Jurik RSX, MESA Ehlers, Regression (linear, quadratic, logarithmic, exponential - all in one)

[fordewind]

Your job then becomes finding the regimes of price where it's not random.

And what are the best tools for that in your view?Are those mentioned not suitable?

 

[ironchef]

General principles. Although I do find Ehlers work interesting and sometimes useful.

Maintain your perspective about technical analysis as well as financial modeling. Both Ehlers and Jurik model market prices as signals with additive noise, and use digital signals processing (DSP) to analyze price movement. A quick word on models: all models are wrong, even the ones developed by Nobel prize winners in Economics.

To view prices as a signal plus noise is wrong, as is the random walk model. It's just a matter of the degree to which they are wrong.

Thank you.

I can believe that stock price has two components: A random walk component because most of us buyers and sellers have no clue where it is going. 49.99% think it is going up and 49.99% think it is going down. Then there is a .02% non random component because a few have profound knowledge. How do you pick it out of a sea of noise with DSP? In applying DSP you got to know what to look for.

 

[traider]

At least with Ehlers work, you can find the source code and modify according to your needs. Although with some searching, could probably find Juriks' code as well.

Just keep in mind, that if you believe price movement is best modeled by a random walk (lognormal distribution with skew and kurtosis), then technical analysis is charlatanism. Your job then becomes finding the regimes of price where it's not random.

That can be done by comparing distribution of returns, when the distn is not normal it could be trending. Problem is it might be too late to enter then. Any advice on how to overcome this?

 

[panzerman]

To address the past few responses, it is kind of a philosophical debate you must have with yourself on which market model is best suited to your trading approach. Even random walks can appear to form trends. Really the best way to find a trading system with positive expectancy is to do proper back testing.

However, you could literally spend years backtesting linear indicators and not find one that is profitable, all due to random price movements. So should you trade options based on probability of price movement calculations? Should you trade off of fundamental data, or news? Maybe, maybe not. I can't tell you which indicator or which market model is best, that's something you must discover for yourselves, based on your trading style.

About me personally, I've spent years as a break-even retail trader pursing reversion-to-mean/extreme value strategies in an effort to find non-random regimes. I am a scientist by training, and so my interest in trading is based solely on analysis of price data, finding repeatability in the data, and what can be programmed into an algorithm. I don't care to pour over balance sheets or read the daily financial news. That is just my mentality and what I hope to get out of the markets.

 

[fordewind]

To address the past few responses, it is kind of a philosophical debate you must have with yourself on which market model is best suited to your trading approach. Even random walks can appear to form trends. Really the best way to find a trading system with positive expectancy is to do proper back testing.

However, you could literally spend years backtesting linear indicators and not find one that is profitable, all due to random price movements. So should you trade options based on probability of price movement calculations? Should you trade off of fundamental data, or news? Maybe, maybe not. I can't tell you which indicator or which market model is best, that's something you must discover for yourselves, based on your trading style.

About me personally, I've spent years as a break-even retail trader pursing reversion-to-mean/extreme value strategies in an effort to find non-random regimes. I am a scientist by training, and so my interest in trading is based solely on analysis of price data, finding repeatability in the data, and what can be programmed into an algorithm. I don't care to pour over balance sheets or read the daily financial news. That is just my mentality and what I hope to get out of the markets.

Give us dudes some rod, not fish.