VWAP sd30×2 reversion with weekday + hour-of-day + post-opening 15m skip filters. Tests whether 1-month seasonality patterns add value.
This strategy has not been validated against TradingView multi-year history yet — RM/day, trades/day, and monthly projection are only computed from authoritative TV stats. The 1-month sweep below is in-sample only and should not be used as a profit forecast.
“Applies weekday + hour-of-day filters from seasonality analysis. Preliminary — 1-month sample. In-sample the filters HURT stats (parent SQN 1.98 → -0.34); likely the seasonality signal is noise given N≈4 observations per weekday.”
The featured config below was not directly covered by the walk-forward grids (different dataset or timeframe, or the in-sample SQN was too low to promote to OOS evaluation). Treat the in-sample numbers with appropriate caution — no unseen-window validation yet.
This strategy: Seasonal-filter variant; not in the walk-forward grids. In-sample only — parent SQN 1.98 degraded to -0.34 under filters on the same legacy_5m 1-month window.
Did not show a stable edge in the sample — treat with caution.
{
"dataset": "legacy_5m",
"tf": "5m",
"params": {
"stdevPeriod": 30,
"stdevMult": 2,
"filterWeekdays": "Mon,Wed,Thu,Fri",
"filterHours": "10:00-12:30,14:00-17:00",
"skipFirstMinutes": 15
},
"trades": 17,
"winRate": 47.06,
"netProfitMYR": -610,
"profitFactor": 0.82,
"sqn": -0.34,
"maxDrawdownMYR": 1885
}