Wallet Intelligence

Score wallets, detect exploitable patterns, identify bots, and auto-fade bad traders.

Pro Feature. Requires a Pro or Ultra subscription. Get started at api.mathematicalcompany.com

Wallet Intelligence

Horizon’s wallet intelligence module provides three capabilities built on Polymarket’s public Data API:

Wallet Scoring

Evaluate any wallet’s trading performance and compute a composite score from -1 (terrible) to +1 (excellent).

Adversarial Analysis

Detect exploitable patterns in a wallet’s trading: timing regularity, sizing predictability, directional bias, and more.

Bot Detection

Scan a market to classify wallets as bots vs humans with strategy type identification and confidence scores.

---

Wallet Scoring

Score any wallet’s historical trading performance.

import horizon as hz

score = hz.score_wallet("0x1234abcd...")
print(score.composite_score)  # -1 to +1
print(score.win_rate)         # 0.0 to 1.0
print(score.sharpe)           # Sharpe estimate
print(score.total_pnl)       # Cumulative USDC PnL

Parameter	Type	Default	Description
address	str	required	Wallet address (0x…)
limit	int	500	Max positions to analyze

Score Interpretation

Composite Score	Interpretation
> 0.5	Excellent trader
0.2 to 0.5	Good trader
-0.2 to 0.2	Average
-0.5 to -0.2	Poor trader
< -0.5	Terrible trader (fade candidate)

WalletScore Fields

Field	Type	Description
wallet	str	Wallet address
win_rate	float	Fraction of profitable positions
avg_pnl_pct	float	Average PnL % per position
sharpe	float	Sharpe estimate (PnL% / stddev)
total_pnl	float	Cumulative USDC PnL
trade_count	int	Number of trades
position_count	int	Number of positions analyzed
composite_score	float	-1 (terrible) to +1 (excellent)

---

Adversarial Analysis

Detect exploitable patterns in a wallet’s trading behavior.

import horizon as hz

report = hz.analyze_wallet("0x1234abcd...")
print(f"Vulnerability: {report.vulnerability_score:.2f}")

for pattern in report.patterns:
    status = "EXPLOITABLE" if pattern.exploitable else "safe"
    print(f"  {pattern.name}: {pattern.value:.2f} [{status}]")
    if pattern.exploitable:
        print(f"    -> {pattern.counter_strategy}")

Parameter	Type	Default	Description
address	str	required	Wallet address
trade_limit	int	500	Max trades to analyze

Detected Patterns

Pattern	Metric	Exploitable When	Counter-Strategy
timing_regularity	CV of inter-trade intervals	CV < 0.3	Front-run before expected trade times
sizing_predictability	CV of trade sizes	CV < 0.3	Size counter-trades to match
directional_bias	Buy ratio (buys/total)	|ratio - 0.5| > 0.25	Fade the dominant direction
reaction_speed	Median reaction time (seconds)	< 60s	Pre-position before triggers
concentration	Herfindahl index across markets	> 0.3	Target concentrated markets
stop_levels	Density of exit price clusters	> 30%	Hunt stops near cluster levels

---

Bot Detection

Scan a market to identify which wallets are bots and classify their strategy type.

import horizon as hz

bots = hz.scan_bots("0xcondition_id_here")
for bot in bots:
    if bot.is_bot:
        print(f"{bot.wallet[:10]}... {bot.strategy_type} "
              f"(confidence={bot.confidence:.0%})")
        for e in bot.evidence:
            print(f" - {e}")

Parameter	Type	Default	Description
condition_id	str	required	Market condition ID
trade_limit	int	1000	Max trades to analyze
min_trades	int	5	Min trades per wallet

Strategy Types

Type	Description	Key Signals
market_maker	Provides two-sided liquidity	Balanced buy/sell, regular timing, consistent sizing
grid_bot	Places orders at regular intervals	Very regular sizing, balanced, moderate frequency
momentum	Follows price direction	Strong directional bias, high frequency
sniper	Fast, large, few trades	Sub-second intervals, large sizes, low count
human	Manual trader	Irregular timing, varied sizes, low frequency

BotDetection Fields

Field	Type	Description
wallet	str	Wallet address
is_bot	bool	True if classified as bot (confidence > 0.6)
confidence	float	0.0 to 1.0
strategy_type	str	Classification (see table above)
evidence	list[str]	Human-readable reasons
trade_count	int	Number of trades analyzed
avg_interval_seconds	float	Mean time between trades
size_regularity	float	0 (random) to 1 (identical sizes)

---

What’s Next?

Fade a Wallet

To fade a wallet, use copy_trades(inverse=True) on the Copy-Trading page. reverse_copy() and reverse_copier() are convenience wrappers around this.

Deep Profiling

Wallet Profiler builds on score_wallet() with 4-dimensional behavioral analysis (temporal, market selection, order flow, edge) and automatic strategy generation.

Wallet intelligence uses Polymarket’s public Data API. Results are based on historical behavior and may not predict future performance. Use responsibly and ethically - these tools are designed for competitive market analysis, not harassment or manipulation.