Random Plot Generators: Do They Actually Help Crime Fiction Writers?

We have a confession to make. It is 2am, we should be asleep, and instead we are feeding half-formed detective duo ideas into an online random plot generator and cackling at the results. “A disgraced maritime lawyer teams up with a psychic ferret handler to solve a smuggling ring in Milton Keynes.” It is nonsense, it is glorious, and it has absolutely no chance of surviving contact with a real draft. So we started asking ourselves a more serious question: do these tools actually help crime fiction writers, or are they just a very entertaining way to avoid writing? After months of tinkering with them for our own story prompts, we have opinions.

What Random Plot Generators Actually Do

Strip away the slick interfaces and most of these tools are doing something remarkably simple: they are trope remixers. They hold a database of character types, crimes, settings, and complications, then shuffle them together using rules or probability weightings. The output looks like a logline, but it is really a collage of familiar pieces arranged in a slightly new order.

The algorithm behind the madness

Tools like RanGen and Plot Generator typically work from hand-built lists. You will find entries like “cynical detective,” “by-the-book rookie,” “cold case,” “witness protection,” and “corrupt superior officer,” combined via templates such as “[Character A] must team up with [Character B] to [goal] before [complication].” ChatGPT-style engines work differently, predicting plausible-sounding text from patterns learned across millions of words. The first approach is transparent and easy to game; the second is more fluent but far less predictable.

From Mad Libs to machine learning

The lineage here is basically Mad Libs with a crime obsession. Early generators were fill-in-the-blank party games. Modern ones dress the same idea in natural language, which makes the results feel more like real pitches than they deserve to. A machine learning model can produce a sentence that reads like a commissioning brief, but reading like a brief is not the same as being a story.

What they can and cannot generate

In our experience, generators are genuinely good at a few things and hopeless at others:

• Good at: unexpected juxtapositions, instant logline shapes, breaking you out of your default genre habits
• Good at: volume — thirty bad ideas in two minutes costs you nothing
• Bad at: UK policing procedure, regional dialect, and anything resembling research
• Bad at: character interiority — motives arrive pre-labelled and paper-thin
• Bad at: knowing the difference between a premise and a plot

The Case for Using Them: Beating the Blank Page

Here is the thing: a blank page is the real enemy, not a mediocre logline. Writers on shows like Line of Duty and Broadchurch have spoken about using improvisation and structured writing exercises to break story blocks — throwing characters into rooms together and seeing what falls out. A random plot generator is a cheaper, sillier version of the same technique. It gives you something to react against.

Killing the inner critic

Nothing silences the inner critic faster than an idea so daft you cannot possibly take it seriously. When the generator hands you “a retired traffic warden and a disgraced sommelier investigate a poisoning at a village fete,” your brain stops guarding the gate and starts playing. That loosening is the actual value. The idea itself is disposable.

Generating detective duos you would never invent

We all default to the same partnerships. The grizzled veteran and the eager graduate. The cynic and the idealist. Feeding those assumptions into a generator and letting it scramble them can surface combinations you would never reach for on your own — a forensic accountant paired with a former army medic, say, or two characters who actively dislike each other for reasons that have nothing to do with the case.

Using random loglines as first-draft scaffolding

Treat the output as scaffolding, not architecture. A generated logline gives you a starting shape: who wants what, who is in the way, what is at stake. You then rebuild every one of those elements with your own knowledge. By the time you have finished, almost nothing of the original remains — and that is exactly the point.

Where They Fall Short: Clichés, Cops, and Cringe

Now the criticism, because it is deserved. Most generators are recycling the same mismatched-partner dynamic that has been running since Lethal Weapon and was lovingly sent up in Hot Fuzz. That trope is not broken, but it is thoroughly worn, and a generator cannot tell the difference between a classic and a cliché.

The mismatched-partner problem

Ask for a detective duo and you will get opposites: messy versus tidy, reckless versus cautious, rule-breaker versus bureaucrat. It is the default setting of the entire buddy cop genre, and it arrives without any of the specificity that makes it work on screen. What made Hot Fuzz sing was not the mismatch itself but the precise English village weirdness surrounding it. The generator gives you the mismatch and none of the village.

Americanisms and procedural errors

This is where UK writers will wince. Generated prompts routinely hand you “homicide detectives,” “the DA’s office,” “Miranda rights,” and guns drawn on routine calls. British policing does not work like that. We have custody sergeants, PACE codes, the Crown Prosecution Service, and a startling shortage of shootouts in provincial towns. A generator trained largely on American film and television will keep serving you American procedure dressed in British place names.

Why a logline is not a story

A logline is a promise. A story is the delivery. Between them sit theme, character arc, structure, subplot, and about ninety thousand words of decisions the generator cannot make for you. Confusing the two is the single most common mistake we see writers make with these tools.

How We Tested Them for Our Own Story Prompts

We ran an informal experiment: the same seed idea fed into three generators, with the outputs scored side by side. It was not science, but it was revealing.

Our scoring criteria

1. Originality — had we seen this exact premise before?
2. UK plausibility — could this actually happen under British policing and law?
3. Duo chemistry — did the partnership feel like two people rather than two labels?
4. Story potential — could we see a full novel or series here, or just a scene?

The best and worst outputs

The best result involved a rural crimes officer and a Manchester-based fraud investigator circling the same case from opposite ends — genuinely usable, with a natural tension between urban and rural policing. The worst was a transatlantic mess featuring a “homicide detective” with a handgun, a “perky” sidekick, and a climax in a warehouse. We have read that story a hundred times. We did not need a machine to write it again.

What we actually kept

Almost nothing survived verbatim. What we kept were fragments: a setting we had not considered, a professional pairing that created interesting friction, and one throwaway line about a Scottish Highlands cold case that we have since built into something far stranger and more specific.

A Better Workflow: Generators as One Tool Among Many

Our recommendation is simple: use generators for what they are good at and do the real work yourself.

Pairing generators with story prompts

Take a generated logline, then apply a constraint of your own. Force the story into a specific UK location — a Manchester tower block, a fading Highland town, a Norfolk caravan park. Constraints are where originality actually lives. The generator supplies the spark; the constraint supplies the story.

Editing the output like a script doctor

Read every generated line and ask three questions: is this plausible here, is this specific, and does this character want something I can name? If the answer to any of them is no, rewrite it. You are not accepting a pitch, you are interrogating one.

When to walk away and write

There is a point — usually about twenty minutes in — where you are no longer generating ideas, you are just pressing a button and giggling. That is the moment to close the tab and open your draft. The generator has done its job. Everything after that is craft.

FAQ

Are random plot generators worth using for serious crime fiction?

Yes, but only as warm-up tools. They are excellent for breaking a blank page and producing unexpected combinations, and poor at procedure, character depth, and anything specific to UK settings.

Why do generated detective duos all feel the same?

Because the mismatched-partner dynamic is the dominant trope of the buddy cop genre, from Lethal Weapon through to Hot Fuzz. Generators default to the most statistically common pattern rather than the most interesting one.

Can I use a generated logline as the basis for a novel?

You can, but plan to rebuild it entirely. Treat the logline as scaffolding: keep the shape if it works, then replace every character, motive, and detail with researched, specific choices of your own.

Which is better, a rule-based generator or an AI chatbot?

Rule-based tools like RanGen and Plot Generator are more transparent and easier to steer. Chatbot-style engines produce more fluent prose but drift toward Americanisms and procedural errors, so they need heavier editing.

How do I make generated ideas feel authentically British?

Anchor them in real places and real procedure. Set the story in Manchester or the Scottish Highlands, swap “homicide detective” for the correct UK rank, and remove the guns. Local knowledge is the one thing no generator can supply.

Random plot generators are a genuinely fun spark, and we will keep using them at 2am. But a spark is not a fire. The craft, the character work, and the local knowledge that make British crime fiction and buddy cop films sing still have to come from you — and that is very good news for anyone who wants to write something a machine could never have produced on its own.