NQ Kalman Trend Cross
Trend cross of price against a self-adjusting 'Kalman' smoothing line.
Last validated: 2026-10-16 (baseline; weekly cadence starts after launch) · Validation status: Flagged — Year 1 lost money; weak recent form; single timeframe.
1 full-size NQ contract. Test window 2024-09-23 to 2026-09-23 (2 years). Slippage 4 ticks per order, $2.50 commission per contract per side ($5 round trip); $100k account. 4-tick slippage per side = NQ $20, ES $50, GC $40 before commission.
Plain-English rules
- Market
- Nasdaq-100 E-mini (NQ)
- Timeframe
- 30m — 30m only
- Direction
- Long & short
- Style
- Trend following
- Frequency
- about 66 trades a year
- In stacks
- —
- Long when the close crosses above the Kalman line (process noise 0.01, measurement noise 1.0), short when it crosses below, with the shared NQ filters.
- Exits: shared stop and trailing stop, the opposite cross (70 of 132 trades), end-of-day flatten (52), stops (9).
- Shared NQ filters: Regular session only; entries from 20 min after the open to 14:30 ET; long and short, one position at a time, max 2 trades/day; trend and volatility filters must agree (price vs 100-EMA and its slope, ADX(14) ≥ 18 with DI agreeing, ATR(14) ≥ 10 pts); initial stop 2.2×ATR then a 2.8×ATR trailing stop; flat at 16:00 ET.
Summary of the rules. The full parameter-level rules document is delivered to subscribers (documents only — no code).
Equity curve & drawdown
Year-by-year
| Period | Net | PF | Trades |
|---|---|---|---|
| Year 1 (exits to 2025-08-31) | -$3,875 | 0.92 | 68 |
| Year 2 (from 2025-09-01) | $52,390 | 2.01 | 64 |
108% of the 2-year profit came in year 2. Year 1 = exits 2024-09-23 to 2025-09-22; year 2 = exits 2025-09-23 to 2026-09-23 (saved trade list clipped to the window: 132 trades, +$48,515, PF 1.49; closed-trade DD $10,490). Per TV4REAL's NQ/GC stack files.
Recent form
Last 3 months (exits 2026-06-24 to 2026-09-23): 13 trades, +$10,840, PF 2.07. An earlier cut (extended list, last 3 months: 13 trades, PF 0.40, −$8,705. The 4 trades after the window lost $5,810.) ended on a later date and told a different story — recent form is unstable on small samples.
Robustness
Robustness checks not run. 15m has PF 1.2 with a $27.8k drawdown; 1h and 5m lose.
Other timeframes
| Timeframe | Net | PF | Win % | Trades | Max DD | Result |
|---|---|---|---|---|---|---|
| 15m | +$36,475 | 1.22 | — | — | $27,812.50 | DD over the cap |
| 1h | −$4,780 | 0.93 | — | — | — | loses |
| 5m | −$75,695 | — | — | — | — | loses |
Flags for this strategy
- FRAGILE (Tier B). Year 1 lost money (PF 0.92); 107% of the 2-year profit came in year 2.
- Recent form weak (3-month PF 0.40).
- Single timeframe (30m only).
- Inside the NQ stack tests Kalman adds only +$5,410 on 47 taken trades, and with DPO it pushes the stack's PF down. Kept out of the NQ Stack.