush, Photoshop, Substance 3D and MotionBuilder for assets, with “VFX and lighting… primarily handled natively in RE ENGINE” — and names the reason Capcom keeps using it: “the ability to share knowledge and features Capcom has accumulated across multiple titles.”

The part worth knowing: this game made Capcom extend its own engine. Animation team lead Hiroshi Yugen told Game Informer that the Dynamic Dismemberment System — which calculates where your blade actually passes through an enemy or an object, rather than playing a canned cut — “is not something that’s existed in any other title before, and it’s something that, on a technical level, needed to be added to the RE Engine” by Capcom’s engineers. He also says the team hit no ceiling: “We didn’t actually have any limitations posed upon us when developing Onimusha using the RE Engine.” Per the 80.lv interview, a single sword clash weighs “enemy attack types, weapon vectors, and terrain slopes in real time.”

Why this matters practically. RE Engine is why the port behaves the way the benchmarks above show, and it is also why the community fixes on this page work: the modding ecosystem around it — praydog’s REFramework above all — is mature, which is how 32:9 ultrawide and a field-of-view slider exist at all before Capcom ships them. It also explains the one recurring complaint: several players in the crash threads blame RE Engine generally, noting the same stutter pattern in other recent Capcom titles ⚠ player opinion, no measurement either way.

Denuvo, honestly

Short version: the PC build ships Denuvo Anti-tamper, the first launch needs an internet connection, and Steam’s DRM notice caps you at five different machines within a day. Whether it costs frames is unmeasured — the long forum argument about it contains zero measurements on either side, and the demo cannot settle it because the demo shipped without Denuvo. Full breakdown, including what the activation limit does and does not restrict, on the Denuvo page.

Frequently Asked Questions