2.4) How to Extract Optimal Parameter Values in Optimizations using Statistical Power Analysis

This is episode 4 of a 4-part tutorial series that considers the process of extracting the parameters with the best possible edge from an optimization process using a technique called ‘Statistical Power Analysis’.

The quantitative study undertaken illustrates just how important ‘sample size’ – the number of uncorrelated trades - is to an effective optimization process.

It looks at how we need to dispel the myth that we can always be successful in finding the best parameters from an optimization.

Instead it considers how algo traders need to operate like statisticians and think in terms of probabilities of obtaining a certain level of statistical significance if they are truly going to succeed in extracting maximum potential from their trading strategies.

If algo traders are relying on small sample sizes in their optimizations, this study shows how they won’t be getting this maximum potential from their algos.

-----------------------
IMPORTANT REQUEST: Please please please.. if you find this content useful, please do consider liking and sharing it on YouTube, Twitter, Facebook, LinkedIn and whatever other social networks you have circles in.

Darwinex relies almost exclusively on organic growth, primarily through recommendation via informative content.

YouTube’s algorithms measure the quality of Darwinex content on the basis of:
- Reach
- Engagement
- and several other related variables

With seemingly small actions such as:
- Clicking the Like button
- Clicking the Subscribe button
- Clicking the Share button (on YouTube) and distributing our content
- etc

… YOU inform YouTube’s algorithms of your sentiment towards Darwinex, thereby directly helping Darwinex MASSIVELY in achieving organic growth.

Thank you very much for your kind consideration!
-----------------------

Risk disclosure:

** Fancy joining a vibrant community of algorithmic traders, quants and data scientists focused on financial hacking? Join the Darwinex Collective Slack Workspace:

#algorithmictrading #poweranalysis#alpha #trading #finance #backtesting #darwinex

[leonjbr]

Very good episode Martyn. I am impatient to see the next one.

[tdb]

These series on extracting optimal parameter values is truly valuable. One question I have though is how does one go about quantifying the percentage edge of one's trading system? The demonstrations given in these videos assume an edge of 10%. But how does one calculate that edge exactly?

[Shoz_]

What a unique level of great quality of usefull content
Really good job, and thank you for share ♥

[kevinalejandro]

But how You measure the edge of your trading system ??

[gadeichhorn]

Thanks for the information, very well presented. Can you share the math behind your tools? What will happen on a multi-parameter edge system, can you simulate all at the same time? how would you measure the power for this type of a system?

[Thetaquo]

Is it possible to know if your system has an actual edge before looking at what your sample size should be? Also, what kind of timeframe are you on that you're able to test 18, 000 trades, or are you testing it across various instruments?

[batatakhizou]

at the end of the video you said that you are going to show us how to get a sample size of this magnitude, well i can't find this video... please point me to the videon thank you in advance @Trade Like A Machine

[elmaestroruben]

If I have a equally weighted weekly rebalance strategy, do you think I can take all of its trades as uncorrelated? Or it will be better to treat all trades of a week rebalance as one unique trade?

[elmaestroruben]

And another question I have is if the sample size needed to have statistically significant results depends on the number of parameters of our strategy, or only on the number of trades?