the reason indicators don't work isn't because they're lagging

[1a2b3cppp]

Most traders don't get this. They say oh adx or whatever doesn't work is because it's lagging.

It doesn't work because it's not predictive.

Because an indicator does something when a trend happens or reverses or whatever doesn't mean that when the indicator does that, a trend is happening. Of course it's lagging, it does calculate based on what price did. But it has no predictive value so it's worthless.

This shocks me to say, but I think volume may actually be a predictive indicator. In some cases.

 

[fan27]

Indicators can be used to help measure a pull back in a trend or other price characteristics. I find that indicators can be useful filters. I agree though that trying to single out an indicator that has predictive qualities is the wrong approach. Better to start out with a strategy hypothesis, find the correct data, indicators, etc to test that hypothesis and then proceed accordingly.

 

[comagnum]

Indicators are derived from price & volume - that's why they lag.

Only momentum and volume precedes price.

* Volume on equities that is -I don't think this applies to futures *

 

[speedo]

I use an oscillator to assess momentum dynamics across my three time frames to filter and qualify my entries. This works for me, if they don't work for you or you don't know how to use then don't use them. Categorical statements like...they can't predict the future or are lagging make one sound foolish. ALL data is lagging or there would be nothing to print and no one since Nostradamus has been able to predict the future (and I have my doubts about him). We are in the field of potential and of managed risk, not fortune telling.

 

[Xela]

It doesn't work because it's not predictive.

it has no predictive value so it's worthless.

As is so with so many other threads in this forum, most of the argument that will promptly ensue, in the thread, will be directly attributable to the vagueness and ill-definition of the word "predictive".

My own take on this ever-thorny subject: "predictiveness" is a probability function.

No indicator is "predictive" in an individual case, but it's possible accurately to make some much more precisely "predictive" statements such as "with this instrument, on this time-frame, when this line crosses that line and so-and-so happens at the same time, then if you open a trade during RTH with a target of twice the ATR and a stop-loss of twice the ATR, it wins 55% of the time, with an error factor of 3% and confidence index of 97%".

Collective instead of individual, and all parameters defined. Such statements can be independently proved or disproved. They "mean something".

If that's "predictive", then indicators can be predictive.

If it isn't, then they can't.

But if you just bandy about the word "predictive" with no explanation at all of what you mean by it (again), all that can really follow is floundering discussion and unproductive talking at cross-purposes.

And - rightly or wrongly - that's my "prediction" for the general direction this conversation.

But hey ... that's how forum conversations go all the time.

Personally, I trade without indicators and don't like them, but I also have colleagues who make their full-time livings by depending on them, and however much you might imagine that the indicators they use "have no predictive value", you can't detract from that observation, because it's objective, factual and verifiable (otherwise they wouldn't be allowed to).

 

[panzerman]

I think Xela has the best explanation of what prediction means in regard to financial indicators.
John Ehlers has an indictor that he labels as predictive, and there is a method from digital signals processing called linear prediction, which extrapolates data into the future. However, I have never seen any back testing data to show if either methods has positive expectancy.

 

[ironchef]

And I have seen folks used Kalman filter, modern estimation and optimum control, Fourier transform, phase conjugation, phase lock loop, digital signal processing, digital filtering, small signal theory.... to predict. None worked as far as I know. Finance is harder than engineering and science.

 

[lawrence-lugar]

I have seen folks used Kalman filter, modern estimation and optimum control, Fourier transform, phase conjugation, phase lock loop, digital signal processing, digital filtering, small signal theory.... to predict.
None worked as far as I know.

Finance is harder than engineering and science.

Finance/trading isn't necessarily harder than engineering and science...it's just that it's highly ambiguous, and constantly requires a dynamic, open mind to understand its movements and gyrations.

Not many people can keep up with the market's somewhat weird and erratic temperament.
Every day in the market feels like a dance...between Play-Doh, Jell-O, and a Chess game.

All of those labels above for market theories sound and seem and look useless and ridiculous.
Studying and analyzing the market is not like digging up and studying dinosaur fossils or ocean currents or doggy behavior.

Basically the market is part art, part science...and that's really much more profound than what it may seem.
The market is way more complex...than the discovery of splitting atoms and creating a nuclear bomb, in a somewhat vague sense.

 

[IronFist]

Indicators are derived from price & volume - that's why they lag.

They don't lag. They show information how they should. Cci for example doesn't lag, it shows how far trades are from a moving average, and this information is correct, albeit not predictive in any way that is advantageous to traders as all known signals are not profitable (ie reaching a certain distance away doesn't mean it's going to reverse and moving from under to over doesn't mean it's going to extend in that direction).

 

[tomorton]

Finance is harder than engineering and science.

Certainly harder for engineers and scientists. All the ways they learned of being good engineers and good scientists are opposite to being a good trader.