Backtest the impact of Bitcoin with Support Level, from 2022 to now.
11/4/2025 04:26am
Below is your event-based back-test of “Bitcoin ⸺ Support-Level Bounce” (Low price pierces below daily S1 support and closes back above it) covering 1 Jan 2022 – 3 Nov 2025.
```jgy-json-canvas
{
"show_type": "jgyNewLowcode",
"config": {
"type": "iwencai/stockBackTesting",
"version": "0.0.2",
"url": "https://cdn.ainvest.com/frontResources/s/aime-frontend-foiegras-outer/stockBackTesting/0.0.2/stockBackTesting@0.0.2index.js"
},
"data": {
"datas": [
{
"Backtest Object": "BTCUSD.UDC",
"Backtest Event": "Daily low < S1 & close ≥ S1 (first touch)",
"Backtest Period": "2022.01.01 – 2025.11.03",
"backtest_result": "https://cdn.ainvest.com/backtest/agent/session/784562/64fdf2c5-67db-4587-8bae-f1a424177b2e/b1ace945-97b9-43c4-b3f2-77e484a0f3e2/btc_support_backtest_vis.json"
}
]
}
}
```
Key take-aways:
• Signal definition (auto-assumed):
– An event is recorded when Bitcoin’s intraday low falls below the daily Pivot S1 level but finishes the day at or above S1, and the prior day’s low stayed above its S1 (first touch).
– Rationale: this pattern often signals immediate demand absorption at a key support, a classical “support bounce”.
• Back-test scope (auto-filled): 1 Jan 2022 to 3 Nov 2025, daily closes, 30-day forward holding window, no leverage or stop-loss/take-profit constraints.
• Events detected: 367.
• Performance summary (vs. BTC buy-and-hold of the same periods):
– Average 1-day excess return: +0.21 ppt, win-rate 51 % (statistically insignificant).
– Cumulative excess after 20 trading days: +0.05 ppt; after 30 days: –0.22 ppt (not significant).
– Best relative edge appeared around day 20 (excess ≈ +0.05 ppt) but remains within noise.
Interpretation:
1. The simple “support bounce” rule offers only a marginal—and statistically insignificant—edge over passive holding. The slight positive drift (<0.3 % per day) quickly converges to market performance.
2. With 367 signals (~1 every 3 trading days) the strategy is very active; after fees/slippage it would likely underperform.
3. No distinct post-event drift pattern emerges; both win-rate and average excess return stay near 50 % and 0 % respectively, indicating randomness.
Recommendations:
• Combine support-bounce with trend filters (e.g., 200-day MA direction) to avoid taking signals in strong down-trends.
• Consider stronger support definitions (e.g., S2 touches or multi-day confirmation) to reduce noise.
• Incorporate risk controls (e.g., stop-loss 5-8 %, max hold 10 days) and re-test.
• Evaluate alternative intraday horizons; support reactions may resolve within hours rather than days for crypto.
Feel free to explore refinements or ask for additional tests.
