You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Walk Forward Optimization in Trading Algorithms
- Thread starter flipflopper
- Start date
Quote from thstart:
It is about the basis, the foundation on you build. "Falsify" is not exactly the case but if we use the term - to "falsify" or "prove" you have to have a right basis, a standard. You cannot prove or disprove on basis of false presumptions.
I do not make any sense of what you are saying. I will try to make it simpler for you. Backtesting is used to determine if a trading system is a failure. If it is not a failure based on past data, no conclusion can be made about the future.
" If the data looks reasonable under the null hypothesis, then no conclusion is made. In this case, the null hypothesis could be true, or it could still be false; the data gives insufficent evidence to make any conclusion."
http://en.wikipedia.org/wiki/Null_hypothesis
These are well-known and understood. Obviously, some science reporters just discovered them, including some ET members.
Quote from flipflopper:
So we all know markets cycle. Identifying what cycle (RTM or Trending) we are in and using appropriate trading strategies is one part of the puzzle.
The other piece to the trading puzzle is if you can you improve real time results by incorporating periodic walk forward optimization in your trading algorithm that tries to adjust parameters based on these identifiable market cycles.
In its simplest terms I want my trading strategies to automatically re-optimize my parameters periodically based on real time data.
Does anyone know of any books or studies done on this topic? This is something I have never heard discussed.
flipflopper:
I use walk forward and a Genetic Optimizer adding the Optimization period and the Test period as items to consider in the optimization process
Quote from intradaybill:
I do not make any sense of what you are saying. I will try to make it simpler for you. Backtesting is used to determine if a trading system is a failure. If it is not a failure based on past data, no conclusion can be made about the future.
" If the data looks reasonable under the null hypothesis, then no conclusion is made. In this case, the null hypothesis could be true, or it could still be false; the data gives insufficent evidence to make any conclusion."
http://en.wikipedia.org/wiki/Null_hypothesis
These are well-known and understood. Obviously, some science reporters just discovered them, including some ET members.
No - what you don't understand is that following the "official" rules you will go where the "climate change" is going - it was "official", right? Based on statistics, right? Let stop here - this is out of topic
You think you understand what "null" hypotesis means. Explain please how do interpret "normal" distribution, "confidence" intervals, etc. - e.g the cornerstone of statistics. Just a hint they are "idealized" representation of reality. By the way I have a MS degree and spend many years studying and applying math. Statistics has its place when applied right. Also wikipedia is not the highest source of knowledge.
Bottom line - you cannot "prove" or "disprove" nothing on false presumptions - e.g. no basis for "null" or "indefinite" hypothesis.
Unsurprisingly, Wikipedia's explanation is opaque and confusing. To clarify -- a p-value is the likelihood of the data given the null, nothing more.Quote from intradaybill:
" If the data looks [sic] reasonable under the null hypothesis... the data gives [sic] insufficent [sic] evidence to make any conclusion."
And it is hard to trust a poster who apparently thinks the word "data" is singular. The Wikipedia article appears to have been crowd-sourced from a gang of illiterates.
And you might look up the difference between e.g. and i.e.
Quote from Kevin Schmit:
Unsurprisingly, Wikipedia's explanation is opaque and confusing. To clarify -- a p-value is the likelihood of the data given the null, nothing more.
And it is hard to trust a poster who apparently thinks the word "data" is singular. The Wikipedia article appears to have been crowd-sourced from a gang of illiterates.
And you might look up the difference between e.g. and i.e.
Hi Kevin,
I see you are institutional trader (I think). Could you please tell me - do your institution is using statistics, where and how? It would be interesting to know its usage.
Thanks
Quote from thstart:
By the way I have a MS degree
I am sorry but "having an MS degree" nowadays is like "having a dog".
All the rich people I know were bad students, got low grades, never went to college but they know how to make money. That is all they know. They do not give a flying !@#% about distributions and statistics. Common sense is all it takes.
Quote from intradaybill:
I am sorry but "having an MS degree" nowadays is like "having a dog".
All the rich people I know were bad students, got low grades, never went to college but they know how to make money. That is all they know. They do not give a flying !@#% about distributions and statistics. Common sense is all it takes.
good luck and listen to "fairy tales" - no you don't personally know rich people - you are reading about them. Having MS makes sense after you spend time learning. I learned - are you?
I am the Cycle Man. It is what I have been working on and trading with all my life. I had to develop it on my own because of the spartan sets of information on this topic.
I built my own cycle indicators and cycle trading systems. All my indicators and strategies run off of the same âcycle engineâ to stay in tune with each other.
The answer to your walk forward testing question is yes strategies can be written to trade in market cycles.
I built my own cycle indicators and cycle trading systems. All my indicators and strategies run off of the same âcycle engineâ to stay in tune with each other.
The answer to your walk forward testing question is yes strategies can be written to trade in market cycles.
Quote from flipflopper:
So we all know markets cycle. Identifying what cycle (RTM or Trending) we are in and using appropriate trading strategies is one part of the puzzle.
The other piece to the trading puzzle is if you can you improve real time results by incorporating periodic walk forward optimization in your trading algorithm that tries to adjust parameters based on these identifiable market cycles.
In its simplest terms I want my trading strategies to automatically re-optimize my parameters periodically based on real time data.
Does anyone know of any books or studies done on this topic? This is something I have never heard discussed.