Walk Forward Optimization in Trading Algorithms

F

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.
 
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.
More...

We are following this method:
1) long term cycles
2) short term trend
3) tape reading

1) is based on natural cycles with a common denominator 365 calendar days reduced to actual trading days and natural ratios.
2) Pivot Points and support/resistance based on them in following time spans - 1 day, 1 week, 1 month
3) real time - 1st hour of trading

1 and 2 and 3 give you enough information to make decision what to do.

You can drop me a PM if you wish - I can share more.
 
Quote from thstart:

We are following this method:
1) long term cycles
2) short term trend
3) tape reading

1) is based on natural cycles with a common denominator 365 calendar days reduced to actual trading days and natural ratios.
2) Pivot Points and support/resistance based on them in following time spans - 1 day, 1 week, 1 month
3) real time - 1st hour of trading

1 and 2 and 3 give you enough information to make decision what to do.

You can drop me a PM if you wish - I can share more.
More...

I think you misunderstood my post. Not interested in what anyone else is doing.

I want to learn about how to incorporate self optimizing parameters into my own trading strategies.
 
Quote from flipflopper:

I think you misunderstood my post. Not interested in what anyone else is doing.

I want to learn about how to incorporate self optimizing parameters into my own trading strategies.
More...

OK I wanted to help. Inevitably it is always from my "own experience".
 
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.
More...

Here is some interesting info about some things you can stay away:
http://www.sciencenews.org/view/feature/id/57091/title/Odds_Are,_Its_Wrong

Basically: "The “scientific method” of testing hypotheses by statistical analysis stands on a flimsy foundation. "
 
Quote from flipflopper:

Very interesting. I like it!!
More...

Basically this leaves a lot of books describing "back-testng" irrelevant. Statistical methods as they are used are "illusion" for the matter we deal with - the stock market.

One corner stone of statistics methods is a "normal" or other well formed data distribution. Even if you have an "unlimited" set of data you cannot see a "normal" or other predictable "textbook" distribution. It follows that the formulas for "confidence intervals", mean distribution, etc. are meaningless. The stock market data are not only with "non-normal" distribution but if there is a meaning in the term "distribution" it changes dynamically.

All the talk about "algorithmic" trading, the way how it is popularized in the press has nothing to do with the actual mechanized trading as performed from the big players.
 
Science reporters do not understand science. That article is pure nonsense.

Backtesting, or testing in general, is used to falsify hypotheses/theories not to prove them. This is how science works, by falsification. No hypothesis can be proven true. Everyone knows that. Only science reporters do not know it. They should be banned from writting articles about science.
 
You must log in or register to reply here.
Back
Top