Prior Day Sweep Portfolio Overlay
An independent prior-day sweep reclaim sleeve layered onto the MNQ active-RTH speed anchor. The sleeve is not a duplicate of the anchor: it adds 121 unique trades and improves the combined prop-eval profile while preserving stress resilience.
Why this is a better next move than broad searching
This branch adds a low-correlation, non-overlapping sleeve to the strongest existing anchor. That is more useful for prop automation than another standalone strategy because the governor can choose between small independent edges while controlling total daily risk.
| Stress | Trades | Win | ExpR | PF | Max DD | Eval pass | Eval fail | Median | p95 DD |
|---|---|---|---|---|---|---|---|---|---|
| Baseline | 592 | 61.0% | +0.321R | 1.91 | 5.03R | 99.1% | 0.7% | 25d | $1,430 |
| +0.05R added cost | 592 | 60.6% | +0.271R | 1.73 | 5.38R | 96.0% | 2.5% | 29d | $1,717 |
| +0.10R added cost | 592 | 60.1% | +0.221R | 1.57 | 5.78R | 90.8% | 4.7% | 33d | $1,872 |
| Random 10% miss + 0.03R cost | 544 | 59.9% | +0.269R | 1.73 | 5.75R | 96.9% | 1.5% | 29d | $1,655 |
Entry, stop, target, and governor logic
The public page shows the research-grade operating shape. Exact thresholds, route weights, private code, and automation files remain private buyer material.
| Item | Validated public rule | Execution note |
|---|---|---|
| Market | MNQ 116, MES 5 | MNQ dominates the sleeve; retest before increasing contract size. |
| Direction | Long-only prior-day sweep reclaim | No short-side assumption is included in this validation. |
| Time window | 07:31-09:28 ET | Signals cluster in the pre/open transition; disable around major scheduled news until separately modeled. |
| Entry | Closed-bar reclaim after prior-day liquidity sweep | Intended next-bar execution only; cancel stale entries. |
| Stop | Immediate bracket stop from sweep/reclaim structure | No widening, averaging down, or rescue logic. |
| Target | Positive-R bracket, primarily 1.25R sleeve target | Do not stretch targets manually unless retested. |
| Portfolio governor | $200 eval risk, global daily stop and profit lock inherited from active-RTH speed anchor | One global risk governor must control all sleeves. |
| Data span | 2023: 19, 2024: 39, 2025: 48, 2026: 15 | 2026 holdout has 15 sleeve trades, enough for watchlist promotion but still not live-proof. |
What this validation is based on
The page is a research summary, not audited live performance. These are the internal files and rule assumptions behind the published metrics.
| Area | Source | Test note |
|---|---|---|
| Combined tradebook | work/independent_sleeve_candidate_validate_20260629/combined_anchor_plus_independent_sleeve.csv | 592 accepted trades after deduping sleeve against the anchor. |
| Unique sleeve tradebook | work/independent_sleeve_candidate_validate_20260629/sleeve_unique_trades.csv | 121 unique trades, zero overlap with the active-RTH anchor. |
| Validation summary | work/independent_sleeve_candidate_validate_20260629/independent_sleeve_validation_summary.csv | 5,000-path baseline, added-cost, missed-fill, and periodic-miss stress runs. |
| Original sleeve source | work/prop_alpha_tactical_sweeps_20260625/top_trades/10_X_prior_day_sweep_reclaim_021efa0d.csv | Prior-day sweep reclaim candidate found by the independent sleeve scout. |
| Rule template | work/prop_alpha_lab.py Topstep-style 50K template | Includes max loss, daily lockouts, payout-day logic, commission, slippage, and Monte Carlo reshuffling. |
| Data source | work/databento_validation_normalized | Normalized MES/MNQ continuous-contract intraday bars used by the local research lab. |
Risk note
Strategy research is educational software and trading research, not financial advice. This is a paper/demo research candidate, not a live-funded approval. Live markets can break backtests through fees, slippage, latency, platform behavior, missed fills, rule changes, and regime shifts.