System that does this? Does it exsist?

[TheSpreadBetter]

I would be interested in paying for the following system if it exists in some form. I need it to do the following:

- When I add a function or technical analysis it automatically adds it into the big picture or trend forecasting. By this I mean that if I add RSI and Stochastic it calculates their pointers as one and creates a new trend line based on the two technical functions.
So if I for example add 12 technical analysis models, it "adds them all up into one" and generates a new trend based on the selection(s).

- Also this system has to allow user input in the sense of a trend predicted by the user based on that the user(human) receives inputs in form of news that a program would not be able to interpret. This would allow for a greater flexibility. Such could be numbers predicted by the markets in the respective day for a respective sector or simply a user predefined formula that the user adds to the bigger picture. Now this would be a nice thing to have.

This will allow us to add any existing technical analysis and "merge" them together to get a broader picture of everything calculated as one. RSI and Stochastic is probably not the best selection, but trend prediction formula's such as the Tom Demark (look that up on the internet to see his technical analysis) and other that I do not know about myself.

The goal is of course to generate a trend line that has multiple pointers and hopefully can give us the bigger picture of how everything can move throughout the day. I dont know if this a good solution, but at least it would be something "new"

 

[TheBramble]

You've already got any trend information - in the price.

What you're looking for can't exist for the simple reason you couldn't 'add' for instance, an MA to a Stochastic value.

All indicators are derived from price (and volume and time) so why muddy the waters?

Of course, if you're hell bent on splicing inds together, go for it! But it wont help your trading directly.

 

[TheSpreadBetter]

Instead of looking at them separately going against each other sometimes it would be better as I see it to have them all into one and have a new line created from that.

 

[TheBramble]

Do you not think the 'going against each other' aspect is something you'd want to preserve?

The example I gave in my earlier post, combining an SMA and a Stochatics level. How would you like them combined? How would you like to plot the resultant vector? How would you interpret it?

I suspect if you address those questions, you'll find yourself in a strange place...

But in any event, as I don't think any such system can exist I don't believe I can help you further with your quest.

Good luck.

 

[TheSpreadBetter]

Look here: Tom DeMark - The best technical analysis? - ITalkCash.com - Forum for financial investments, HYIPs, Stocks, E-currency, Forex, Credit Cards, Hedge Funds and Offshore Companies
Apparently having 5 or six of those technical analysis would do the trick from my point of view.

 

[Trader333]

More indicators will not make you more profitable in my view. This has been tried on numerous occasions and to date it has not proven more successful than trading a far more simple approach.


Paul

 

[Hoggums]

More indicators will not make you more profitable in my view. This has been tried on numerous occasions and to date it has not proven more successful than trading a far more simple approach.


Paul

I can vouch for that - I tried it years ago - and it didn't work - you just end up with one of the indicators dominating the rest. A MA will move slowly - RSI faster - even weighting them you'll end up with the "trend" being dictated by the faster indicator.

 

[TheSpreadBetter]

I kind of agree that all that math can never count for the real underlying reasons for the movements. That just leaves me with a big questions. How are we ever supposed to know the trend lines on indices and commodities in most cases?
What are you guys doing?

 

[TheBramble]

Is the price going up, down or sideways for the timeframe in question? That's it.

 

[nine]

Is the price going up, down or sideways for the timeframe in question? That's it.

Yes.

Are the swings making new highs and lows? Or, if not, is it in a horizontal channel or a contracting or expanding triangle?

Or in market profile land are the value areas moving up, down, or not getting through?

Or is price sitting above the ma or below.

Or has price broken out of the channel its been in ... in which case is this just a pause for continuation in a slower channel or is it a reversal?

Lots of good ways to do this with few indicators involved. Fewer indicators = more chance that you will start to perceive what's happening and, sometimes, why. And that's good.

 

[dcraig1]

It sounds a bit like what you are after is something like a neural network predictor. You throw a bunch of inputs at it and out pops a prediction of something N bars in the futures. The something that you might predict could be a trading signal, moving average etc etc. There are ANN based products available such as TradingSolutions, Biocomp and others. Also ANN plugins for Tradestation, Excel etc.

Availability of products might be the good part, but in the hands of the less than expert they are likely to be worse than useless and many question their value even when applied by accomplished traders/system developers.

The first problem is choosing the imputs for the ANN and preprocessing them in appropriate fashion. It has been observed that if you can do that properly, you don't need an ANN anyway. There is probably some truth to this.

The second problem is that ANNs are magnificent curve fitters and can indeed produce some impressive results on their test set and even sometimes on walk forward testing. However markets change and with the change can come spectacular failures of the models which in effect have been seriously overfitted to the training data.

Another inherent problem is there is no way of knowing what an ANN is doing - you can't get inside it and determine it's "reasoning". You therefore cannot know the relative significance of each of the inputs to the model because the model is non-linear. This is an uncomfortable place to be if money depends on it.

It might be interesting stuff to experiment with, but as for developing trading systems, be extremely cautious and prepared to put in a lot of time and effort. There are probably easier tools to use.

I would be interested in paying for the following system if it exists in some form. I need it to do the following:

- When I add a function or technical analysis it automatically adds it into the big picture or trend forecasting. By this I mean that if I add RSI and Stochastic it calculates their pointers as one and creates a new trend line based on the two technical functions.
So if I for example add 12 technical analysis models, it "adds them all up into one" and generates a new trend based on the selection(s).

- Also this system has to allow user input in the sense of a trend predicted by the user based on that the user(human) receives inputs in form of news that a program would not be able to interpret. This would allow for a greater flexibility. Such could be numbers predicted by the markets in the respective day for a respective sector or simply a user predefined formula that the user adds to the bigger picture. Now this would be a nice thing to have.

This will allow us to add any existing technical analysis and "merge" them together to get a broader picture of everything calculated as one. RSI and Stochastic is probably not the best selection, but trend prediction formula's such as the Tom Demark (look that up on the internet to see his technical analysis) and other that I do not know about myself.

The goal is of course to generate a trend line that has multiple pointers and hopefully can give us the bigger picture of how everything can move throughout the day. I dont know if this a good solution, but at least it would be something "new"

 

[TheSpreadBetter]

What you are basically saying is that to many technical analysis is to much analysis and thus becomes obsolete as its prediction becomes to mathematical compared to the real movements. I can agree with that and I guess that you are right when saying it would be much fun trying it out and testing such, even though we can predict that failure would be big. Sticking to few things is better, and following the news beats any technical mambo jumbo.

So what are we supposed to do to get ANY sort of indication for the markets instead of the typical "gut" feeling?
I would say this place could offer some stuff: SignalWatch: Welcome! (but I havent tried out their intra day signals yet, dont know if its worth it).

I also just found this: Barcharts Morning Call

And I also found this: ClearStation : Welcome to ClearStation! (but I dont know if this is good for day trading/spread betting the indexes up and down). Basically what I am trying to say is that a day trader is ****ed as there seems to be NO indicators or "experts advice".

 

[nine]

A day trader is not ****d if he learns to trade instead of wasting his time searching for yet another disappointment (someone else's free or paid advice).

Get past this stage as fast as you can.

 

[TheSpreadBetter]

I disagree! Its like saying that you alone can predict better what is happening without any method out there. You can just sit in front of your screen and day trade without any inputs.

That means that if you need some inputs all you have to do is to find the "right" inputs and go for those. Its correct you develop your own inputs and feelings about it, but it has to be based on something, or you might as well go to the casino.

 

[nine]

But to base it on someone else's opinion is the signature of a losing newbie. People selling signals do so because they can't trade; if they could then they wouldn't need to sell signals: so why would anyone be foolish enough to take signals from a "can't trade".

To any newbie reading the thread: this is a stage you must get past if you are ever to succeed.

 

[0007]

But to base it on someone else's opinion is the signature of a losing newbie. People selling signals do so because they can't trade; if they could then they wouldn't need to sell signals: so why would anyone be foolish enough to take signals from a "can't trade".

To any newbie reading the thread: this is a stage you must get past if you are ever to succeed.

So True. If you never learn and understand the very basics, you will never be in a position to use, assess or reject indicators / opinions / advice. The dangerous period for newbies is when you get advice etc from all quarters - some of it will be good, some of it less so. You have to work through all of this: otherwise you end up like the moronic sat-nav follower who gets his juggernaught stuck up a one-way country lane.

If you don't understand black boxes then you'll always be at their mercy. This is not how you want to be in the markets. Believe me.