Does the Fair Value Gap (FVG) Strategy Work?

DiscardedNot in this test. Both versions of a popular YouTube FVG strategy, which claims a 70% and an 81% win rate over one month, won 28% and 35% of trades over almost 5 years of data (2018-2022) on the three US indices, with costs. Neither beats random entries with the same exits, and the day trading version loses all five years.

Tested on · Updated · José Martínez Hernández

Key facts

Markets
Nasdaq 100, S&P 500 and Dow Jones (CFDs)
Period
April 2018 to December 2022, 1-minute data
Trades
1,184 scalping and 1,282 day trading
Win rate
28% and 35% (the video: 70% and 81%)
Profitable years
2 of 5 (scalping) and 0 of 5 (day trading)
Compared with
200 runs of random entries with the same exits and costs
Costs
The broker’s spread and commission on every trade
The original video
One month of backtest chosen by the author, no costs

What was tested

A Fair Value Gap strategy from a YouTube video with more than 400,000 views. The video gives two versions, one for scalping and one for day trading, and backs them with a one-month backtest chosen by its author, with no costs: a 70% win rate for scalping and 81% over 16 trades for day trading. Both were tested, with their rules exactly as taught.

The rules

Both start the same way. Take the first candle of the New York session, from 9:30, and wait for a large candle to break its high or low and leave a Fair Value Gap. The first gap of the day sets the direction.

  • Scalping: the first 5-minute candle (9:30-9:35) and the gap on the 1-minute chart. Wait for price to come back into the gap and for an engulfing candle, then enter at market. Stop below the pullback low and target at 3 times the risk.
  • Day trading: the first 15-minute candle (9:30-9:45) and the gap on the 5-minute chart. Limit order at the edge of the gap, stop below the first candle of the gap and target at 2 times the risk.

What the video leaves out was fixed before looking at any data: one trade per day, entries until 11:00 New York time, the scalping trade is cancelled if a candle closes on the other side of the gap, and anything still open is closed at 15:55.

Markets, years and costs

  • Nasdaq 100, S&P 500 and Dow Jones, as Darwinex CFDs, with 1-minute data.
  • From April 2018 to December 2022: almost 5 years, against the video’s one month.
  • The broker’s spread and commission on every trade.

The comparison with random entries

The strategy was compared with 200 runs of random entries in the same window, with the same stop, exits, costs and order type: for day trading, the random entries also use a limit order. To pass, the strategy had to land clearly above what random entries achieve.

Results

  • Day trading: wins 35% of trades, against the video’s 81%. For every $1 it loses, it makes back 85 cents, and it loses all five years. It sits below what random entries typically do: the limit order gets filled precisely on the days when price keeps going against it.
  • Scalping: wins 28% of trades, against the video’s 70%. Its curve runs inside the random band, close to the edge, and it is profitable in only two of the five years. The stop is so tight that spread and commission take on average 10% of what each trade risks, and more in the early years. Without costs it would make a little, but so would the random entries.

Neither version beats random entries with the same exits. Both are discarded before coding the bot and without risking any money.

Day trading: the curve against 200 runs of random entries
2019202020212022Break-even
  • Strategy (day trading)
  • Random entries, median
  • 90% of random runs
  • Middle half of random runs

Cumulative result from April 2018 to December 2022 on the three indices, with costs. The curve ends below the random median, close to the lower edge of the band. Random entries lose too: they pay the same costs.

Scalping: the curve against 200 runs of random entries
2019202020212022Break-even
  • Strategy (scalping)
  • Random entries, median
  • 90% of random runs
  • Middle half of random runs

Same period and markets. The curve runs inside the random band, close to its upper edge, without ever breaking out of it.

Limits of this test

The video

The video is in Spanish.

Frequently asked questions

What is a Fair Value Gap (FVG)?

A gap in price: three candles in a row where the middle one moves so fast that the first and the third don’t overlap. The space between them is the gap, and the idea is that price tends to come back to it. It comes from the ICT method and is one of the most popular ideas on trading social media.

Why does the original video claim an 81% win rate?

Because it is a one-month backtest chosen by its author, with 16 trades and no costs. With so few trades there’s no way to tell an edge from luck. Over almost 5 years, with 1,282 trades and costs, the same version wins 35% of trades.

Why compare it with random entries?

Because a strategy can win or lose for reasons that have nothing to do with its signal: the stop, the target, the costs or the market’s trend. Random entries keep all of that the same, except the signal. If the strategy doesn’t beat them, its rules add nothing.

Does it work on gold or forex?

It was never measured, because it failed the test on the indices, which is where it is taught. The cost estimate made beforehand already showed that the scalping version, with such a tight stop, would not cover its costs on gold or forex.

What about the ICT Silver Bullet?

It is almost the same day trading version half an hour later, with a liquidity sweep filter. It was tested too, on indices, gold and forex, and it doesn’t beat random entries either.

Source

Original video tested: «The Best FVG Trading Strategy (That Actually Works)» (Casper Trading, YouTube).This study tests these rules, on these markets and over this period; it does not judge the rest of the author’s work.

Backtest results with costs, not live trading results. Past results do not guarantee future results. This is not investment advice.

Free course

Learn to test any idea like this, with AI

The free course “Introduction to Algorithmic Trading with AI” shows the full process, from a video idea to a verdict, before you risk any money.

Start the Free Course

More strategies tested

See all the strategies tested →

Trading in financial markets involves a high risk of capital loss and is not suitable for all investors. Leveraged products such as futures, CFDs, or options may result in losses greater than the amount invested. The information and materials provided on this page are for educational purposes only and do not constitute financial advice or investment recommendations. Each individual is responsible for their own investment decisions and must ensure that the products they trade are appropriate for their risk profile. GreaterWaves and its representatives do not accept any liability for direct or indirect losses arising from the use of the information shared here.