Multi-Timeframe Curve-Matching Forecast Indicator

I am looking for an experienced MQL5 engineer to build a custom MetaTrader 5 indicator that scans historical price data and finds the best curve/shape match to the current price action, then projects the historical continuation forward as a forecast.

This is not a standard bar-by-bar indicator and not a moving average/trendline tool.
The goal is to compare the shape of price action curves across history, and optionally across multiple timeframes, to find similar past structures.

What I Want the Indicator to Do

1. Take the most recent price action window from the current chart.

2. Convert that recent price action into a normalized curve/shape representation.

3. Scan historical data to find the best matching curve from prior history.

4. If no strong match exists on the current timeframe, also scan other timeframes.

5. Return the best historical analog and project what happened next after that analog.

6. Plot:

   • the matched historical curve

   • the projected continuation/forecast

   • optional confidence/similarity score
     
     

Important Clarification

I do not want simple candle matching or standard technical indicators.
The system should focus on curve similarity / shape similarity / structure similarity, where bars are only a sampling method of the curve.

The matching logic should prioritize:

• overall curve shape

• swing structure

• peaks/troughs

• slope changes

• curvature / path similarity

It should be able to compare patterns even when:

• price scale differs

• volatility differs

• timeframe differs

• the shape is slightly stretched/compressed

Technical Requirements

• Platform: MetaTrader 5 / MQL5

• Deliverable: working custom indicator source code

• Clean, readable, well-structured code

• Efficient enough for practical chart use

• No repainting / no future-data cheating

• Historical scanning must only use data available up to the matching point

Preferred Matching Approaches
I am open to the engineer’s recommendation, but the solution may involve one or more of:

• normalized curve comparison

• correlation on normalized sequences

• dynamic time warping (DTW)

• swing-point matching

• pattern similarity scoring

• multi-timeframe analog search
  
  

Functional Features
The indicator should allow configurable inputs such as:

• current pattern length

• search depth (for example previous 500 / 1000 / more bars)

• future projection length

• timeframes to include in search

• similarity threshold

• number of best matches to evaluate

• smoothing / normalization settings

Expected Output on Chart

• current price path

• best matched historical curve

• projected future path

• optional alternative matches

• optional score/label showing:

  • matched timeframe

  • location in history

  • similarity/confidence score

Nice to Have

• support for top 3 matches instead of only best 1

• weighted average forecast from best matches

• confidence zone/band

• performance optimization for large scans

• modular design so future ML/Python integration is possible later

Important

Please do not apply if your approach is only based on:

• moving averages

• trendlines

• simple candlestick pattern matching

• repainting logic

I need someone who understands historical analog matching by curve/shape, preferably with multi-timeframe scanning.