Ron wrote:
> One belief that helps to achieve the 70% barrier is that the market is not random, but > has definable patterns that occur with some degree of regularity...Another helpful
> belief for the 70% achievement is that the definable patterns are irregular on a single > chart multi-year daily or weekly chart.
Ron,
I think anyone who builds a trading system needs to believe the markets are non-random. The challenge has to be that they seem so nearly random, and anytime a signal seems to be measured there are seemingly random shifts in its character.
I think there is something to the second point, about scale. In my own work, I sort parameters into two groups which I call "walking parameters" and "standing parameters". Standing parameters are defined by two characteristics,
1) They are optimizable and remain stable for long periods of time (at least whole market cycles).
2) Changes to them create proportionate changes to the performance of the system.
Walking parameters have the opposite characteristics. The optimal values seem to change suddenly and at random intervals, and a change to the value of a walking parameter often/usually has a disproportionate impact on the performance of the system. The best I can do to "optimize" walking parameters is narrow down a range of values where the variation in performance constitutes an acceptable risk/reward.
I bring this up because of what you say above about irregular features. In my experience walking parameters are typically those which are sensitive to scale-irregular features. A classic example might be a lookback number used to collect data for analysis (not for smoothing). I often see the performance of a system vary disproportionately with small changes in such a lookback and (correspondingly) see the "optimal" lookback value change in what seems to be random and unpredictable ways.
On examination, what causes this is that the system is sensitive to patterns that are changing in scale from occurrence to occurrence, or where the elements of the pattern are spaced differently from occurrence to occurrence even if the pattern as a whole reoccurs at the same scale.
If some points in the pattern fall outside the data collection boundaries, the system fails to recognize it, or if the data collection boundaries are too large it is as if the system has stepped back too far from the pattern, filled with noise, and will fail to see it. There's a "goldilocks" number for the amount of data collected relative to the size of the pattern, which varies randomly as the scale of the pattern varies for different occurrences. In summary, I see problems in that patterns 1) change their internal scale from occurrence to occurrence; and 2) as a whole do not repeat necessarily on a constant scale.
Problem (1) means pattern recognition needs to be sensitive to high-level rather than low-level features; and problem (2) means that a walking parameter set to a value which is sufficient to read a pattern on one occurrence may be too short on its next occurrence and thrown off by incomplete information or too large and therefore thrown off by noise.
This is a very difficult problem. I think one of the more difficult in measuring markets. My approach is to use Bayesian methods for reasoning under uncertainty and with incomplete information, and also to run measurements at multiple scales simultaneously. I'd love to hear if others see this problem in their own work and how they tackle it.
---In amibroker@yahoogroups.com, <ronj@...> wrote :
Guys In the discussions here many have expressed that they have difficulty breaking above the 50% profitable trades barrier. Along the way they have also mentioned some of their beliefs that probably contribute to that. One belief that helps to achieve the 70% barrier is that the market is not random, but has definable patterns that occur with some degree of regularity. A simple example, that many are familiar with is the head and shoulders pattern. Another helpful belief for the 70% achievement is that the definable patterns are irregular on a single chart multiyear daily or weekly chart. Perhaps others could share the underlying beliefs they find supportive to create powerful systems.
Regards Ron RonJaenisch.com "Success is almost totally dependent upon drive and persistence. The extra energy required to make another effort or try another approach is the secret of winning." — Denis Waitley: American motivational speaker |
All of my emails are for Educational or Entertainment purposes only. If you need professional advice please contact an appropriate professional, none is given via my emails.
Posted by: rosenberggregg@yahoo.com
__._,_.___
Posted by: rosenberggregg@yahoo.com
| Reply via web post | • | Reply to sender | • | Reply to group | • | Start a New Topic | • | Messages in this topic (2) |
**** IMPORTANT PLEASE READ ****
This group is for the discussion between users only.
This is *NOT* technical support channel.
TO GET TECHNICAL SUPPORT send an e-mail directly to
SUPPORT {at} amibroker.com
TO SUBMIT SUGGESTIONS please use FEEDBACK CENTER at
http://www.amibroker.com/feedback/
(submissions sent via other channels won't be considered)
For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/
This group is for the discussion between users only.
This is *NOT* technical support channel.
TO GET TECHNICAL SUPPORT send an e-mail directly to
SUPPORT {at} amibroker.com
TO SUBMIT SUGGESTIONS please use FEEDBACK CENTER at
http://www.amibroker.com/feedback/
(submissions sent via other channels won't be considered)
For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/
.
__,_._,___
EmoticonEmoticon