Pattern finder robot.

[OntradingX]

Hello friends, I am trying to develop an EA that finds repeated patterns on the chart daily.
Has anyone done or know where to start this project?
I'm starting this idea as the best idea to develop a day trading robot.
Tips accepted.
Basically, this robot will have to check how many times a certain pattern repeats during the day/month/year, and based on that an EA will be developed to mathematically take advantage of stop loss and take profit targets from this repetition.

[OntradingX]

@ontradingx What I understand as a pattern (something that always happens) doesn't matter the day, the time, but it always happens.
It can't just be a simple pattern, it can be for example:

1. 10:00 AM candle on EURUSD currency is always bullish and above 500 point.
2. A sequence of 3 candles always happens at 13:00 in AUDCAD.
3. A Red candlestick will always cross the central 20-period Bollinger Band at 07:30 on XAUUSD.

The idea is to generate a counter, and whenever that certain pattern that was (pre-defined) by the EA happens, it registers in this counter, then as many patterns as possible in the EA would be registered, and so Backtest, depending on the result, let's say that what else has occurred (has happened), a second step would be analyzed, (after this pattern is located) what did the market do? (did it go down, did it go up?) (how many points) and based on that a strategy would be elaborated, optimized and put to the test in a real market.
The biggest problem is that I have no idea where to start doing this.
If you have any ideas I would like to know.

[TipsyWisdom]

This is the difference between what is building an EA and what is designing an EA. Where most people think looking at a chart filled with indicators and either adding or eliminating is how you "optimize" or build an EA. That entire thought process is wrong and is not how EA's are to be designed.

What you have thought of is a price action pattern that your goal should now be to replicate the most perfect trade that you see happen with that pattern. When you design and frame that trade, you should have a best case stoploss and a worst case profit target in mind (the lowest profit, with the highest stop loss that you would accept on that specific setup on that specific time frame).

With that information, now you go through history on several pairs and time frames and find how many times that pattern repeated itself. When you optimize to find that information, if you have any winning scenarios leftover, that means you have found that many # of trades that gives you that many # of profitable scenarios at a minimum with that maximum accepted stoploss.

Now throw in a command that terminates the EA after 20 consecutive failed trades and you can sort through "optimizations" very quickly.

[OntradingX]

@ontradingx Sorry my friend, maybe because of language I don't understand if in your view, I'm (right) or (wrong) in wanting to develop this EA.
Can it be clearer?

[l'andorrà]

@ontradingx You can do that but that can be an immense amount of work for potential little use. INHO you should find a reliable custom indicator doing that job instead. You will save time and suffering, trust me.