The size anomaly detector currently emits a base 0.2 confidence signal
whenever a market is "niche" (low volume OR niche-prone category with
unknown volume) — even for $5 trades. In production this floods the
alert pipeline with nothing-burgers, because the niche-prone category
set covers `science / tech / finance / other`, which matches a huge
chunk of Polymarket's long tail.
This adds a `DEFAULT_NICHE_MIN_TRADE_SIZE = $500` floor that applies
ONLY to the niche-only path: if a trade exceeds the volume or book
thresholds, the guard does not block it. So real anomalies still come
through, but tiny niche trades get filtered out before reaching the
risk scorer.
The threshold is configurable via `niche_min_trade_size` in
`SizeAnomalyDetector.__init__`.
Tests added:
- niche-only below floor → suppressed
- niche-only at/above floor → still emits 0.2 base
- small trade with high volume_impact → guard does not block
Co-authored-by: schrodinger01 <schrodinger01@users.noreply.github.com>
- Use contextlib.suppress instead of try/except/pass (SIM105)
- Prefix unused fixture arguments with underscore (ARG002)
- Replace asyncio.TimeoutError with TimeoutError (UP041)
- Apply ruff formatting to all files
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add sniper cluster detection system that identifies wallets exhibiting
coordinated "sniper" behavior - consistently entering markets within
minutes of their creation.
Key features:
- SniperDetector class using DBSCAN clustering algorithm
- Tracks wallet entries across markets with timing analysis
- Feature vector: market hash, entry delta, log position size
- Identifies clusters of wallets with similar timing patterns
- Generates SniperClusterSignal for detected cluster members
- Configurable entry threshold (default 5 minutes), cluster size, DBSCAN params
- Confidence scoring based on cluster size, entry speed, market overlap
Also adds:
- SniperClusterSignal model to detector/models.py
- scikit-learn and numpy dependencies for ML clustering
- 27 comprehensive tests covering clustering logic and edge cases
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Implement RiskScorer that combines signals from multiple detectors into
a unified risk assessment with weighted scoring and deduplication.
Features:
- SignalBundle for collecting signals for a single trade
- RiskAssessment dataclass with complete scoring metadata
- Configurable weights for each signal type
- Multi-signal bonus (1.2x for 2 signals, 1.3x for 3+)
- Redis-based deduplication (1 hour window by default)
- Alert threshold configuration (default: 0.6)
- Batch assessment for processing multiple trades
- A/B testing support via dynamic weight updates
Default weights:
- fresh_wallet: 0.40
- size_anomaly: 0.35
- niche_market: 0.25
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Implement SizeAnomalyDetector for identifying trades with unusually
large position sizes relative to market liquidity.
Features:
- Volume impact analysis (trade size / 24h volume)
- Order book impact analysis (trade size / book depth)
- Niche market detection using category heuristics
- Confidence scoring with configurable thresholds
- Batch analysis for processing multiple trades
The detector gracefully handles missing volume/book data by falling
back to category-based heuristics for identifying niche markets
where large trades are more significant.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add FreshWalletDetector class that identifies trades from fresh wallets
and generates alert signals with confidence scores.
Features:
- FreshWalletSignal dataclass with trade/wallet data and confidence
- Configurable thresholds for nonce, age, and minimum trade size
- Confidence scoring based on wallet freshness and trade characteristics
- Batch analysis support with parallel processing
- Integration with WalletAnalyzer from profiler module
Confidence scoring:
- Base: 0.5 (fresh wallet detected)
- +0.2 if nonce == 0 (brand new)
- +0.1 if age < 2 hours (very young)
- +0.1 if trade size > $10,000 (large trade)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add pyproject.toml with core and dev dependencies
- Configure Ruff for linting and formatting
- Configure MyPy for strict type checking
- Configure pytest with asyncio support
- Create src/polymarket_insider_tracker package structure
- Add py.typed marker for PEP 561 compliance
- Add pre-commit hooks configuration
- Add comprehensive .gitignore
- Create test directory structure with basic tests
Closes#25🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
