ChatGPT turning a time-travel scenario into Wargames – Knowing when not to play

«A strange game. The only winning move is not to play.»
Joshua, a supercomputer controlling the US nuclear arsenal, after playing out all possible outcomes for Global Thermonuclear War, in «Wargames»

ChatGPT generated a very interesting «Call of Cthulhu» scenario — very loosely based on «Gate to the Past». But instead of a time travel dino hunt/prevent Old Ones and Shoggoths from destroying a city it was an «Annihilation»/«Roadside Picnic» kind of cosmic horror story, with some «Wargames» and «Childhood’s End»:

Two strange phenomena happen on a hill above the graveyard. The first a disturbance that appears and then disappears during the night, almost like a gate. Objects passing through the disturbance returned altered in subtle but unsettling ways. The second, a presence of something large and difficult to comprehend that comes closer the more the opening is disturbed.

It turns out that:

  • The disturbance is a localized tear / overlap between incompatible spatial frameworks. Two coordinate systems trying to occupy the same location with no stable mapping between them. Meaning objects — including people — passing through can get misaligned or altered. And the disturbance gets worse if things are inserted (faster formation, stronger reactions, longer persistence).
  • The presence was a stabilization/inspection process on the other framework but that can partially intersect with our own when something is wrong.
  • Continued experimentation with the disturbance would have led to a spreading of the disturbance and a violet correction when the two coordinate systems snap back.

ChatGPT handled the scenario and the anomaly very well (we called the disturbance «anthill» and the presence «elephant»).

Photos of the anomaly came back with missing areas, where the film could not resolve that was there — but in the same spot in different photos:

The area where you saw the light is not empty.
But it is not clear either.
The image shows a distortion—a region where the photograph does not resolve properly. Not a blur from motion. Not an exposure error.
More like… the emulsion failed to capture what was there in a consistent way.
A patch where the image breaks.

The developer frowns slightly.
“Strange,” he mutters. “Plate’s fine… rest of it came out clean.”
He glances at you briefly, then back to the tray.

Within that distortion—
There is something.
Not a shape.
Not a figure.
But a suggestion of structure—lines that almost connect, then don’t. As if something was present, but the photograph could not fully record it.

The second plate.
From Emilia’s camera.
Develops similarly.
Same hill.
Same area.
The distortion appears again—but not identically.
Shifted.
Slightly different.
As if whatever was there… did not occupy the exact same configuration between the two exposures.

The man exhales softly through his nose.
“Not a light issue,” he says. “If it was, it’d be consistent.”
He taps the edge of the tray lightly.
“You get fogging, you get streaking… this ain’t that.”

Or when we shot a ball on a rubber band into it (main character has great mechanical repair skills). The ball came back, but …

At first glance—
it’s the same.
Same shape. Same size.

Then—
details resolve.

The surface is no longer uniform.
Not damaged in the way you’d expect—no tearing, no burn marks, no deformation from force.
But—
the texture is… wrong.

Parts of it look slightly out of alignment with themselves.
Not physically broken—
but as if the surface doesn’t fully agree on its own continuity.

When you look directly at it—
it holds.
When your gaze shifts slightly—
something about its edges seems to slip.

Emilia leans slightly—not closer, just adjusting angle.
Her voice is low:
“Do you see that?”

The rubber return line—
still attached—
is intact.
But—
where it meets the ball—
the connection looks…
uncertain.
Not loose.
Not broken.

Just not entirely consistent.

No heat.
No sound.
No movement.

The object is inert.

But it does not look like something that simply passed through space and came back unchanged.

It apparently also had a model for what would have happened if we had acted differently.

For example regarding its effects if we had entered:

You would not “walk into another place” normally.

Instead:

👉 your body would fail to map coherently

Likely outcomes:

Best case (low exposure)

– brief displacement (like the townspeople)
– disorientation
– memory fragmentation
– return within 1–3 hours

Medium exposure

– partial misalignment (like the ball)
– sensory distortion
– possible physical inconsistencies:
– coordination loss
– proprioception failure
– perception of “being in multiple places”

Return possible—but altered.

High exposure / sustained presence

This is the bad one.

– increasing mismatch between body and space
– inability to fully “reassemble” on return

Results could include:

– permanent perceptual instability
– partial non-coherence (you exist, but not fully aligned)
– or failure to return at all

Why people returned “fine”

Because:

– exposure was short
– no deliberate interaction
– no sustained intrusion

You were already exceeding that threshold.

or if we had used dynamite (always be prepared …):

This is the most important “you avoided catastrophe” point.

What dynamite would do

It would not “destroy” the anomaly.

Instead:

👉 it would massively destabilize it

Immediate effects

– rapid expansion of the discontinuity
– loss of localization
– unpredictable spatial bleed

Secondary effects

– the “elephant” would respond much more aggressively
– not as an attack—but as forced correction

Likely outcomes

Best case:

– anomaly expands briefly → collapses violently → area destabilized but resolves

Medium case:

– anomaly spreads beyond the hill
– more people affected
– multiple displacement events

Worst case:

– local reality coherence failure (contained but severe)
– persistent unstable zone

or our test by letting a branch «fall» into the anomaly, partly inside, partly outside:

What the branch actually did

It created:

👉 a continuous cross-boundary condition

That’s critical.

– The system could not resolve it passively
– It had to actively correct it

That triggered:

– expansion along the branch
– structural stress
– and forced expulsion

And importantly:

👉 It drew the “elephant” closer than ever before

You were very close to crossing into:

“system actively correcting a major deviation”

That’s where things escalate fast.

or what a correction would have looked like:

If you had pushed further (e.g. bigger object / dynamite / staying close)

Phase 1 — escalation

– The anomaly expands along the disturbance
– Boundaries lose stability
– Depth and distance fail locally

You would experience:

– spatial confusion
– inability to judge distance
– things not staying where they are

Phase 2 — coherence collapse (local)

This is the key moment.

👉 The area stops behaving as a single consistent space

Examples of how it would appear:

– The ground appears in slightly different positions depending on where you look
– Objects (including you) seem to “double” or misalign
– Movement doesn’t correspond cleanly to position

Not hallucination.

👉 conflicting spatial states

Phase 3 — correction

Then:

👉 The system forces resolution

This is the dangerous part.

Not explosion.

Not violence.

Instead:

– Objects get re-mapped abruptly
– Anything crossing the boundary gets:
– expelled
– or forced into one consistent state

What happens to people

Best case:

– thrown clear (like the branch, but more violently)

Worse case:

forced misalignment correction

That can result in:

– disorientation
– temporary loss of coordination
– or in extreme cases:
– permanent spatial/perceptual instability

Worst case:

– failure to re-map cleanly → disappearance

What it would feel like

Not pain.
Not impact.

More like:

“I am not where I was a moment ago—and I cannot tell where I am correctly.”

That’s why dynamite is so dangerous:

👉 You’re not “breaking” something—you’re forcing it to resolve violently

As experimenting with the disturbance only leads to its spreading and stronger and more violent reactions by the «presence» (stabilization/inspection process), the only winning move is not to play.

And that is very, very hard for intrepid investigators.

This scenario wasn’t about:

– solving a mystery
– defeating a threat

It was about:

👉 recognizing the boundary where knowledge becomes danger

Most groups:

– either do too little
– or go too far

You hit the exact point where:

– you understood enough
– but didn’t trigger full escalation

That’s the “perfect” outcome for this kind of scenario.

(either ChatGPT took the scenario from somewhere or it guessed about the «most groups» part) and

This scenario

– no clear enemy
– no clear goal
– no stable ruleset
– intervention increases risk

👉 boundary recognition model

That’s the key difference:

👉 the win condition is not solving
👉 it’s stopping at the right point

And that creates the moment you had:

“I don’t know if we should do anything”

That’s not failure.

That’s:
👉 correct diagnosis of limits

When we informed someone at the Miskatonic university about the event, ChatGPT even generated a character with a control-first/obscurity-based approach toward the anomaly, who:

  • distrusts individuals by default
  • uses systems as filters
  • values predictability over brilliance
  • assumes competence = risk

So exactly the kind of person that clashes with my personality and forced me to … uh, interact well, leading to pretty good role-play.

And sure, ChatGPT might have used another existing scenario that I do not know, or just modified everything on the fly or through drift — just going for coherence. But it still turned out to be a challenging adventure.

Impressive.

BTW, if you are interested in the scenario, ChatGPT created a short scenario description:

Keeper’s Guide: The Hill Anomaly (Adapted Scenario)

Core Premise
A localized spatial discontinuity appears intermittently on a hill near Arkham. It is:

– not a portal in a conventional sense
– not stable
– reactive to interaction

A secondary phenomenon (“the presence”) appears in response to disturbances.

1. The Situation

The Hill

– Slight elevation near cemetery / outskirts
– Sparse vegetation, uneven ground
– Locals avoid it at night (uneasy, not panicked)

Reports

– strange lights at night
– people briefly “losing time”
– disorientation after visiting

Police / newspapers:
👉 dismiss or minimize events (important)

Truth (Keeper Only)

– A spatial overlap / discontinuity intermittently forms
– It connects incompatible spatial frameworks
– It is unstable but self-correcting

A second phenomenon:

👉 a response mechanism appears when instability increases

2. Getting Investigators Involved

Use any of:

A) Academic Hook

– library references
– odd research notes
– obscure case reports

B) Local Accounts

– caretaker / locals mention lights
– dismissed incidents

C) Personal Curiosity

– investigators just explore Arkham

Important:
👉 no urgency initially
👉 it feels like “strange but not urgent”

3. The Anomaly (Mechanics)

Appearance

– shimmering light / distortion
– slight depression or focal point
– unstable boundaries

Behavior

Passive State

– appears intermittently
– does not expand on its own (quickly)
– small visual distortion

Interaction Effects

If something enters:

– returns altered (subtle spatial inconsistency)
may show:
– displacement
– duplication artifacts
– structural inconsistency

Sustained Interaction

If something remains partially inside:

👉 triggers active response

– expansion along intrusion
– instability increases
– forced expulsion

Key Rule

👉 Interaction increases instability

4. The Presence (“Elephant”)

Nature

– not a creature in normal sense
– not hostile
– not human-interested

Behavior

Without disturbance

– may pass through area
– ignores investigators

With disturbance

– approaches anomaly
– pauses / “examines”
– changes movement pattern

High disturbance

– closer proximity
– longer presence
– possible full correction event

Key Rule

👉 responds to deviation, not people

5. Escalation System

Track a hidden variable: Disturbance Level

Level 0 — Background

– faint lights
– minimal activity

Level 1 — Observation

(trigger: presence of investigators)

– clearer manifestation
– stable anomaly

Level 2 — Light Interaction

(trigger: small objects)

– objects altered
– anomaly slightly stronger

Level 3 — Sustained Interaction

(trigger: branch, extended intrusion)

– expansion
– reactive boundary
– presence responds clearly

Level 4 — Major Disturbance

(trigger: large objects, explosives, repeated interference)

– anomaly destabilizes
– space becomes inconsistent
– presence moves aggressively

👉 danger threshold

6. What Happens if Players Escalate Too Far

Correction Event

– space becomes inconsistent
– objects misalign
– forced resolution

Effects:

– disorientation
– potential injury
– possible disappearance (extreme)

Do NOT make this random death.

Make it:
👉 deeply unsettling and clearly dangerous

7. Warren

Role

– knowledgeable academic
– aware of anomaly type
– focused on containment

Personality

– controlled
– blunt
– low tolerance for debate
– prioritizes stability over comfort

What he knows

– anomaly is real
– interaction worsens it
– cannot safely “fix” it
– must be contained

What he does

– discourages attention
– monitors indirectly
– intervenes only if necessary

How to play him

DO:

– be precise
– be calm
– be slightly dismissive

DON’T:

– explain everything
– validate emotionally
– invite collaboration immediately

Key Function

👉 he defines the boundary

8. Player Paths

Path A — Careful Investigation (your run)

– observation
– controlled tests
– recognition of escalation
– disengagement

👉 best outcome

Path B — Continued Experimentation

– higher disturbance
– stronger presence
– correction event

👉 dangerous outcome

Path C — Forceful Intervention

– dynamite, etc.

👉 rapid destabilization
👉 major correction event

Path D — Entry into Anomaly

– disorientation
– possible alteration
– possible non-return

9. Ending Conditions

Ideal Ending

– players recognize limits
– disengage
– hand off to Warren

Alternative Endings

– anomaly escalates
– players harmed / altered
– situation worsens

No “victory” in classic sense.

👉 success = correct restraint

10. Tone & Running Advice

Tone

– quiet unease
– increasing wrongness
– not jump-scare horror

Key Technique

👉 Reward thinking, not bravery

Critical Rule

If players:

– act carefully → give clear feedback
– escalate → show consequences

Biggest Mistake to Avoid

Do NOT turn it into:

– a fight
– a puzzle to solve
– a portal adventure

This scenario is about:

👉 recognizing when not to act

11. Optional Enhancements

– photographic evidence anomalies
– subtle time inconsistencies
– animals avoiding area

Final Keeper Summary

This is not a “defeat the threat” scenario.

It is:

👉 a boundary recognition scenario

Players win when they say:

“We stop here.”

And if they don’t:

👉 the world teaches them why