OperatorReady-made rule sets you can copy and change, instead of starting from a blank page.When: When you need a fast-start for a new market type.

Details

What this page is forBrowse pre-built strategy templates ready to clone and customize.

When to use itWhen you need a fast-start for a new market type.

Powered bystrat.maker_tight strat.cross_market_arb

Strategies · Templates

Nobody types a strategy. They answer questions about one.

We built a scripting language so a strategy could be expressed precisely. That was the right call for the engine and the wrong call for the person. Almost nobody who wants to trade a prediction market wants to learn a DSL first, and the ones who do still get it wrong in ways that cost real money.

So the script stops being the input and becomes the output. A template is a strategy that already works, with the hard thinking done. The user opens it, answers a handful of plain questions on screens built for the specific bot being configured, reads the whole thing back in English, watches it run on last week’s data, and only then arms it. The DSL is generated underneath and stays visible for anyone who wants it.

The same strategy, two ways to say it

Left is what we ask a developer to write today. Right is the same four decisions, asked as questions. Both produce the identical bundle.

What the DSL asks for

Precise. Also: eleven ways to be silently wrong.

# maker-tight-election.strat
bind cluster US_ELECT_2028_A
engine strat.maker_tight {
  half_spread_bps  = 10
  quote_size_pusd  = 200
  cooldown_s       = 10
  requote_on       = [mid, depth]
}
guard risk.exposureguard {
  max_pos_market   = 500
  max_pos_cluster  = 5000
}
guard risk.drawdownguard {
  daily_stop_pct   = 8
  scope            = strategy
}
route exec.smartrouter {
  objective        = balanced
  max_latency_ms   = 200
}

→

What the builder asks for

Four screens. Every number has its consequence next to it.

Which markets?4 election markets

How tight do you quote?Balanced

Most on one market500 pUSD

Most across the cluster5,000 pUSD

Stop for the day if down8%

Getting filled matters more than price

Six answers. The other nine fields are derived, and shown read-only.

The user never sees half_spread_bps. They see a price ladder with two handles and a line that reads “you will offer to buy at 41¢ and sell at 43¢.” The 10 is what we write down afterwards.

How authoring works, end to end

Five steps. The user can leave at any point and the draft survives — nothing is executed until step five.

Pick a template

Eight starting points. Each one says in a sentence what it is trying to do, and how many decisions it will ask of you.

you choose: 1 of 8

Pick your markets

Search or take the whole cluster. Live count of how much work that implies. Hard cap at 50.

you choose: which markets

Tune each bot

One screen per bot, each built for what that bot does. A guard gets a sentence and a bar. An execution bot gets three named choices.

you choose: 4–8 answers

Read it back

The whole strategy as one paragraph of English. Every number is a link back to the control that set it.

you choose: confirm or go back

Dry run, then arm

Replayed against last week. You see what it would have done, and every guard that would have fired, before anything is live.

you choose: arm it, or don’t

See it working

The whole flow is built out and clickable — every screen, every control, for all four bots in Maker-Tight Election.

Walk the click-through → Control vocabulary for every bot → The older 5-step form

The eight templates

Each card says what the strategy is trying to do in words a newcomer can hold, then lists exactly what it will ask of them. If a template needs more than eight decisions it is too complicated to ship as a template.

Maker-Tight Election

v1.4.2

Sit on both sides of an election market and earn the spread. You are not betting on who wins — you are being the shop that buys and sells while other people make up their minds.

4 decisionsUS_ELECT_2028_Astrat.maker_tightrisk.exposureguardexec.smartrouterrisk.drawdownguard

What you actually decide

Which election markets to quote on
How tight to quote (three presets, or drag the handles)
The most money you will ever have riding on one market
When to stop for the day

Walk through it, screen by screen → every step of this one is built below

Sum-to-One Crypto

v1.1.0

When a market has several outcomes, their prices should add up to 100%. When they don't, buy the cheap side of the mistake. Purely arithmetic — no opinion needed.

3 decisionsCRYPTO_Q2strat.maker_tightexec.smartrouterrisk.marketqualityguardrisk.exposureguard

What you actually decide

Which multi-outcome crypto markets to watch
How far off 100% it has to be before you act
The most money in play at once

Open in the builder → Its bots’ controls

NBA Mean-Rev

v1.2.1

After a big scoring run the crowd overreacts and the price swings too far. Take the other side and wait for it to settle back.

6 decisionsSPORTS_NBAstrat.fade_meanrevintel.macro_sentimentrisk.exposureguardrisk.drawdownguard

What you actually decide

Which games and which markets
How big a move counts as an overreaction
How long to wait before giving up on the snap-back
Position size, daily stop, and whether to trade in the last 2 minutes

Open in the builder → Its bots’ controls

Fed Rate Carry

v1.0.3

The prediction market's view of the next rate decision drifts away from the futures market's view. Hold the gap and collect as it closes.

5 decisionsFED_RATESintel.macro_sentimentstrat.maker_tightrisk.exposureguard

What you actually decide

Which meeting dates to trade
How wide the gap has to be to be worth it
Which data sources you trust for the futures side
Position size and daily stop

Open in the builder → Its bots’ controls

AI Frontier Scout

v0.9.8

Watch the news for the words that move AI markets, and get in before the price catches up. The most opinionated template here — it acts on headlines.

8 decisionsAI_FRONTIERdisc.headline_scannerintel.macro_sentimentexec.smartrouterrisk.drawdownguard

What you actually decide

The words and sources you want watched
How confident the read has to be before it acts
How fast to get in, and how long to hold
Everything about sizing and stopping

Open in the builder → Its bots’ controls

Exposure Guard Wrapper

v1.3.0

Not a strategy on its own — a lid you put on top of other strategies so a whole cluster can never run away from you at once.

2 decisionsUS_ELECT_2028_Arisk.exposureguardrisk.killswitch

What you actually decide

Which strategies sit under the lid
The cluster-wide ceiling

Open in the builder → Its bots’ controls

Smart Router Arb

v1.1.5

The same contract trades in two places at slightly different prices. Buy in the cheap one, sell in the dear one, in the same breath.

4 decisionsCRYPTO_Q2exec.smartrouterrisk.stalebookguardrisk.exposureguard

What you actually decide

Which venues to route between
The smallest gap worth crossing for
How stale a book can be before you refuse to trade on it
Position size

Open in the builder → Its bots’ controls

Compliance Gate Alpha

v1.0.1

A wrapper that refuses to let anything through without a valid builderCode and a passing geofence check. Boring on purpose.

1 decisionUS_ELECT_2028_Arisk.compliancegaterisk.killswitch

What you actually decide

Which jurisdictions you are cleared for

Open in the builder → Its bots’ controls

Where templates come from, and who can publish one

A template is a frozen, versioned bundle: an engine binding, a set of guards, a default parameter map, and a decision manifest that says which of those parameters the user is allowed to see and change. The manifest is what the builder renders — the builder has no hard-coded knowledge of any strategy.

Templates are published by us, not by users, and go through the same promotion ladder as bots: written down, demo-wired, shadow, runtime, production. A template can only reference bots that are at the same rung or higher. That is why AI Frontier Scout sits at v0.9.8 — disc.headline_scanner is still frozen.

User-authored templates are a later milestone and will need a review queue. Cloning an existing template into a private draft is the substitute for now.

Polytraders Operator UX Mock · INTERNAL · demo-wired ≠ production-live · how this works · words we use · the plan · reason codes · shadow-mode pipeline