Best indicators ?????????????

Quote from Humpy:

Which indicators do YOU recommend and why ?

I have been using Macd, RSI, Mov averages etc. for years with mixed results.

Perhaps you favour a mixture of 2 or more indicators to compliment each other ??
More...

Im using ADX indicator..Because it measures the strength of the trend and it indicates high readings as a strong trend and low readings as a weak trend....

tower light uk
 
Quote from Humpy:


Perhaps you favour a mixture of 2 or more indicators to compliment each other ??
More...

I see there are some that share the notion that indicators of indicators is a no-win decreasing returns but

try dropping the stochastic oscillator indicator on the Macd indicator. Surprisingly it is quicker than the Macd !

Perhaps there are some others too ?
 
Unless you have strict rules of how your indicators will be used, you are only making your job as a trader more difficult. I'm developing a second system that uses 4 indicators. Each one has 1 purpose, and one purpose only, that is applied to every trade.

stochs with 34/10/10 settings
macd
5 atr
boll bands

The reason I'm developing a second system is so that I have one thats better suited for day trading , as my first one requires too much analysis to make quick decisions.

The guys that say they are price action only, I don't think are telling you the whole story. You think they read, and analyise every candle?..no. MOST of them, at least the ones I know, trade price using fibs. A lot of guys here on et hate them with a passion, mostly because they dont understand them. But in FX circles, fibs are followed like a damn religion, and are also very accurate. I love fibs, but my problem has always been timing...hence the indicators.

Here is my 2 cents of advise to indicator traders:

1. Like I said above, every indicator should have a specific purpose for being on your charts that will be used on EVERY trade..what do I mean by this?. My stochastics is used to give me a buy or sell signal, however I use macd to confirm wether or not I should take it...if macd is red and stochs say "buy" i ignore. The boll bands are only there for me to calculate stop losses. Stops are at the low or high of the bands ( i dont use the 20sma), and the ATR is used to calculate targets.

So you see, I cover all bases: entry, stops, and targets. Some thing needed on EVERY trade.

2. Have an indicator that confirms the signal indicator. Having a buy/sell signal isn't enough. A confirmation indicator is needed to help ignore the bad trades. DONT take buy/sell signals off your confirmation indicator unless this is part of your plan!.

3. Play with the values of the indicators. I find that a 34/10/10 stochastics works great for ME, but try other values to see what you like, and works for your system.

I'll be nice and share my system with anyone interested, but its still very new. So do your own backtesting.

KON
 
Quote from konviction:

Unless you have strict rules of how your indicators will be used, you are only making your job as a trader more difficult. I'm developing a second system that uses 4 indicators. Each one has 1 purpose, and one purpose only, that is applied to every trade.

stochs with 34/10/10 settings
macd
5 atr
boll bands

The reason I'm developing a second system is so that I have one thats better suited for day trading , as my first one requires too much analysis to make quick decisions.

The guys that say they are price action only, I don't think are telling you the whole story. You think they read, and analyise every candle?..no. MOST of them, at least the ones I know, trade price using fibs. A lot of guys here on et hate them with a passion, mostly because they dont understand them. But in FX circles, fibs are followed like a damn religion, and are also very accurate. I love fibs, but my problem has always been timing...hence the indicators.

Here is my 2 cents of advise to indicator traders:

1. Like I said above, every indicator should have a specific purpose for being on your charts that will be used on EVERY trade..what do I mean by this?. My stochastics is used to give me a buy or sell signal, however I use macd to confirm wether or not I should take it...if macd is red and stochs say "buy" i ignore. The boll bands are only there for me to calculate stop losses. Stops are at the low or high of the bands ( i dont use the 20sma), and the ATR is used to calculate targets.

So you see, I cover all bases: entry, stops, and targets. Some thing needed on EVERY trade.

2. Have an indicator that confirms the signal indicator. Having a buy/sell signal isn't enough. A confirmation indicator is needed to help ignore the bad trades. DONT take buy/sell signals off your confirmation indicator unless this is part of your plan!.

3. Play with the values of the indicators. I find that a 34/10/10 stochastics works great for ME, but try other values to see what you like, and works for your system.

I'll be nice and share my system with anyone interested, but its still very new. So do your own backtesting.

KON
More...

Thx Kon - interesting view. I tend to use the shotgun ( scattergun ) approach with 13 indicators.
 
Quote from Jerry030:

Not really.

The software application looks at history, be it market prices time series, pervious acoustic signals from geological formations that have produced oil or failed to produce it or previously intercepted voice patterns of known terrorist suspect then identifies patterns and relationships that are currently unknown and unseen by a human reviewer and learns them. If they hold up on unseen data not subject to the same learning and discovery process and on future unknown data (works in the real world), then it's of value as something unknown has been discovered and verified.

Attached is an example of predictive analytics applied to FOREX.

Goal: trend detection.

Data Structure: modified Landmark

Analytic tool: neural network

Market: Forex EUR/JPY hourly bars

Training Set 24,000 bars

Stable out of Sample Performance: about 4,000 bars

Retraining time: about a week.

As you can see it doesn’t catch the exact trend turn point but it does get signals bars sooner than conventional linear rule systems, which havce to wait for the price to move as they aren’t pedictive.

The other drawback is as you mentioned in non-linear/asymmetric systems about 55% of the time there is nothing to detect as the market is chaotic or at least the analytics process could not learn a pattern structure. So sometimes it doesn't trade for a few days.

Jerry
More...

So your software is a type of market sonar?

If I am understanding it the software takes non-correlated data from a variety of sources. Then, much like a sonar would, it transmits a topographical image of the targeted terrain, which in this case is EUR/JPY, correct?

If so, interesting concept.

:)
 
Quote from Redneck trader:

Straight up answer – profitable Sir - to make a living on

RN
More...

That's excellent as not all that many people do.
I tried you approach years ago and it didn't work for me.
So I went on to find one that did.
 
Back to basic indicators, another combination is to use the MACD(12,26,9) and a 12 period SMA. Trade using pin bars, hammers, price failures, whatever u call them in the direction of your now defined trend.

Take your trade metrics from a volatility study.
 
Quote from RCG Trader:

Not in a visual sense if that's what you are asking.

Landmark and similar feature extraction logic algos were developed when the price of data storage technology made it possible to essentially save every bit of measurement information ever acquired or created from a particular process. Storing millions of huge data streams was affordable (voice prints of cell phone traffic, medical EKG/EEG readings, etc) but there was a new problem: how to find commonality in all these waveforms (not unlike a market price wave form in some ways) to extract useful information. One of the initial uses was to read the waveforms in EKGs to find markers that would indicate the possibility of specific medical conditions.

The process takes the raw waveforms and finds abstracted pattern markers of various kinds. The closest analogy in the market I can think of is Japanese Candlesticks, except it’s far more complex. However this isn’t visual but mathematical, the raw waveforms are fractured into something similar to the system used to organize life forms: genus, species, etc, etc.
by using a large number of archetype structures. All this is of course automatic. Dealing with this level of complexity with human analysis and discovery would take a long, long time. Add to this the complicity each market has a unique WP (Wave Print) at some level in the hierocracy tree, which evolves over time, like a living organism.

So to summarize, sonar isn’t’ the right analogy as a skilled human can read a sonar screen and tell a whale from a submarine.

Also this process uses only one data source: the price of the instrument traded. There is no variety of non-correlated data sources. It could be done of course but the complexity
would require some really heavy iron in terms of computer capacity.
More...
 
You must log in or register to reply here.
Back
Top