WickCheck / Trading school / Blog
Strategy research

ICT New York Session Strategy: Liquidity Sweep + FVG Tested

18 November 2026 · 4 min read · by the WickCheck team

This is one of the most popular ICT-style intraday models: during the New York session, wait for price to sweep liquidity (take out a recent high or low), then enter on the fair value gap that forms as price reverses. The sweep sets the direction.

We tested it for 3 years on 15-minute data across five markets, with costs, a news filter and 1% risk per trade. It made money on two of the five.

Key takeaways
  • Gold: 67 trades, 36% win rate, profit factor 1.16, +6.9% at 1% risk, max drawdown −9.1%.
  • GBP/USD did best: 55 trades, profit factor 1.48, +17.9%.
  • It lost on EUR/USD (0.89), Nasdaq futures (0.80) and S&P 500 futures (0.83).
  • Small samples and two losing years in a row on gold: promising on some markets, not proven.

In this guide: The exact rules · Results on gold · Results on 5 markets · What this means for ICT traders · What to test next

The exact rules

Between 13:30 and 16:00 London time, wait for a liquidity sweep, then wait for a fair value gap. Enter on the FVG. Stop above the sweep. Target liquidity, at least 2R. Don't trade on days with red news.

How WickCheck reads it: trades both ways within 13:30 to 16:00 London time. Step 1 is a sweep of any untaken high or low (swings, equal highs and lows, the previous day, sessions). Step 2 is a 15-minute fair value gap within 20 bars of the sweep. Entry is a limit order at the FVG edge, the stop goes beyond the sweep wick and the target is the nearest liquidity at least 2R away. Days with high-impact US news in the session (CPI, NFP, PPI, JOLTS, FOMC) are skipped.

Results on gold

From November 2023 to November 2026 on 15-minute gold (XAU/USD): 67 trades, 36% winners, a profit factor of 1.16, +0.11R per trade, +6.9% at 1% risk and a maximum drawdown of −9.1%.

By year: 2023 +2.6R (7 trades), 2024 −0.7R, 2025 −3.1R and 2026 +8.7R. Most of the profit came in 2026, which is a warning sign: an edge concentrated in one year is fragile. Numbers are from the WickCheck strategy library as of 18 November 2026. The library re-runs every test as new candles arrive, so the live pages may differ slightly.

Results on 5 markets

Same model, very different outcomes. On EUR/USD and the index futures the win rate fell to 26% to 28%, too low for targets of about 2R to make up after costs.

What this means for ICT traders

The concepts are not the edge on their own. The market, the session and the filters are. Compare this with the ICT killzone test, where a sweep + FVG model in the 07:00 to 10:00 New York window broke even or lost on Nasdaq futures, gold and EUR/USD, and the Silver Bullet test, which lost on both Nasdaq and S&P 500 futures.

With 55 to 79 trades per market, none of these results is conclusive. They are a starting point for your own testing, not a green light. Our losing streaks guide explains why a 36% win rate means long runs of losses even when the edge is real.

What to test next

Practise the setup bar by bar in chart replay before trading it live, and see our ICT backtesting page for more ICT tests.

Bottom line. An ICT-style New York sweep + FVG model made money on gold and GBP/USD and lost on EUR/USD and US index futures. Test it on your market before you trust it.

Test the NY sweep + FVG model yourself

Same rules on any market, or write your own version in plain English.

Open the test

Questions

Does the ICT New York session strategy work?
In WickCheck's 3-year test with costs, the sweep + fair value gap model was profitable on gold (profit factor 1.16) and GBP/USD (1.48) but lost on EUR/USD, Nasdaq and S&P 500 futures. Samples were 55 to 79 trades per market.
What time is the New York session in London time?
This test used 13:30 to 16:00 London time, which covers the US open and the first hours of the New York session.
Is ICT trading profitable?
It depends on the exact rules, the market and the costs. Most textbook ICT setups in WickCheck's library lost money once every signal was counted, but some versions held up on specific markets.

Keep reading

Which Trading Strategies Actually Work? 15 Backtested on Real Data

We backtested 15 popular trading strategies, from ICT setups to RSI and moving averages, on years of real data with costs. Here is what held up.

Losing Streaks in Trading: How Many Losses in a Row Are Normal?

How long losing streaks get at different win rates, the exact odds over 100 trades, what they cost at 1% risk, and how to survive them.

How to Pass a Prop Firm Challenge: Rules, Risk and a Realistic Plan

How prop firm challenges work, why most traders fail, and a realistic plan to pass: risk per trade, daily limits, drawdown buffers and testing first.

EMA Crossover Strategy on Gold: What the Backtest Shows

We tested the 9/21 EMA crossover with a 200 EMA filter on gold, EUR/USD and the Nasdaq, with costs included. Gold lost money after costs over 10 years.

Follow WickCheck for a new strategy test every day: Instagram · TikTok · X

Educational content, not investment advice. Trading involves risk of loss.

One tested strategy a week

A popular strategy backtested on real data, with the honest numbers. Free, one email a week, unsubscribe any time.