if your strategy can be adjusted according to external conditions any time, then it will never fail....make 5%- 30% easily

per day

Vlastimil, not per day. Forget that. That strategy could not do that, not even close. But what was presented yes.

If they find a strategy that doesn't fail, they don't publish the paper :D

Pablo, even Mr. Buffett some 30 years ago revealed his trading methods to anyone who would listen. And then concluded with: it does not matter, no one is going to listen anyway!

Day strategy's is ok but hardly to be for long run

Hi Aurel, true. If I wanted a daily trading systems, I would program something different, at least differently. The strategy presented made over 58,000 trades over the testing period. That necessitated a lot of entries and exits on an almost daily basis.

Note that the underlying trading principles are relatively simple. You have a channel (which could be designed using about anything) where its inside is a non-trading zone, and you roll over your ongoing inventory on rising prices with the intention of not always selling it all. This way you accumulate shares for the long term while also taking profits over the short, mid and long-term.

A simulation gives if a trading strategy could have been possible, at least over its past. It says if the methodology you want to use make profitable sense.

The real question is: will stock prices continue to fluctuate in the future? If you answer yes as I do, then you too can draw a channel.

Hi Guy,your indicator is bases on DSP ?

Great article. Thanks. Some may find an article I write 16 years ago of interest. It's titled "Trading System Diversity"​.

@Ludo, if DSP is for digital signal processing, then, not in that program. What was used however were objective functions. I wanted to guide and control the long term outcome, even under uncertainty, even if partially.

You have a random amplitude and random phase shifting signal (the price). But, due to its short term randomness characteristics very hard to predict. So, you don't even try.

However, you note that a long term portfolio has a sigmoid function for its holding time. Short term, the probability of a portfolio level win might be close to 0.50, but as you increase the time horizon, it will approach asymptotically one, following this sigmoid function. And you can easily take advantage of that. It might be the only thing you know that will happen!

@John, thanks. Why not provide a link to your article. I would be interested in reading it.

Most trading argos go through periods of gains and losses. If the loss period is substantial and the developer has not established risk containment solutions, then the algo may experience substantial drawdown. Are there ways to address this.. yes. Are there ways to turn losses into profits.. yes. Do most algo designers understand these concepts.. I doubt it.

I would say for algo trading to win over broker is much easier than market

@Brad, all portfolios have ups and downs. You can always reduce volatility, but it has for side effect to reduce overall performance. However, for a long term player, short term fluctuations in rising prices might not be seen as a deterrent, but as an opportunity to increase profits just like in the trading strategy presented. And, as said, this can be done in many many ways.

@Kenneth, you design an EOD automated trading strategy. It will execute its job in just a few minutes every day. It will not be bothered by whatever intention a broker might have, or by anybody else for that matter. You are just executing a program.

What you want to know is will the future resemble the past. Will stock prices fluctuate in a similar manner going forward? The simulation only showed that over a 20-year trading interval it succeeded in outperforming the market, brokers, and most everyone else.

Anyone could design something similar to DEVX8 and outperform averages.

Degrading don't write them. Tip humans don,t change I will say staying dynamic is smart. Math tied to emotion repeats endlessly.