Adding separate criteria for longs and shorts automatically increases the strategy’s degrees of freedom and makes it excessively prone to curve-fitted solutions.
1 Like
- What is so dangerous about over-optimization?
We don’t want to generate trading strategies with absolutely astonishing results that will not be achievable going forward. - How long should a testing period be if you are serious about building a profitable trading strategy?
*Time frames should be greater than 30 minutes
*Ideally 9-11 years of data should be used for the process in order to ensure that a large amount of market conditions become available. - Why should you avoid asymmetric trading signals?
You end up matching up with previous market cycles. More probability of making a mistake or over optimizing
1 Like
-
What is so dangerous about over-optimization?
adjusting our strategy to past result might cause that in wont work in real live conditions. -
How long should a testing period be if you are serious about building a profitable trading strategy?
9-11 years is ideal period so wee make sure that all kinds of events happened and our strategy works most of the time. -
Why should you avoid asymmetric trading signals?
Adding separate criteria for longs and shorts automatically increases the strategy’s degrees of freedom and makes it excessively prone to curve-fitted solutions.
1 Like
-
Over optimization can lead to curve fitting, this is dangerous because its basically customizing your strategy to fit past events and future events are rarely (if ever) identical as the past.
-
Ideally ~10 years of data.
-
It will lead to curve fitting as asymmetric signals work for past data as there were other parameters put in play and we cannot be sure it will work in the future.
1 Like
- Overfitting train data will produce excellent results on that specific data, but much worse results on future market data. You can greatly overestimate your strategy, which will lead you to lose money.
- 9-11 years, averaging along such a long period will greatly reduce the risk of overfitting.
- Doing asymmetric trading signals might be in itself overfitting the strategy towards a specific side of the market cycle.
1 Like
- What is so dangerous about over-optimization?
Over - optimization often happens when strategies are over - fitted to a specific back-test data set.
One may not be able to replicate back-test results with realtime data.
- How long should a testing period be if you are serious about building a profitable trading strategy?
A data set of about 10 years is suitable to minimize over-fitting.
- Why should you avoid asymmetric trading signals?
Because asymmetric trading signals increase the degrees of freedom of a strategy and is a sign of over-fitting your system.
1 Like
- Over optimization lead to curve-fitting, which leads to fit specific past data.
- 9-11 years
- It works for past data but we can’t be sure for future data.
1 Like
-
What is so dangerous about over-optimization?
Because it can lead to curve fitting, trying to mimic past results. -
How long should a testing period be if you are serious about building a profitable trading strategy?
Testing period should be 9-11 Years -
Why should you avoid asymmetric trading signals?
Adding separate criteria for longs and shorts automatically increases the strategy’s degrees of freedom and makes it excessively prone to curve-fitted solutions.
1 Like
1 The more optimized the model, the more likely it is to fail outside the optimization time frame.
2 As long as possible. Using statistics; the larger the sample, the closer to reality. Thus a sample period above 1000 time units might be sufficient. As this is exponential, prolongations will be less and less valuable and more and more time consuming. I do not have the optimum number yet…
3 I do not fully agree with Daniel Fernandez. As the current inflationary system is biased to higher absolute values, increase is generally more probable than decrease. However pinpointing any specific asset some may behave differently. If you were to trade assets in real value, it would be the opposite, but in general we are not.
1 Like
1.Curve-Fitting
2 Optimizations should be carried out for long periods of time, ideally 9-11 years of data should be used for the process in order to ensure that a large amount of market conditions become available. If a simple strategy yields profitable results across a ten year period then the probability of curve fitting is greatly reduced as the system has limited degrees of freedom to artificially “fit” all those different market conditions.
3.Keep your system symmetric. One of the first ideas new traders have when they start analyzing system development and mathematical expectancy results is to have a separate criteria for entering and exiting short and long trades (for example using an indicator cross at 20 for long entries but 15 for shorts). Although it is true that under past data up and down trends might have developed differently in currencies this cannot be guaranteed to continue in the future as these differences rely on interest rate differentials or such similar macro economic variables that inevitably change through economic cycles. Adding separate criteria for longs and shorts automatically increases the strategy’s degrees of freedom and makes it excessively prone to curve-fitted solutions.
1 Like
-
If we over-optimize, we run the risk of curve-fitting our data and coming up with a trading strategy that only works with the data that we used backtest.
-
Testing should be done over long periods of time, with data segments of ideally 9-11 years.
-
Adding separate criteria for longs and shorts automatically increases the strategies’ degrees of freedom and makes it prone to curve-fitting.
1 Like
Over-optimization can lead to trying to force-fit a current situation from an earlier situation. Somebody trying to justify their hypothesis can keep trying to make a case for something without a valid basis
The testing period should be substantial enough to filter out random gyrations, and to show the long term results. One year is reasonable.
Assymetric trading signals can result in curve-fitting.
1 Like
-
Over-optimization leads to curve-fitting which makes a simulation not a realistic representation of future pricing (extrapolation).
-
Ideally, 9 to 11 years of historical data should be used.
-
Because having more criteria in the model automatically increases the strategy’s degrees of freedom and makes it excessively prone to curve-fitted solutions. More criteria, more likely an attempt to achieve a better approximation to the data set (curve fitting).
2 Likes
- What is so dangerous about over-optimization?
It is dangerous to adjust a strategy that fits perfectly to historical curve/ data. It will never happen exactly the same twice. The goal is to create a strategy that is acting similar than previous data and that will give good enough results in future.
- How long should a testing period be if you are serious about building a profitable trading strategy?
For a long period of time, ideally 9 to 11 years.
- Why should you avoid asymmetric trading signals?
Using asymmetric trading signals could easily lead to curve fitting strategies.
2 Likes
-
What is so dangerous about over-optimization?
It will start adopting algorithm of past events and lead to "Curve-Fitting" -
How long should a testing period be if you are serious about building a profitable trading strategy?
Min. of 5 years of data, but the ideal would be with 10 years of data so it can be tested to a wider rage of changes in the market. -
Why should you avoid asymmetric trading signals?
You open yourself to over-optimization and that leads to curve-fitting .
1 Like
1.) What is so dangerous about over-optimization?
A.) Overoptimization can be dangerous in its self-deception “Curve-fitting”. It can create a false sense of EA profitability that encourages the GREED within us to try our “Utopian” EA on a real account –and thereby subject our account to the flame of what in fact is a truly terrible system that can burn it the ground. (You start to beleive your own bullshit" and it will end in tears.
2.) How long should a testing period be if you are serious about building a profitable trading strategy?
A.) . Optimizations should be carried out for long periods of time, prefferably 9-11 years of data should be used for the process in order to ensure that a large amount of market conditions become available. If a simple strategy yields profitable results across a ten year period then the probability of curve fitting is greatly reduced, as the system has limited degrees of freedom to artificially “fit” all those different market conditions.
3.) Why should you avoid asymmetric trading signals?
A.) Asymmetrical risk is the concept of taking a risk that will produce a return that far surpasses the risk taken. This is a really important concept that can change your quality of life greatly and a risk worth taking if you can afford to loose, should it not go to plan. 
1 Like
1. What is so dangerous about over-optimization?
Over-optimization is dangerous because it trasforms the strategy in a curve-fitting of past data that will be never identic to the future data.
2. How long should a testing period be if you are serious about building a profitable trading strategy?
9 to 11 years of data
3. Why should you avoid asymmetric trading signals?
Because on long time frames some variables might change, compromising the trading system.
1 Like
-
Over-optimization can lead to curve-fitting which is when you adapt a bot to past data and it becomes useless with future data.
-
Testing periods should be more than 10 years.
-
Because it increses the strategy’s degrees of freedom which makes it more prone to curve-fitting.
1 Like
- What is so dangerous about over-optimization? curve fitting
- How long should a testing period be if you are serious about building a profitable trading strategy? At least 5 years
- Why should you avoid asymmetric trading signals? Because economic variables are likely to change in the future.
1 Like
-
It leads to curve-fitting, which means we can generate great results for the market in the past but those results cannot be maintained going forward.
-
Optimizations should be carried out for long periods of time, ideally 9-11 years.
-
Asymmetric trading relies on economic variables that might change through future economic cycles. That makes the system prone to curve-fitted solutions.
1 Like