Forex strategy feedback

[VictoriaAlgo]

Hi all,

Trading, coding and designing trading strategies can be a very lonely experience with very little feedback, but we don't complain because we love it. Anyway, I am starting this thread, so I can discuss ideas and perhaps get a little feedback about my designs and trading.

Thank you for reading

 

[VictoriaAlgo]

Okay. Let me kick off. What do you think of the performance of this back-test? It is a strategy I have been working on for nearly 12 months, the question is...can it operating like that into the future? Yes, the relative drawdown is very high indeed, but we can just use less risk and a larger account size.

Feedback please.

 

[Rufus_Leakey]

Okay. Let me kick off. What do you think of the performance of this back-test? It is a strategy I have been working on for nearly 12 months, the question is...can it operating like that into the future? Yes, the relative drawdown is very high indeed, but we can just use less risk and a larger account size.

Feedback please.

View attachment 294923

Going from $100 to $58,043,449.44 for the almost 15 years of the backtest is a compound annual growth rate of about 145.1 percent.

It's an ok return for Trade2Win (this isn't EliteTrader afterall), but not quite as good as

From $1 to Million - Results Updates

Hi everyone, I decided to share my journey on this forum. Below are the results so far: https://www.myfxbook.com/members/onlineaddict/from-1-million/6354122 Trading history is open, nothing is hidden(other than open trades), so feel free to explore and hit me up with any questions. And yes, I...

www.trade2win.com

If you think it's realistic, try it with a small amount of real money.

 

[tomorton]

Feedback please.

View attachment 294923

Its not possible to evaluate a strategy from the equity curve. For example the equity curve of a Martingale system could look superb, rising steadily and for a long period of time - until it crashes to zero in 10 easy trades. The flight records and 27-year operating career of the Concorde were both superb, until it crashed.

Its also obviously impossible to optimise a strategy without knowing its component elements.

On top of which, there's a general rule in finance that you should never put your money in something you cannot understand.

So you're showing us a stellar performance but excluding anyone from getting interested. Why?

 

[CavaliereVerde]

For every 10.000 backtests you can find in forums ... you can find ONE live trackrecord.
😁
Profit Factor 1.38 is possible but there is a relevant probability that the parameters are overoptimized.
Usually the higher the performance the lower the robustness.
A backtest out of the context doesnt' tell nothing about robustness.
Lenght and n. of trades is ok but we should know the number of parameters and the optimization method.

With a growing backtest you have 1% of what you need to beat the market.
You are still missing:

1. patience
2. risk tolerance
3. persistence

 

[CavaliereVerde]

@Lightning McQueen
@Trader333
I think this topic should be moved to Algo Trading forum .

 

[VictoriaAlgo]

@Lightning McQueen
I think this topic shoud be moved to Algo Trading forum .

Thank you for the replies.

How do I move it?

 

[CavaliereVerde]

How do I move it?

You can't.
Only moderators can do that.

 

[VictoriaAlgo]

You are still missing:

1. patience
2. risk tolerance
3. persistence

That is so true!!! Not enough is written about those three points and algo trading.

 

[VictoriaAlgo]

Its not possible to evaluate a strategy from the equity curve. For example the equity curve of a Martingale system could look superb, rising steadily and for a long period of time - until it crashes to zero in 10 easy trades. The flight records and 27-year operating career of the Concorde were both superb, until it crashed.

Its also obviously impossible to optimise a strategy without knowing its component elements.

On top of which, there's a general rule in finance that you should never put your money in something you cannot understand.

So you're showing us a stellar performance but excluding anyone from getting interested. Why?

True again. MGale is limited to a max of 2 open positions at a multiplier of 2 and min distance of 30 pip.

 

[CavaliereVerde]

That is the reason there are a lot of algo bloggers and algo teachers and very few algo traders.
Even if you develop a strategy with the most robust or sophisticated process there is no risk and endurance involved. They just talk.
The trader walks the talk.
The trader invests his time and his money in his strategy.

 

[VictoriaAlgo]

That is the reason there are a lot of algo bloggers and algo teachers and very few algo traders.
Even if you develop a strategy with the most robust or sophisticated process there is no risk and endurance involved. They just talk.
The trader walks the talk.
The trader invests his time and his money in his strategy.

I learnt it the hard way, and I am not there yet. My algos actually do very well; the failure has come from manual intervention!!! The other aspects of algotrading simply didn't occur to me, and I tell you, they were not mentioned in any of the books and copious research material I read. As far as I can remember, there in no mention of it in Robert Pardo's book, for example.

 

[counter_violent]

Okay. Let me kick off. What do you think of the performance of this back-test? It is a strategy I have been working on for nearly 12 months, the question is...can it operating like that into the future? Yes, the relative drawdown is very high indeed, but we can just use less risk and a larger account size.

Feedback please.

View attachment 294923

So what kind of strategy is it? bearing in mind that there is nothing new under the sun. Nothing that thousands of others have not considered before.

 

[CavaliereVerde]

Robert Pardo's book,

So you know the importance of robustness.
Usually people the show a backtest without context just want to sell you an EA or a course...
The process is much more important than the backtest chart.

 

[VictoriaAlgo]

So you know the importance of robustness.
Usually people the show a backtest without context just want to sell you an EA or a course...
The process is much more important than the backtest chart.

Yes. Only Pardo robust strategies make it to strategy evaluation.

 

[VictoriaAlgo]

So what kind of strategy is it? bearing in mind that there is nothing new under the sun. Nothing that thousands of others have not considered before.

I don't discuss rules. I think there are new things under the Sun if the field is very new. I was really just looking for the feedback like "1.38 profit factor" is possible. Just common sense checks.

I am up to speed with the theory and practice however I'm quite new to extreme growth models.

 

[CavaliereVerde]

There is nothing to invent but there are infinite new ways to combine old stuff.
What about crossvalidation?
Is this ruleset profitable also on other assets and timeframes?

 

[VictoriaAlgo]

There is nothing to invent but there are infinite new ways to combine old stuff.
What about crossvalidation?
Is this ruleset profitable also on other assets and timeframes?

I've done all the standard stuff, so multi market, asset, time frame, monte carlo, WFA, cross validation, OOS. It passes all of them but not stellar performance across the board and the very low timeframes generate too many trades.

 

[CavaliereVerde]

So my suggestion is to trade it on 1000 with one microlot and follow it on myfxbook.
If it works for one year you can scale up.

 

[LongVision]

I was really just looking for the feedback like "1.38 profit factor" is possible. Just common sense checks.

What is pip expectancy?
What is average holding period?
Did you considered slippage for order execution?
Did you considered swap & commission for backtest?

If pip expectancy is less then 5pips then this profit factor will be difficult to sustain/profit for live trading.