How-to-leave-only-strategies-with-smooth-equity-curve

Check out the metrics in the Risk/Reward group of the Extended Scorecard (Backtest Results tab).

I use a combination of metrics: first I look to the SysQ scorecard (https://wealth-lab.com/extension/detail/finantic.SysQ), high SysQ tend to show stable profitability vs risk, then I look for high Sharpe in combination with a low Expected DD, a low DD and a high SysQ, and of course a large % of profitable trades. Then I run different backtests amongst the best cases and I look for satisfactory monthly results.

An idea: I will look for something like a Standard Deviation across the equity curve, Maybe the slope.

Wouldn't everyone love that? One way I found to smooth out the equity curve is to trade a handful of strategies, not just one. I run my MetaStrategy at the end of every trading day and get all my signals at once.

I recommend to read Laurens Bensdrop's book "Automated Stock Trading Systems: A Systematic Approach for Traders to Make Money in Bull, Bear and Sideways Markets".

He has laid out his trading approach exactly so that he can profit in all market situations. In his book he presents 6 stock trading approaches (Trend Momentum + Mean-Reverse Short/Long) designed to keep the equity curve as linear as possible.

Springroll, the table looks awesome. If you did a strategy using this system, how did it behave in recent years and what is the maximum drawdown?

You can check out his entire portfolios and their metrics at https://www.turingtrader.com/portfolios/. Bensdrop seems to be a very transparent guy...

Springroll, I'm aware of this site, didn't know it was his. Such openness, of course, is admirable. From what I see there, last years such more passive management systems show worse results. Even compared to the Glitch yield curve shown above.

Glitch, what's your maximum drawdown?

Upd: As for the topic question, haven't found a way in Evolver to filter strategies with unprofitable years at least.

Just a single system is too error-prone and will sometimes produce unprofitable years. The key is to combine several non-correlating systems to ensure constant profits.

To demonstrate this, two simple mean-reverse systems on leveraged ETFs. The short system had its great days until September 2022. After that in October 2022, the losses were covered by the long system, which can be seen in the combined view.

This is the way.

Still haven't found a way to cull the strategies that earn most of the time.

Such a score is especially important if this improvement is to be implemented
https://www.wealth-lab.com/Discussion/Strategy-Evolver-Detect-Convergence-Stop-Record-and-Restart-9212

For example a very simple ratio formula: (the number of profitable months or weeks/total)*100

QUOTE:
I run my MetaStrategy at the end of every trading day and get all my signals at once.

How do you send these signals to WealthSignals?

QUOTE:
I recommend to read Laurens Bensdrop's book "Automated Stock Trading Systems: A Systematic Approach for Traders to Make Money in Bull, Bear and Sideways Markets".

All systems from this book are now available as published strategies for WL8.
(see https://www.wealth-lab.com/Strategy/PublishedStrategies)

QUOTE:
For example a very simple ratio formula

You should consider the "Formula ScoreCard" available with the finantic.Scorecard extension.
(see https://www.wealth-lab.com/extension/detail/finantic.ScoreCard)

It implements many Performance Metrics which gauge "Smoothness of Equity Curve" like Lars' K-Ratio.