[Famous-PA] Darvas box breakout
public
@djgambhirBTCUSDT · 1d · 2021-12-28 → 2026-08-31 · 1,708 bars in 1.9ms
Underperformed the market. Holding BTCUSDT returned +65.29% over the same period, against this strategy’s +21.04%.
Only 17 trades. Too few to separate skill from luck. Treat every statistic here as provisional.
Net P&L
+21.04%
$2,104.43
Max drawdown
-8.92%
Max run-up
+35.11%
Trades
17
52.9% win rate
Profit factor
2.49
2021-12-28drawdown · worst -8.92%2026-08-31
TrainingValidationOut-of-sample
| Segment | Period | Return | Max drawdown | Trades | Win rate |
|---|---|---|---|---|---|
| Training | 2022-01-01 → 2024-10-18 | +12.19% | -6.95% | 10 | 60% |
| Validation | 2024-10-18 → 2025-09-24 | +8.77% | -4.65% | 2 | 50% |
| Out-of-sample | 2025-09-24 → 2026-08-31 | -0.80% | -6.39% | 5 | 40% |
Risk-adjusted
- Sharpe
- 0.58
- Sortino
- 0.89
- Calmar
- 0.47
- Return ÷ max drawdown
- 2.36
- CAGR
- +4.17%
- Exposure
- 30.5%
Per trade
- Expectancy
- $123.79 · +4.76%
- Average win / loss
- $390.38 / −$176.12
- Payoff ratio
- 2.22
- Largest win / loss
- $1,128.24 / −$290.25
- Longest losing streak(winning: 6)
- 4 trades
Drawdown over time
- Max drawdown
- -8.92%
- Longest time under water(still going at the end)
- 622 days
- Worst drawdown recovered in
- not recovered
Costs and mechanics
- Fees paid
- −$47.30
- Funding
- −$513.13
- Liquidations
- 0
- Warmup bars(loaded before the start date)
- 19
- Benchmark(buy and hold)
- +65.29%
Settings
- hh20 length
- 20
- ll10 length
- 10
- volsma length
- 20
- atr14 length
- 14
- Long entry: volume > ▢ × volsma
- 1.5
- Stop loss (× atr14)
- 2
- Leverage (×)
- 1
- Position size (% of equity)
- 25
BTCUSDT with every trade
Loading prices…
▲ long opened▼ short opened● closed at a profit● closed at a loss
Trades
17 trades
| # | Side | Entry | Exit | Held | P&L | Return | Closed by |
|---|---|---|---|---|---|---|---|
| 1 | long | 2022-07-20 00:00 @ 23,390.58 | 2022-08-20 00:00 @ 20,820.84 | 31d | −$290.25 | -11.61% | signal |
| 2 | long | 2022-10-27 00:00 @ 20,762.55 | 2022-11-08 00:00 @ 19,401.35 | 12d | −$164.89 | -6.79% | stop loss |
| 3 | long | 2022-12-14 00:00 @ 17,780.66 | 2022-12-16 00:00 @ 16,699.00 | 2d | −$148.97 | -6.24% | stop loss |
| 4 | long | 2023-01-12 00:00 @ 17,939.19 | 2023-02-10 00:00 @ 21,788.94 | 29d | $474.80 | +20.21% | signal |
| 5 | long | 2023-02-16 00:00 @ 24,337.57 | 2023-03-03 00:00 @ 22,537.73 | 15d | −$193.46 | -7.84% | stop loss |
| 6 | long | 2023-03-15 00:00 @ 24,687.44 | 2023-04-22 00:00 @ 27,247.15 | 38d | $230.96 | +9.55% | signal |
| 7 | long | 2023-06-21 00:00 @ 28,296.36 | 2023-07-25 00:00 @ 29,157.97 | 34d | $54.27 | +2.19% | signal |
| 8 | long | 2023-10-24 00:00 @ 33,080.61 | 2024-01-19 00:00 @ 41,314.94 | 87d | $503.66 | +20.22% | signal |
| 9 | long | 2024-02-10 00:00 @ 47,125.52 | 2024-03-20 00:00 @ 61,961.71 | 39d | $700.35 | +26.77% | signal |
| 10 | long | 2024-10-16 00:00 @ 67,057.41 | 2025-02-03 00:00 @ 97,644.97 | 110d | $1,128.24 | +40.42% | signal |
| 11 | long | 2025-05-09 00:00 @ 103,248.65 | 2025-05-31 00:00 @ 103,929.11 | 22d | $4.04 | +0.13% | signal |
| 12 | long | 2025-07-11 00:00 @ 116,002.40 | 2025-08-02 00:00 @ 113,222.15 | 22d | −$96.44 | -3.14% | signal |
| 13 | long | 2025-10-02 00:00 @ 118,575.71 | 2025-10-10 00:00 @ 113,395.41 | 8d | −$142.34 | -4.67% | stop loss |
| 14 | long | 2026-01-06 00:00 @ 93,841.06 | 2026-01-20 00:00 @ 88,686.65 | 14d | −$174.32 | -5.78% | stop loss |
| 15 | long | 2026-03-05 00:00 @ 72,655.93 | 2026-03-23 00:00 @ 67,817.13 | 18d | −$198.30 | -6.67% | signal |
| 16 | long | 2026-04-18 00:00 @ 77,046.01 | 2026-05-17 00:00 @ 78,088.68 | 29d | $40.94 | +1.40% | signal |
| 17 | long | 2026-08-20 00:00 @ 69,323.96 | 2026-08-31 00:00 @ 78,533.89 | 11d | $376.13 | +12.83% | end of data |
How this was run
Orders fill at the next bar’s open, never at the close that produced the signal. When one bar’s range covered both the stop and the target, the stop was taken, because the data cannot say which came first. Fees, slippage and funding are all charged.
Warmup history was loaded from before 2022-01-01 so the requested window is fully tradable.