Negative after costs

ETHUSDT · 60m · Verdict Card

RSI overbought/oversold fade · rsi_fade · hold 24 · 2025-06-01 → 2026-10-11 · 11929 bars · bybit · 2026-10-12 12:48

Rule: Long when RSI(14) ≤ 30, short when RSI(14) ≥ 70
Cost model: 5.5bp/side fee + 3bp/side slippage + 0.06bp/bar funding

Trades

169

Win rate

50.9%

Avg / trade

-21bp

Gross avg

-3bp

Total

-3550bp

t-stat

-0.81

Permutation p

0.5527 · T0.553/S0.541

95% CI of mean

[-72bp, 28bp]

Max drawdown

48.0%

Exposure

40.9%

Profit factor

0.84

Worst trade

-1694bp

Median

12bp

Equity curve (log)

Buy & hold: -0.9% · missing 0 (max gap 0) · OHLC complete

Stability by segment

Seg	From → To	Trades	Avg / trade	Win rate	t-stat
Seg 1	2025-08-25 → 2025-11-27	43	16bp	49%	0.31
Seg 2	2025-12-01 → 2026-03-08	43	-81bp	44%	-1.55
Seg 3	2026-03-13 → 2026-06-25	43	4bp	60%	0.08
Seg 4	2026-06-30 → 2026-10-08	40	-23bp	50%	-0.40

In-sample/Out-of-sample	Trades	Avg / trade	Win rate	t-stat
In-sample 60%	101	-23bp	50%	-0.69
Out-of-sample	68	-18bp	51%	-0.44

Trade log · last 12 trades

Time	Side	Entry px	Exit px	Bars held	Exit	Net
2026-10-08 19:00	LONG	2536.04	2504.97	24	time	-141.7bp
2026-10-07 09:00	LONG	2655.82	2576.8	24	time	-320.5bp
2026-10-05 06:00	SHORT	2729.75	2714.9	24	time	36.1bp
2026-09-23 22:00	LONG	2674.85	2656.53	24	time	-87.2bp
2026-09-21 17:00	SHORT	2719.94	2750.01	24	time	-128.4bp
2026-09-20 10:00	LONG	2572.72	2660.84	24	time	318.3bp
2026-09-19 00:00	SHORT	2596.54	2649.2	24	time	-219.2bp
2026-09-17 23:00	SHORT	2469.21	2583.75	24	time	-471.9bp
2026-09-15 04:00	SHORT	2569.86	2406.53	24	time	638.2bp
2026-09-13 17:00	LONG	2484.04	2523.03	24	time	137.3bp
2026-09-11 21:00	SHORT	2629.27	2543.6	24	time	312.8bp
2026-09-10 20:00	LONG	2413.08	2510.39	24	time	376.9bp

Method & known limits

Exit: fixed 24-bar hold; fill=signal bar close
Costs: 5.5bp/side fee + 3bp/side slippage + 0.06bp/bar funding drag (same both sides; funding sign not modeled)
Non-overlapping: signals during open positions are dropped
Significance: t-stat on per-trade net returns. Two nulls — (1) timing permutation: entry times randomized, direction mix kept, same trade count, same non-overlap suppression, p=0.5527; (2) sign-flip permutation: each trade's realized move kept, direction re-rolled (variance-matched, stricter for fades), p=0.5412. Verdict uses the stricter p=0.5527 (2000 trials, seed 42, reproducible)
The null must match the live procedure trade-for-trade: every surrogate fills the same count (169 entries, 2000 effective trials). A surrogate that only fills ~9 entries widens the null about 4x and can fake p=0.30 out of p=0.01
Uncertainty: mean 95% bootstrap CI [-72bp, 28bp] — spanning 0 means no claim of edge
Sample: 11929 bars of 60m, 0 missing (max gap 0), OHLC complete
Splits: first 60% in-sample, rest out-of-sample, plus 4 time blocks to catch single-segment flukes
Params: period=14 lo=30 hi=70 side=both warmup=2000 — if tuned on this same history the verdict carries selection bias; collect out-of-sample before believing it
Known limits: no order-book depth so slippage is assumed; liquidation/margin/liquidity not modeled; rejected or partial fills not modeled
Null sensitivity: the sign-flip test keeps each trade's realized move and re-rolls only the direction, so the long/short allocation itself counts as evidence — in a one-way market that can inflate significance. Hence the verdict requires both nulls to clear, a positive net mean and a positive out-of-sample mean; miss one and it degrades to "not separated from luck"

This card only judges whether a rule has statistical evidence under stated costs. Not investment advice; past results do not predict future returns.