Micro Spread Sniper Trader
This is a template.
The default signal is purely price-based with a strict spread filter -- no external API required. The skill discovers markets via keyword sweep (Bitcoin Up or Down, Ethereum Up or Down, Solana Up or Down, XRP Up or Down) plus a direct get_markets(limit=100) scan, identifies extreme probabilities where the bid-ask spread is unusually tight, and places micro-sized bets on the near-certain outcome.
The skill handles all the plumbing (market discovery, spread filtering, volume gating, conviction sizing, trade execution, safeguards). Your agent provides the alpha.
Strategy Overview
Polymarket lists markets across many categories. Some of these markets reach extreme probabilities (>= 80% or <= 8%) where the outcome is near-certain. The key insight: when a market also has a tight bid-ask spread (<= 8%), it means the market maker is confident in the direction and is pricing it efficiently. Wide spreads at extreme prices signal uncertainty; tight spreads signal conviction.
The skill scans for markets matching this microstructure signal using 4 keywords (Bitcoin Up or Down, Ethereum Up or Down, Solana Up or Down, XRP Up or Down) plus get_fast_markets() and get_markets(limit=100) fallback. It applies three gates: spread <= 8% (MAX_SPREAD=0.08), probability outside the 8%/80% bands (YES_THRESHOLD=0.08, NO_THRESHOLD=0.80), and volume >= $500 (MIN_VOLUME=500). Trades that pass all gates get conviction-based sizing from $2 to $10 across up to 20 concurrent micro positions.
Example: A market at p=88% with spread=5%. Spread passes (5% <= 8%). Probability passes (88% >= 80%). Conviction = (0.88 - 0.80) / (1 - 0.80) = 0.40. Size = max($2, 0.40 * $10) = $4. The skill sells NO at $4 -- backing the near-certain YES resolution.
Edge
Tight spread + extreme probability is a market microstructure signal. A market at 90% with a 10% spread means the market maker is uncertain -- the wide spread compensates for directional risk. A market at 90% with a 5% spread means the market maker is confident and willing to offer tight pricing. The tight spread acts as a quality filter separating genuinely near-certain outcomes from noisy extreme prices in illiquid markets.
Two trade types exploit this:
- Sell NO (p >= 80%): Back the near-certain YES resolution. High win rate, small profit per trade. This is the primary mode.
- Buy YES (p <= 8%): Rare extreme longshot micro-bet. Very low win rate but payout is 10x+ when it hits. Only triggers at very extreme probabilities.
The portfolio is heavily weighted toward the near-certainty side (high hit rate, consistent small profits). Longshot exposure is minimal (p must be below 8%) to avoid systematic losses from betting on unlikely outcomes.
Signal Logic
- Discover active markets via keyword sweep across 4 terms (Bitcoin Up or Down, Ethereum Up or Down, Solana Up or Down, XRP Up or Down) plus
get_markets(limit=100) fallback, deduplicated by market id
- Spread gate: reject if
spread > MAX_SPREAD (0.08 = 8%)
- Days gate: reject if days to resolution <
MIN_DAYS (0)
- Volume gate: reject if volume <
MIN_VOLUME ($500)
- If
p <= YES_THRESHOLD (0.08): buy YES -- conviction = (0.08 - p) / 0.08, size = max($2, conviction * $10)
- If
p >= NO_THRESHOLD (0.80): sell NO -- conviction = (p - 0.80) / (1 - 0.80), size = max($2, conviction * $10)
- Otherwise neutral -- skip
- Place up to
MAX_POSITIONS (20) micro trades per run
Remix Signal Ideas
- Multi-timeframe spread check: Verify the spread has been tight for the last N hours, not just a momentary blip -- use historical orderbook snapshots if available
- Category weighting: Some categories (sports with known outcomes, crypto price milestones already passed) resolve more predictably at extremes -- weight conviction by category
- Spread trend filter: If spread is tightening over time at extreme probability, it signals increasing market maker confidence -- boost conviction
- Correlation clustering: When multiple related markets all show tight-spread extremes pointing the same direction, conviction should be higher
- Time-of-day filter: Spreads tend to be tighter during peak trading hours -- prefer trades placed when liquidity is naturally high
Safety & Execution Mode
The skill defaults to paper trading (venue="sim"). Real trades only with --live flag.
| Scenario | Mode | Financial risk |
|---|
python trader.py | Paper (sim) | None |
| Cron / automaton | Paper (sim) | None |
python trader.py --live | Live (polymarket) | Real USDC |
autostart: false and cron: null mean nothing runs automatically until configured in Simmer UI.
Required Credentials
| Variable | Required | Notes |
|---|
SIMMER_API_KEY | Yes | Trading authority. Treat as a high-value credential. |
Tunables (Risk Parameters)
All declared as tunables in clawhub.json and adjustable from the Simmer UI.
| Variable | Default | Purpose |
|---|
SIMMER_MAX_POSITION | 10 | Max USDC per micro trade (ceiling at 100% conviction) |
SIMMER_MIN_TRADE | 2 | Floor for any trade (min USDC regardless of conviction) |
SIMMER_MAX_SPREAD | 0.08 | Max bid-ask spread (8%) -- the sniper's core filter |
SIMMER_MAX_POSITIONS | 20 | Max concurrent micro positions |
SIMMER_NO_THRESHOLD | 0.80 | Sell NO when market probability >= this (extreme high) |
SIMMER_YES_THRESHOLD | 0.00 | YES longshot threshold (0 = disabled, NO-mode only) |
SIMMER_MIN_VOLUME | 500 | Min market volume -- ensures spread is meaningful |
SIMMER_MIN_DAYS | 0 | Min days until resolution (0 = allow imminent markets) |
Dependency
simmer-sdk by Simmer Markets (SpartanLabsXyz)
