Topic: system quality number
Curious what system quality number means? I assume it means the higher the number the 'better'/'cleaner' the strategy, but I am not sure. Thoughts?
- Popov
- Lead Developer
- Offline
Re: system quality number
The author of System Quality Number is Van Tharp: http://www.vantharp.com/tharp-concepts/sqn.asp
Standard interpretation of the values:
Score: 1.6 – 1.9 Below average, but trade-able
Score: 2.0 – 2.4 Average
Score: 2.5 – 2.9 Good
Score: 3.0 – 5.0 Excellent
Score: 5.1 – 6.9 Superb
Score: 7.0 - Keep this up, and you may have the Holy Grail.
Or the higher - the better.
You can use SQN in repository to sort the records. It does a good job.
3 2016-08-08 15:51:20 (edited by Freesby 2016-08-08 15:51:44)
Re: system quality number
Damn, I do not get to 7.0, I will stop at 6.3.
I can also become a millionaire?
- Irmantas
- Licensed Member
- Offline
Re: system quality number
HI,
there is very simple few rules to follow to avoid curve fitted systems, good example what is curve fitted is posted in last freesby post 
And these all high statistic like high SQN is meaningless until you follow something like this:
1. Less variables is better. Less degree off freedom will not curve data so much. 2-5 variables per systems is max (at least for me and some book authors of system creation)
2. Try to have at least 50 trades for each variable to have some statistical meaning.
3. System needs to show similar good results on unseen data. Use time horizon tool to walk forward. Acceptable Walk Forward efficiency for me is 0.4, it means if system shows 0.4 of optimized profits on unseen data.
Hope it helps to save tons of time of learning that best back test is not the best future money maker....