The NY Liquidity Model
One instrument, one session, one pattern. The system watches the first hours of the New York cash session on the Nasdaq-100 future and trades a single event: a raid on an overnight liquidity level followed by a fair value gap in the direction of the day. Everything below is the complete rule set of version 3.2, validated on 5 November 2026.
The trading day, in order
- Direction
Two readings have to agree before anything is traded: the pre-market move from 07:00 to 09:30 New York time, and the colour of the 09:30 to 10:00 candle. If they disagree, the day is off. On the test year this filter alone separated a profit factor of 1.4 from 2.1.
- Levels
From 09:30 the system watches six untouched levels: the Asia high and low (20:00 to 00:00), the London high and low (02:00 to 05:00), and the previous day's high and low.
- Raid
A 5-minute bar that trades through one of these levels is a raid. A raid against the day's direction sets up a reversal; a raid with the direction sets up a continuation. Both are traded, in the direction of the day only.
- Entry
Between 10:00 and 12:00, the first 5-minute fair value gap after the raid, in the day's direction, triggers a market order on the close of the bar. The raid expires after 18 bars without a gap.
- Stop
Beyond the extreme of the sweep (reversal) or of the three-candle formation (continuation), plus four ticks. A setup whose risk exceeds 150 points is skipped.
- Target
The nearest untouched liquidity level, if it offers at least 1.5 R; otherwise two times the risk. At plus one R the stop moves to break-even.
- Time
A trade still below its entry after 60 minutes is closed. Nothing is held past 15:55 New York time, and nothing overnight.
- Day limits
At most three trades a day, stop after two winners, stop after 300 $ of daily loss per Micro contract.
Twelve months on NAS100 5-minute data
| Trades | 87 |
| Profit factor | 2.14 |
| Net result per Micro contract | +4 881 $ |
| Winning trades | 30 % |
| Average trade | +0.28 R |
| Largest drawdown | 573 $ |
| Longest losing streak | 8 trades |
| Months positive | 9 of 13 |
| Worst month | −298 $ |
| Both half-years positive | +2 738 $ and +2 143 $ |
Simulator figures, conservative fills: if stop and target both lie inside a bar the stop counts first; one tick slippage per side; 1.20 $ commission per round trip. The TradingView deep backtest of the same version on MNQ1! (October 2025 to November 2026) shows 83 trades, profit factor 1.86, +3 954 $ and a drawdown of 830 $. The two engines use different data feeds and fill models; both are published here on purpose.
How the 87 trades ended
Trade list and changelog
Every trade of the test year with date, level, direction, entry, stop, target, exit and result: download the CSV (87 rows).
- v3.2, 5 November 2026
Pre-market direction read from the running open; session values in Pine were one bar late and read the previous day. Time exit made price-based (close below entry) instead of P&L-based. - v3, 5 November 2026
Pre-market move must agree with the opening candle. Time exit after 60 minutes. Three trades, two winners a day. Profit factor 1.42 to 2.14, drawdown 916 $ to 573 $. - v2, 4 November 2026
Direction from the 09:30 to 10:00 candle. Market entry on the FVG. 158 trades, profit factor 1.42, +4 845 $, drawdown 916 $. - v1, October 2026
Pure raid-and-reversal without a direction rule. Profit factor around 0.7 in every variant. Abandoned.
What was tested and rejected
- Limit entry inside the gap instead of market entry: worse in every variant.
- A midday stall exit at 12:00: costs money.
- Composite overnight bias (London versus Asia, gap, prior close location), prior-day bias, initial-balance break rule: no benefit.
- A fixed 200 $ daily target: halves the result. 'Stop after 200 $' changes nothing, because only one day in four reaches it with one Micro.
- Tight trailing stops and pyramiding: worse ratio. Scaling only by fixed size, two Micros from about 3 000 $ of account.
- The same pattern on the S&P 500 future (MES): loses in every variant, profit factor 0.89 to 0.99. The Nasdaq is the instrument.
- Opening range breakout (120 variants) and opening retracement without a gap (78 variants): no edge on MNQ.
- London session on NQ: profit factor 0.73 to 1.00 in twelve variants.
- Pre-market window start from 04:00 to 07:00: profit factor 1.9 to 2.2 throughout; from 08:00 it breaks down. The rule is not sitting on a knife edge.