📊 BTCUSDT April 2024 — Polymarket Δprob Backtest

Backtest that uses Polymarket "Bitcoin all-time high" probability shifts (Δprob) as a leading indicator for BTC direction, with a 1% trailing take-profit and 1% hard stop-loss.

The strategy reads Polymarket CLOB daily snapshots (Δprob — change of "yes" probability for crypto-prices markets) from assets/polymarket-backtest-result.json. When |Δprob| ≥ 0.10 and the snapshot is no older than 1h, it opens a position: long on positive Δprob, short on negative. The position runs with a 1% hard stop-loss and 1% trailing take-profit; if the trailing take never arms, the position closes after 24h.

Strategy: apr_2026_strategy | Exchange: ccxt-exchange | Frame: apr_2024_frame

🚀 Quick Start

Link to the source code

npm start -- --backtest --symbol BTCUSDT ./content/apr_2024.strategy/apr_2024.strategy.ts
Copy

🔍 Methodology

The signal source is Polymarket prediction-market probability for BTC-related markets (e.g. "Bitcoin all time high in August?"). Each day's "yes" price shift is a market-implied probability change for a BTC-driven event — when it moves sharply, retail flow into Polymarket precedes BTC spot movement by roughly one day.

To remove look-ahead bias, the strategy uses only the timestamp and Δprob from the source JSON — it does not read entryPrice or exitPrice (which were computed in the original polymarket-backtest.ts using future klines). The decision rule is:

1. For the current candle time when, find the most recent Polymarket signal with timestamp ≤ when.
2. Reject it if older than MAX_SIGNAL_AGE_MS = 1h or |dprob| < MIN_ABS_DPROB = 0.10.
3. Open long/short at market with Position.moonbag (TP off, hard SL = 1%).
4. Close on 1% trailing drawdown from peak profit, or on 24h timeout.

📉 Price Context (April 2024)

Metric	Value
Period	Apr 1 – Apr 28, 2024
Ticker	BTCUSDT
Price range	~$60,600 – ~$71,600
Dominant move	Topping pattern after April highs, ~13% drawdown into mid-month

April 2024 captured BTC's local top near $71,600 (Apr 8) followed by a sharp correction to $60,700 (Apr 30) — a regime where Polymarket sentiment swings were unusually pronounced.

✨ Performance Summary

Backtest result

Metric	Value
Total trades	10
Net PNL	+0.63%
Win rate	70% (7 / 10)
Avg win	+0.69%
Avg loss	−1.40%
Best trade	+1.11% (SHORT Apr 17)
Worst trade	−1.40% (SHORT Apr 6, SL)
Closed by trailing take	7
Closed by hard stop-loss	3
Direction split	8 SHORT / 2 LONG

Risk-Adjusted Metrics

Individual trade PNL values: −1.40, −1.40, −1.40, +0.62, +0.35, +0.89, +0.87, +1.11, +0.57, +0.42

Metric	Value
Mean trade PNL	+0.063%
Std dev per trade	0.981%
Sharpe Ratio (per-trade)	+0.065
Max drawdown (single trade)	−1.40%
Profit factor	1.15
Expectancy per trade	+$0.06

Sharpe Ratio = mean(PNL) / std(PNL), per-trade, no risk-free adjustment (10-trade sample over 28 days makes annualisation meaningless).

Profit factor = gross profit / gross loss = 4.83% won / 4.20% lost = 1.15.

The 70% win rate is encouraging but per-trade Sharpe of 0.065 is statistically indistinguishable from zero — the three SL hits (−1.40% each) almost exactly cancel the seven trailing-take wins. The strategy survives April 2024 but does not produce a robust edge at the current MIN_ABS_DPROB = 0.10 threshold; higher thresholds (≥0.15) or volatility filters are likely needed to suppress the LONG signals during the topping phase (trades 1 and 3 were both LONGs that hit SL on the way down).

📋 Trade Log

#	Opened (UTC)	Closed (UTC)	Held	Dir	Δprob	Open	Close	Peak	DD	PNL%	Exit
1	Apr 5 00:01	Apr 5 02:36	2.6h	LONG	+0.130	$68,465.24	$67,780.59	−0.12%	−1.37%	−1.40%	stop-loss
2	Apr 6 00:01	Apr 6 22:12	22.2h	SHORT	−0.180	$67,811.75	$68,489.87	+0.09%	−1.40%	−1.40%	stop-loss
3	Apr 9 00:01	Apr 9 06:40	6.7h	LONG	+0.140	$71,627.28	$70,911.00	−0.27%	−1.35%	−1.40%	stop-loss
4	Apr 10 00:01	Apr 10 14:21	14.3h	SHORT	−0.110	$69,144.88	$68,438.99	+1.77%	−0.90%	+0.62%	trailing take
5	Apr 13 00:01	Apr 13 03:04	3.0h	SHORT	−0.345	$67,125.52	$66,624.34	+1.47%	−0.40%	+0.35%	trailing take
6	Apr 14 00:01	Apr 14 03:37	3.6h	SHORT	−0.170	$63,910.38	$63,088.37	+2.19%	−0.89%	+0.89%	trailing take
7	Apr 16 00:01	Apr 16 05:19	5.3h	SHORT	−0.135	$63,406.47	$62,605.77	+1.95%	−1.03%	+0.87%	trailing take
8	Apr 17 00:01	Apr 17 13:44	13.7h	SHORT	−0.140	$63,789.80	$62,829.23	+2.15%	−1.39%	+1.11%	trailing take
9	Apr 25 00:01	Apr 25 14:11	14.2h	SHORT	−0.125	$64,275.82	$63,651.86	+1.71%	−1.08%	+0.57%	trailing take
10	Apr 27 00:01	Apr 27 02:35	2.6h	SHORT	−0.170	$63,775.30	$63,254.50	+1.48%	−0.56%	+0.42%	trailing take

📈 Equity Curve

After trade	Cumulative PNL%
1 — Apr 5	−1.40%
2 — Apr 6	−2.80%
3 — Apr 9	−4.20% ← trough
4 — Apr 10	−3.57%
5 — Apr 13	−3.22%
6 — Apr 14	−2.33%
7 — Apr 16	−1.47%
8 — Apr 17	−0.36%
9 — Apr 25	+0.21%
10 — Apr 27	+0.63%

The first three trades all hit the 1% SL, producing a −4.20% drawdown by Apr 9. From Apr 10 onward only SHORT signals fired, and seven consecutive trailing-take exits recovered the deficit by Apr 27.

🚀 How to Run

npm start -- --backtest --symbol BTCUSDT \
  --strategy apr_2026_strategy \
  --exchange ccxt-exchange \
  --frame apr_2024_frame \
  ./content/apr_2024.strategy/apr_2024.strategy.ts
Copy

Add --ui to open the web dashboard at http://localhost:60050:

npm start -- --backtest --symbol BTCUSDT --ui \
  ./content/apr_2024.strategy/apr_2024.strategy.ts
Copy

🌍 Environment Variables

Create a .env file in the project root (copy from .env.example):

# Telegram notifications (optional)
CC_TELEGRAM_TOKEN=your_bot_token_here
CC_TELEGRAM_CHANNEL=-100123456789

# Web UI server (optional, defaults shown)
CC_WWWROOT_HOST=0.0.0.0
CC_WWWROOT_PORT=60050
Copy