123 Forex Trading Strategy: The Mechanical 1-2-3 Pattern, 10-20-50 Filter, and Backtested Realities

The Brutal Reality: Backtested Win Rates and Pair Dependency

Most trading resources present the 123 pattern as a high-probability holy grail. This is false. Industry estimates place the unfiltered 123 pattern win rate around 35% to 40%. However, our own educational backtest applying the strict 10-20-50 filter across six major pairs yielded a 31.88% win rate. This proves that the pattern is inherently low-probability and relies entirely on extreme pair dependency and high risk-reward ratios to survive.

A trader executing the 123 pattern must accept a psychological reality: you will experience losing streaks of 10 to 19 trades in a row. Our backtest recorded a worst losing streak of 19 consecutive trades. If your risk management is not strict, these inevitable losing streaks will destroy your account before the statistical edge can manifest.

This page does not teach you how to find every 123 pattern on a chart. It teaches you how to filter out the 68% of patterns that are destined to fail, using the mechanical 10-20-50 rule, and how to execute the remaining 32% with mathematical precision.

The Anatomy: Defining P1, P2, and P3 Structurally

The 123 pattern is a three-swing price action formation. Unlike generic reversal strategies that rely on single candlesticks, the 123 requires a complete structural sequence. For a bullish 123 (reversing a downtrend):

• Point 1 (The Initial Extreme): The absolute lowest structural low of the downtrend. This is a 5-bar fractal low (a low that is lower than the two bars to its left and the two bars to its right).
• Point 2 (The Intermediate Pullback): A corrective swing high. Point 2 establishes the confirmation level. The pattern does not exist until the price breaks beyond this level.
• Point 3 (The Momentum Failure): The second push downward. In a valid 123, Point 3 must be a higher low than Point 1. This failure to make a new extreme is the mechanical proof that selling pressure is exhausted.

The trade is triggered when the price breaks above Point 2. If Point 3 breaks below Point 1, the pattern is immediately invalidated. The trend is continuing, and the reversal thesis is false.

The 10-20-50 Filter Rule: Eliminating Fakeouts

Taking every 1-2-3 sequence you see is the fastest way to lose money. The 10-20-50 rule provides a strict, mechanical filter to separate high-probability structural setups from market noise.

A 123 pattern that forms over 40 bars with a 30% retracement is statistically far more likely to result in a false breakout at Point 2 than a pattern that forms over 15 bars with a 61.8% retracement. If the setup does not meet the 10-20-50 criteria, skip it.

Three Entry Methods: Standard, Cheat, and Continuation

Depending on your risk tolerance and the market context, there are three distinct ways to enter the 123 pattern.

1. The Standard Entry (Point 2 Breakout)

This is the most reliable method. Wait for a daily or 4-hour candle to close decisively beyond the Point 2 level. Unlike a generic breakout strategy that trades any structural break, the 123 breakout is strictly conditional on the prior momentum failure at Point 3. The stop loss is placed below Point 3 (plus the liquidity buffer, detailed below).

2. The Cheat Entry (Trendline Breakout)

By drawing a trendline connecting Point 1 and Point 2, you can anticipate the breakout. The entry is triggered when the price breaks this trendline before it actually reaches the horizontal Point 2 level. This provides a much tighter stop loss and a superior risk-reward ratio. However, the trade-off is a significantly higher rate of false signals. Only use the Cheat Entry when the 10-20-50 rule is perfectly met and there is strong RSI divergence between Point 1 and Point 3.

3. The Continuation Entry (Point 3 Breakout)

This is a structural variation. If the price approaches Point 2 but fails to break it, and instead turns down to break below Point 3, the reversal has failed. However, this failure creates a high-probability setup: the original trend is continuing. Enter in the direction of the original trend when the price breaks Point 3. The stop loss is placed above the failed Point 2 level. This turns a failed reversal into a trend continuation trade.

Take-Profit Targets: Geometric Midpoint and Fibonacci Extensions

Most resources suggest using an arbitrary 1:2 risk-reward ratio. This ignores the internal geometry of the formation. The 123 pattern has specific, mathematically derived take-profit targets.

The Geometric Midpoint Formula

1. Identify the exact price level of Point 1 and Point 3.
2. Calculate the midpoint between Point 1 and Point 3: Midpoint = (Point 1 + Point 3) / 2.
3. Measure the vertical distance between Point 2 and the Midpoint: Distance = Point 2 - Midpoint.
4. Project that distance upward from the Point 2 breakout level to find the target: Target = Point 2 + Distance.

Alternative Targets: Fibonacci Extensions

Professional traders often scale out of the 123 pattern using Fibonacci extensions of the Point 1-to-Point 2 impulse wave. The highest probability extension targets are:

• 127.2% Extension: An excellent target for scaling out 50% of the position.
• 161.8% Extension: The structural equilibrium for the full reversal. This often aligns perfectly with the geometric midpoint target.

Our backtest data shows the average win was only +0.92R, significantly lower than the 2R geometric target. This indicates that the price often reverses before reaching the full structural equilibrium. Traders should strongly consider scaling out at 1R or the 127.2% Fibonacci extension rather than waiting for the full geometric target.

Stop Loss Placement and the Liquidity Sweep Buffer

Placing your stop loss exactly at Point 3 is a structural error that will cost you money in live trading. Point 3 is an obvious structural level where retail traders place their stops. Institutional algorithms frequently push the price slightly beyond Point 3 to trigger these stops (a liquidity sweep) before reversing.

To survive this, you must add a structural buffer to your stop loss:

• For a Long Setup: Stop Loss = Point 3 Low - (Current Spread + 5 Pips).
• For a Short Setup: Stop Loss = Point 3 High + (Current Spread + 5 Pips).

This buffer ensures that if the price wicks below Point 3 to hunt stops, your order is not triggered. If the price closes structurally beyond your buffer, the pattern is genuinely invalidated, and you want to be out of the trade. Always calculate your position size based on this buffered stop distance, not the raw Point 3 distance. Use the FXGlory margin calculator to ensure your buffered stop aligns with your 1% risk rule.

123 Pattern vs. Head & Shoulders, Double Bottoms, and ABCD

To avoid confusion and strategy overlap, it is critical to distinguish the 123 pattern from other common formations.

No-Trade Conditions for the 123 Setup

The following conditions invalidate the 123 pattern or make the risk too high to justify entry:

• Skip if Point 3 breaks Point 1. The momentum failure thesis is immediately false.
• Skip if the distance between Point 1 and Point 3 is less than 10 bars or more than 20 bars (violating the 10-20-50 rule).
• Skip if Point 2 retraces less than 50% of the impulse wave to Point 1. The pullback is too shallow to confirm a genuine reversal.
• Skip if you are trading EURUSD, GBPUSD, AUDUSD, or USDCHF without extreme confluence. Our backtest proved these pairs yield win rates below 28% for this pattern.
• Skip if the Point 2 breakout occurs during a major fundamental news release. The slippage and spread widening will destroy the structural risk-reward.
• Skip if you cannot add the structural buffer to your stop loss without making the position size violate your 1% risk limit.

Backtesting the 123 Pattern with Fractal Logic

Backtesting the 123 pattern manually is highly prone to hindsight bias. Because the pattern requires identifying swing points, traders will subconsciously ignore the setups that failed and remember the ones that worked. To measure the 123 pattern accurately, the rules must be strictly programmatic using fractal detection:

• Define Point 1 as a 5-bar fractal low (the lowest low of a 5-bar window).
• Define Point 2 as the next 5-bar fractal high.
• Define Point 3 as the next 5-bar fractal low, with the strict condition that Point 3 > Point 1.
• Calculate the 50% retracement of the impulse wave and ensure Point 2 meets this threshold.
• Execute the entry only on the close of the bar that breaks Point 2.
• Apply the structural stop-loss buffer (Spread + 5 pips) to every trade.

Educational Backtest: 123 Pattern Across Six Pairs (2014–2024)

The following results were generated from yfinance public research data using the fractal logic, 10-20-50 filter, and structural stop buffer described above. The data source was public research data, not FXGlory broker execution data.

Combined Metrics — All Pairs, All Sensitivity Runs

Pair-Level Comparison

The backtest confirms the brutal reality of the 123 pattern. Four findings stand out:

1. Extreme pair dependency. The strategy generated a 63.16% win rate and +0.26R expectancy on USDJPY. On EURUSD, the win rate was 22.5% with a -0.54R expectancy. The identical rule model produced completely opposite results based solely on the instrument. A trader who only backtested USDJPY would believe the strategy is highly profitable; a trader who only tested EURUSD would abandon it immediately.

2. The hidden cost of holding time. The average holding period was 35.96 days. Swap and rollover costs were not included in this test. For a trade held over a month, negative swap on certain pairs could easily erase the small positive expectancy on USDCAD and turn the USDJPY edge into a net loss. The 123 pattern is not a quick scalp; it is a structural swing that requires patience and favorable swap conditions.

3. The target is rarely hit fully. The average win was +0.92R, significantly lower than the 2R geometric target. This indicates that the price often reverses before reaching the full structural equilibrium, or the 60-bar max hold forces an exit at a smaller profit. Traders should consider scaling out at 1R rather than waiting for the full geometric target.

4. The structural buffer works, but doesn't save bad setups. The average loss was -0.81R, slightly better than the planned 1R risk. The 5-pip structural buffer kept traders in the trade during minor liquidity sweeps, but ultimately the stop was hit on the majority of setups. The buffer prevents premature stop-outs, but it cannot fix a pattern that lacks genuine momentum failure.