Kakashi postmortem: the strategy worked, the exchange didn't

A Polymarket copy-trading bot deployed to EC2. It tracks top wallets and mirrors high-conviction positions through a risk-gate stack.

Copy-trade flow

Top-wallet tracker: Polymarket leaderboard and position ingest.
Conviction filter: signal only when a bet is at least 10% of that wallet’s 7-day volume.
Risk gate stack: confidence thresholds, exposure caps.
CLOB execution: order placement against Polymarket’s book.

Postmortem


Validated	Paper trading confirmed the copy-strategy edge end to end on EC2.
Blocked	An exchange-side CLOB order version mismatch rejected live orders. A bug on their end, not the strategy’s.
Status	Infrastructure intact and redeployable. Live trading waits on the exchange fix.

From the original write-up (May 2026)

A multi-signal prediction market system with wallet intelligence and risk-gated CLOB execution. The core mission: identify mispriced prediction markets and execute with confidence-weighted position sizing and hard risk controls.

The signal pipeline

Market inputs: ingests Kraken OHLCV, Polymarket order books, top-wallet on-chain activity, and macro/event context in real time via WebSocket.
Feature layer: engineers signals: Bollinger Bands, RSI, momentum, token normalization, wallet conviction ranking, volatility, and edge calculations.
Decision layer: routes signals across concurrent strategies (directional, market-making, wallet-following, arbitrage, event forecasting).
Risk gate stack: validates confidence threshold, edge minimum, exposure cap, and position guards before any order is approved.
Executor (CLOB): creates, routes, and posts orders with stateful retries and fill handling.
Operator feedback: Discord alerts, process monitor, trade logs, win/loss dashboards, and analytics state via PostgreSQL.

Risk gate stack

Every order passes through a multi-layer gate before execution. If any layer fails, the order is blocked and logged. The system never silently skips a check.

Layer	Component	Function
Layer 1	Confidence threshold	Blocks any signal below 60% model confidence.
Layer 2	Edge minimum	Requires at least 3% edge over market implied probability.
Layer 3	Exposure cap	Hard cap on total portfolio exposure. No exceptions.
Layer 4	Deduplication guard	Prevents double-entry on the same market across strategies.
Layer 5	Inventory guard	Tracks open positions and blocks conflicting directional exposure.
Layer 6	Timeout exit	Forces position closure if hold time exceeds configured maximum.

Logic highlight: confidence-weighted sizing

Position size scales with signal confidence and edge. High-conviction signals get larger allocation; weak signals get minimum size or are blocked entirely.

def calculate_position_size(confidence: float, edge: float, portfolio: dict) -> float:
    """
    Scales position size based on signal quality.
    Never exceeds portfolio exposure cap regardless of confidence.
    """
    base_size = portfolio["base_position_size"]
    max_size = portfolio["max_position_size"]

    quality_multiplier = (confidence - 0.60) * edge * 10
    raw_size = base_size * (1 + quality_multiplier)

    return min(raw_size, max_size)

Tech stack

Execution: Polymarket CLOB API, Kraken WebSocket API.
Orchestration: Python asyncio, multi-strategy concurrent runners.
State management: Redis (real-time order book, deduplication), PostgreSQL (trade logs, analytics).
Risk controls: custom risk gate engine with 6 validation layers.
Infrastructure: Docker, environment-based config, unless-stopped restart policy.
Observability: Discord webhooks (live alerts), structured logging, win/loss dashboards.