the future validity of indicator led systems

[jezza888]

We all know that money management and risk are the key to overall success (or at least staying afloat).

When devising and using an indicator based system we backtest it over 6 weeks/months/years... But, how can we be sure it will carry on in the same fashion. however long the backtest of the system is,whats to say the equation and rules will keep working in the same waY?

Or, if we find a system, can we really guarantee our future based on income on the averaged results of the 5 years backtested. If one was to lay themselves on the line and had MM to compensate for numerous losses etc, could one honestly say, with the data they have produced... I know in 6 months I will be still earning an average... per week, as i will in 1 year, 18 months and so on....

 

[neil]

Crystal Balls

We all know that money management and risk are the key to overall success (or at least staying afloat).

When devising and using an indicator based system we backtest it over 6 weeks/months/years... But, how can we be sure it will carry on in the same fashion. however long the backtest of the system is,whats to say the equation and rules will keep working in the same waY?

Or, if we find a system, can we really guarantee our future based on income on the averaged results of the 5 years backtested. If one was to lay themselves on the line and had MM to compensate for numerous losses etc, could one honestly say, with the data they have produced... I know in 6 months I will be still earning an average... per week, as i will in 1 year, 18 months and so on....

You cannot expect a "system" of any sort to be profitable in the future anymore than you can see into the future. There is no way of predicting. All you have are "probabilities." Backtesting is hindsight. Look at all the systems discussed on this and other boards now neglected and lying in dark corners whilst people like you rush around in search of the new trading crystal ball. Sorry but you will never know what tomorrow brings. However, money management ,and the avoidence of systems, might give you an edge provided you also remember that understanding yourself ( emotional control etc.) may count for more.

 

[jezza888]

are you saying that the only way you can definately earn money and have an edge is by, not using systems, but P+V etc and learning the markets?

 

[neil]

Research

are you saying that the only way you can definately earn money and have an edge is by, not using systems, but P+V etc and learning the markets?

To avoid repeating advice for new traders shown on this board I suggest you search the threads under the tabs above ie. "K.Lab etc. It's all free and useful

 

[dbphoenix]

are you saying that the only way you can definately earn money and have an edge is by, not using systems, but P+V etc and learning the markets?

Be wary of any phrase or sentence that contains the words "only way".

But learning the markets, if one is going to trade the markets, would seem to be a reasonable and logical avenue to explore.

--Db

 

[jezza888]

Does make sense really!

 

[mr_cassandra]

Building & testing a mechanical system

First and foremost, imo, any system or concept needs to be first paper traded in real time, then traded with real money. From personal experience I have found that to be light years different from the sterile world of numbers on a spreadsheet.
How will it do going forward in real life depends a lot on how its designed. For instance if it was based on buying when vix hits 50, it would go off a cliff as the vix can and does change ranges over the years.
If it was designed to trigger when macd of mdy hits 2.5 again it could fail in the future because as stock price rises, so does macd range.
A lot of m program is based on 200 day lookbacks that decide 'position in current year range' rather than look for fixed numbers.
On the flip side, negative mclellan summations will always mean a bad time to be long and vice versa.
IN addfition, you do need to test how your system works in bull mkkts, bear mkts and trading ranges. In the USA we had all of those since 1995, when my backtesting starts.
Last but not least, one consistent set of decision code(s) must be uniformly applkied to all the years in your test thru a program or spreadsheet. Its impossible to just 'eyeball ' stock charts and say well I woulda.....

I have a website which answers all the general questions I get about my work at www.angelfire.com/ma4/mr_cassandra/MainPage.html

Another good site is www.vtoreport.com then see RSI link. They have results since 1997 on a simple effective free system that I suspect beats the majority of online investors.

Regards, Steve

We all know that money management and risk are the key to overall success (or at least staying afloat).

When devising and using an indicator based system we backtest it over 6 weeks/months/years... But, how can we be sure it will carry on in the same fashion. however long the backtest of the system is,whats to say the equation and rules will keep working in the same waY?

Or, if we find a system, can we really guarantee our future based on income on the averaged results of the 5 years backtested. If one was to lay themselves on the line and had MM to compensate for numerous losses etc, could one honestly say, with the data they have produced... I know in 6 months I will be still earning an average... per week, as i will in 1 year, 18 months and so on....

 

[mr_cassandra]

One addendum

In the five years and 100's of hours I've put in backloading data and doing what if testing, I've been entreated to many armchair system quarterbacks quoting various euphemisms like

You can't time the market
If it worked everyone would do it and it wouldn't work
must be over optimized
must be underoptimized
must be curve fit
wont work in a bear mkt
won't work in a bull mkt
won't work in real time
then they switch to saying not enough real time.

As you address each one they shift to a different one.

Bottom line is to ask these folks what work of their own they have done, have they checked out a free system like vto thats already working, etc.
I had a long go round once with a man brilliant in his own field but a nudge in system design. He went on an enormous tirade because my vix numbers were off from his by .001 and this was a giant concern. I went and spent 4 hours changing them in order to show him that giant tirade was all for nothing, it changed nothing in my program results. He couldn't see the forest for the trees.

Most of the time you are dealing with someone who read an article or heard a soundbite about 'you can't time the market' rather than someone who has done research of their own.

In sum, if they havent done the work themselves and are taking someone elses word for it, don't waste much of your time about their learned sounding opinion.
[


QUOTE=jezza888]We all know that money management and risk are the key to overall success (or at least staying afloat).

When devising and using an indicator based system we backtest it over 6 weeks/months/years... But, how can we be sure it will carry on in the same fashion. however long the backtest of the system is,whats to say the equation and rules will keep working in the same waY?

Or, if we find a system, can we really guarantee our future based on income on the averaged results of the 5 years backtested. If one was to lay themselves on the line and had MM to compensate for numerous losses etc, could one honestly say, with the data they have produced... I know in 6 months I will be still earning an average... per week, as i will in 1 year, 18 months and so on....[/QUOTE]

 

[commanderco]

We all know that money management and risk are the key to overall success (or at least staying afloat).

When devising and using an indicator based system we backtest it over 6 weeks/months/years... But, how can we be sure it will carry on in the same fashion. however long the backtest of the system is,whats to say the equation and rules will keep working in the same waY?

Or, if we find a system, can we really guarantee our future based on income on the averaged results of the 5 years backtested. If one was to lay themselves on the line and had MM to compensate for numerous losses etc, could one honestly say, with the data they have produced... I know in 6 months I will be still earning an average... per week, as i will in 1 year, 18 months and so on....

Forget guarantees in life Jezza888 and stay with realities, it is much safer.
If you want a foolproof system then try this one on highly liquid futures.
Firstly, look at the direction of the market in a higher time frame ie currencies in a 60 minute frame. Then switch to a 5 minute frame and look for the opposite over extended direction.
Trade into that direction, which will put you contra with your 5 min chart and in harmony with your 60 min chart. You can further add to this by ensuring that your 15 min chart agrees with 60 min.
Keep your accuracy high ( 75%+) and your brokerage low and bank the points.
Let your faith in yourself increase the number of contracts traded and I assure you that you will be happy with the outcome.

 

[mr_cassandra]

I would disagree

There are many all successful ways to trade and hard-working pracxticioners of each can make some good money. Its sad, however, when any one tries to say 'that' way is the only way. very untrue.

Imo what 'is' true is that I can't do what Warren Buffett does and vice-versa. That does not negate his wealth or the results from my own program work.

Anytime you run into a 'harrumph that won't work' poster, be sure and ask him what he bases that on and what research he has done. You'll quickly find that narrows the field about 90%.

are you saying that the only way you can definately earn money and have an edge is by, not using systems, but P+V etc and learning the markets?

 

[jezza888]

Cheers for the input all,

Its now been just over 2 months since I decided to live trade the system I'd tested. Despite the mixed reactions this board brings on how to be successful I halved my account and proceeded with caution.

Thankfully, with decent MM and risk adversity it has done me very well and I am now more than confident that I can earn my income through using this 'system'. I did start posting live on the 'live cable trading' thread but it was a hindrance to my trading and the only person I need prove anything to is myself. After Monday’s lackluster day, the Tuesday was my weekly target from first thing so I no longer posted.

To say that a system will not work is wrong in my opinion. To buy a system for a large amount of money without any experience is not wise but learning the market I was trading and forward testing my plan worked well for ME. Not necessarily for others though.

Any system/plan/method prepared for, be it through price, indicator, breakout etc has only a degree of probability and none has anymore than any other as no one knows where its going next for certain, however you gauge it.

I now appreciate there are no guarantees in trading and now understand that its purely probability and I can confidently trade my plan happily knowing it will be a positive overall outcome.

As for the 'harrumph that won't work' poster... Well… who cares, I know it works now and I’m more than happy! Each to their own!

J

 

[mr_cassandra]

Glad to hear how you're doing

Anyone who says they didn't take some hard lumps learning the game is usually lying. You are doing the best thing, work on your own, learn from your mistakes, keep on going.

In the 5 years and 700+ hours I've spent on the program I now use to trade I've been entreated to a spectacular number of armchair system quarterbacks all knowingly intoning various cliched euphemisms like 'must be curve-fit' or won't work in a bear market or some other blah-blah.

When pressed most of these people turn out to have researched not much if at all, and can't even explain why vto's 5 day rsi system has been making money since 1997.

I ask them a few questions and usually the answers (or indeed the lack therefo) tells me what level of system student I'm talking to.

Good luck with your work.

Cheers for the input all,

Its now been just over 2 months since I decided to live trade the system I'd tested. Despite the mixed reactions this board brings on how to be successful I halved my account and proceeded with caution.

Thankfully, with decent MM and risk adversity it has done me very well and I am now more than confident that I can earn my income through using this 'system'. I did start posting live on the 'live cable trading' thread but it was a hindrance to my trading and the only person I need prove anything to is myself. After Monday’s lackluster day, the Tuesday was my weekly target from first thing so I no longer posted.

To say that a system will not work is wrong in my opinion. To buy a system for a large amount of money without any experience is not wise but learning the market I was trading and forward testing my plan worked well for ME. Not necessarily for others though.

Any system/plan/method prepared for, be it through price, indicator, breakout etc has only a degree of probability and none has anymore than any other as no one knows where its going next for certain, however you gauge it.

I now appreciate there are no guarantees in trading and now understand that its purely probability and I can confidently trade my plan happily knowing it will be a positive overall outcome.

As for the 'harrumph that won't work' poster... Well… who cares, I know it works now and I’m more than happy! Each to their own!

J

 

[jezza888]

thanks mr_cassandra.

good trading to you too.

J

 

[mr_cassandra]

I have spent 7 years working with over a dozen indicators and have heard this concern before. I think its a valid concern as long as its put in a balanced perspective. One example. I had a poster mention that vix makes major range changes over time so it would not be useful. This was an unproductive half-truth. Ie you can't use fixed number targets on vix because this years low may be next years high, or vice-versa. However this doesn't mean drop and slink away, it means find an intelleigent way to use it like with 200 day lookbacks from which you can then determine relative posotion in recent range. I have been using it this way successfully right thru another signal I had 2 weeks ago.
There is also a prevalent meme that says: if any indicator is really reliable, everyone will use it and it will become self-cancelling. Gag me. The human condition will always prevent that. Some people want fast trades, others slow. Other folks won't stay with it, others don't have the capital. There is a list of more reasons at my yahoo group.
Imo, there are many indicators which will always cycle the same as long as there are stocks. Adv-decl, pcr etc.