Future Fractal Pro
- インディケータ
- Su Watanabe
- バージョン: 1.30
- アクティベーション: 5
Future Fractal Pro
What is Future Fractal Pro?
A statistical pattern-matching indicator that scans up to 15,000 bars of historical data, finds the closest matching price patterns using k-Nearest Neighbors (k-NN) methodology, and projects the statistically most likely price scenario as future candlesticks on your chart. It displays a confidence rank (A/B/C), expected return in pips, and risk direction -- all in a single information panel. Designed by a data analyst. M30 timeframe, 6 major pairs. Non-repainting on confirmed bars.
The Core Idea
If you flip a coin 10 times, the results look random. But flip it 10,000 times, and the ratio converges toward 50/50. This is the Law of Large Numbers.
Now imagine tilting that probability to 51/49. Each individual flip still looks like luck. But over hundreds of flips, the cumulative result drifts steadily in your favor. That is the principle behind this indicator: not predicting one trade, but identifying conditions where the statistical edge tilts slightly in your direction -- and doing it consistently.
Modern FX markets are heavily influenced by algorithmic trading. These algorithms leave subtle patterns in the data -- recurring tendencies that are invisible to the naked eye but detectable through statistical analysis. Future Fractal Pro is designed to surface those patterns and present them as actionable visual information.
How It Works
The indicator performs three core processes on every confirmed bar:
1. Pattern Matching (k-NN with Pearson Correlation)
The current 30-bar price pattern is normalized and compared against up to 15,000 historical bars. Only patterns with strong positive correlation are selected. Time-of-day matching and ATR-ratio filtering further refine the candidate pool, ensuring that comparisons are made under similar market conditions.
2. Probability Visualization (Fan Chart)
From the top 30 matching patterns, the indicator calculates what happened next in each case. It then projects the median forecast as a main line, a 50% probability zone (interquartile range), and 80% probability boundary lines. This gives you not just direction, but a visual sense of how dispersed or concentrated the expected outcomes are.
3. Rank Assessment (Confidence + EV + Trend Filter)
The indicator combines pattern match confidence, expected return in pips, skewness (risk direction), and a 200-EMA trend filter to assign a final rank. Only when all conditions align does a high-confidence rank appear.
Rank System
Rank A - High confidence. Pattern match rate is high, expected return exceeds the minimum threshold, trend direction aligns with the forecast, and the 200-EMA filter confirms the macro trend. Alert notification is triggered.
Rank B - Moderate confidence. Conditions are favorable but one or more factors fall slightly below the Rank A threshold. Worth monitoring, but requires additional confirmation from your own analysis.
Rank C - Low confidence or range conditions. Expected return is insufficient, trend is counter-directional, or the market appears range-bound. The indicator recommends staying out.
Rank A thresholds are calibrated individually for each supported symbol based on historical performance characteristics. This is not a one-size-fits-all threshold.
Information Panel
The on-chart panel displays the following in real time:
Direction + Rank - BUY or SELL with Rank A/B/C, plus trend context (Trend-following, Counter-trend, or Range-like).
Pattern Match Rate - A composite confidence score integrating correlation strength, directional consistency, and dispersion. Higher values indicate stronger statistical alignment.
Expected Return - The average projected move in pips if the median scenario plays out. Positive values indicate potential profit in the forecast direction.
Risk Direction - Statistical skewness of the forecast distribution. "Balanced" means symmetric risk. "Upside Risk" or "Downside Risk" warns of asymmetric tail risk in one direction.
Search / Selected - Shows how many bars were scanned and how many candidate patterns were selected for the forecast ensemble.
Built-in Backtest Mode
Hold Ctrl and click any past bar on the chart. The indicator recalculates the forecast as if that bar were the current moment, showing you exactly what the prediction would have looked like at that point in time. Click normally anywhere on the chart to return to the latest forecast.
This allows you to verify the indicator's behavior across different market conditions without relying solely on the developer's claims. You can check for yourself how Rank A signals performed in past sessions.
Rank A History Arrows
When a Rank A signal is confirmed on a closed bar, an arrow is automatically placed on the chart at that location. These arrows persist as you scroll through history, giving you a visual record of where high-confidence signals occurred. Blue arrows indicate bullish Rank A signals, red arrows indicate bearish. These are recorded at the moment of confirmation and do not repaint.
Backtest Results (Text Summary)
To validate the statistical edge of Rank A signals, backtests were conducted under strict conditions: entry at bar open on Rank A confirmation only, fixed TP/SL of 50 pips, spread cost of 2.0 pips deducted. No curve-fitting or optimization was applied.
Results across supported symbols (M30, January 2022 to recent data):
GBPUSD - Profit Factor: 1.87, Win Rate: 64.3%, Max Drawdown: 3.41%
EURUSD - Positive expectancy confirmed
USDJPY - Positive expectancy confirmed
EURJPY - Positive expectancy confirmed
GBPJPY - Positive expectancy confirmed
AUDJPY - Positive expectancy confirmed
All six supported symbols achieved positive expectancy under these conditions. Past performance does not guarantee future results, but these results demonstrate that the logic is based on measurable statistical tendencies rather than random chance.
Why M30 and 6 Symbols Only
Shorter timeframes (M1, M5) have excessive noise that degrades statistical reliability. Longer timeframes (H4, D1) produce too few signals to be practical. M30 offers a favorable balance between forecast accuracy and signal frequency.
The indicator is limited to 6 symbols because Rank A thresholds are individually calibrated for each pair based on extensive testing. Expanding to untested symbols without proper calibration would compromise signal quality. As additional pairs are validated, they may be added in future updates.
Supported symbols: EURUSD, GBPUSD, USDJPY, EURJPY, GBPJPY, AUDJPY
Expected Rank A frequency: approximately 50-80 signals per symbol per year. With 6 symbols monitored, this translates to roughly 300-400 high-confidence opportunities per year (averaging 1-2 per trading day).
Forecast Limitations (Butterfly Effect)
The indicator projects up to 10 bars into the future. However, forecast reliability naturally decreases with each additional bar due to the butterfly effect (sensitive dependence on initial conditions). Bars 1-3 tend to be the most reliable. Bars 4-10 are best used as scenario guidance rather than precise targets. The indicator intentionally does not project beyond 10 bars.
Who This Is For
Recommended for:
- Discretionary traders who want a data-driven filter before entering
- Traders who struggle with overtrading during range-bound conditions
- Traders interested in statistical/quantitative approaches to trading
- Those who want to verify indicator behavior through built-in backtesting before committing real capital
Not recommended for:
- Traders looking for a fully automated Expert Advisor
- Traders expecting signals on every bar or every session
- Traders who only trade timeframes other than M30
- Traders who only trade symbols outside the 6 supported pairs
Parameters
Display Settings
Main forecast line color (default: Gold) - Color of the median forecast line.
Main forecast line width (default: 3) - Thickness of the median forecast line.
50% zone color (default: DarkSlateBlue) - Color of the interquartile probability zone.
80% lines color (default: DarkGray) - Color of the 80% probability boundary lines.
Bullish candle color (default: SpringGreen) - Color for projected bullish candlesticks.
Bearish candle color (default: OrangeRed) - Color for projected bearish candlesticks.
Panel Settings
Show information panel (default: true) - Toggle the on-chart information panel.
Panel background color (default: Black) - Background color of the panel.
Panel X/Y position (default: 0/0) - Offset adjustment for panel placement.
Base font size (default: 10) - Text size for panel elements.
Font preset (default: Segoe UI) - Font selection from preset list.
Custom font name (default: empty) - Override with any installed font name.
Alert Settings
Popup alert (default: true) - MT5 alert dialog when Rank A is confirmed.
Push notification (default: false) - Push to MT5 mobile app. Requires MetaQuotes ID setup.
Email notification (default: false) - Email alert. Requires SMTP setup in MT5.
Alerts fire only on Rank A confirmation. Rank B and C do not trigger alerts. Alerts are processed on confirmed bars only and do not fire during the first calculation cycle after startup.
FAQ
Q: Does the forecast repaint?
A: While the current bar is forming, the forecast updates with each new tick as new data becomes available. However, once a bar closes and a signal is confirmed, that confirmed result does not change. Rank A history arrows are placed at the moment of confirmation and remain fixed.
Q: Most of the time it shows Rank C. Is this normal?
A: Yes. Approximately 70% of market conditions lack clear directional bias. The indicator is designed to filter out these low-probability conditions aggressively. Rank C means the filter is working correctly -- it is preventing entries during unfavorable conditions.
Q: Why does it only work on M30?
A: M30 provides the optimal balance between noise reduction and signal frequency for this type of statistical analysis. Shorter timeframes introduce excessive noise, while longer timeframes reduce opportunity count. The Rank A thresholds are specifically calibrated for M30 behavior.
Q: Can I add more symbols?
A: Currently, only 6 symbols are supported because each one requires individual threshold calibration. Using the indicator on unsupported symbols will display an error message. Additional symbols may be added in future updates after proper validation.
Q: How do I use the backtest mode?
A: Hold Ctrl and click any past bar on the chart. The indicator recalculates the forecast from that point. Click normally (without Ctrl) anywhere on the chart to return to the latest forecast.
Q: What happens during major news events?
A: Technical analysis tends to be less reliable during major economic releases. It is recommended to avoid trading during these periods regardless of the indicator's rank. The indicator does not have a built-in news filter.
Q: Is there a recommended PC specification?
A: The indicator performs intensive statistical calculations (scanning up to 15,000 bars with correlation analysis on every confirmed bar). Recommended specs: Windows 10/11 64-bit, Intel Core i5 / AMD Ryzen 5 or higher, 8GB RAM minimum (16GB recommended). Low-spec environments (Celeron, low-RAM VPS) may experience slowdowns.
Q: Can I use Rank A as my only trading signal?
A: Rank A indicates that statistical conditions appear favorable based on historical pattern analysis. It is designed as a decision-support tool, not a standalone trading system. Combining it with your own analysis and risk management rules is recommended.
Key Features
- K-Nearest Neighbors pattern matching across up to 15,000 historical bars
- Median forecast line with 50% and 80% probability zones (fan chart)
- Future candlestick projection (up to 10 bars ahead)
- Three-tier rank system (A/B/C) with symbol-specific calibration
- Pattern match rate, expected return (pips), and risk direction display
- 200-EMA trend filter with range detection
- Built-in backtest mode (Ctrl+Click any past bar)
- Rank A history arrows on confirmed bars (non-repainting)
- Alert, push, and email notifications on Rank A confirmation
- Time-of-day matching and ATR-ratio filtering for pattern quality
- Skewness calculation for asymmetric risk detection
- Customizable colors, font, and panel position
- M30 timeframe, 6 major pairs (EURUSD, GBPUSD, USDJPY, EURJPY, GBPJPY, AUDJPY)
- No DLL required
About the Developer
Designed and developed by a data analyst specializing in statistical modeling and big data analysis. The indicator applies quantitative methods -- k-Nearest Neighbors pattern recognition, Pearson correlation scoring, percentile-based probability distribution, skewness analysis, and expected value calculation -- to provide a structured, data-driven view of potential future price behavior.
This is not a black-box system with hidden logic. The methodology is rooted in well-established statistical techniques, applied specifically to the challenge of short-term price forecasting on the M30 timeframe.
The developer actively monitors market data and plans to release updates incorporating new findings and expanded symbol support as validation is completed.
Concept: Replace gut feeling with statistical evidence. Trade only where the data supports it.
Disclaimer
This indicator does not provide investment advice. All forecasts, ranks, expected returns, and probability zones are reference information derived from historical statistical analysis. They do not guarantee future price movements or trading profits. Past backtest results do not guarantee future performance. All trading decisions are made at the user's own risk. The author assumes no liability for any losses incurred through the use of this indicator.
