Search results

Whether you use the geometric mean, arithmetic mean, you are still using a lowpass filter, and are subject to group delay, commonly known as lag. Specifically, your moving average is a finite impulse response (FIR) lowpass filter. The group delay of the filter is given by the formula (N-1)/2...

Fractal market structure seems to make intuitive sense, but since the late 80's and early 90's when fractal market analysis started to become a "thing" to study, I have never seen anything published which could be termed a profitable trading system. Fractal market theory seems to be sound, but...

For those wishing to use MAs, try this exponential moving average. alpha*(price+bestgain*(price-EMA[1]))+(1-alpha)*EMA[1] where alpha = 0.25 and bestgain = 0.5 Next, start playing around with different alphas and gain factors. Best of luck. btw, EMAs are just infinite impulse response (IIR)...

Any of the books by John Ehlers: Rocket Science for Traders Cybernetic Analysis Cycle Analytics Ehlers work is about modeling the market as cycles plus noise, using tools from the field of digital signals processing. Some people are of the philosophy that anything from the physics realm...

I tell you what, I would rather run through sand and mud carrying a 500lb log with a group of men who where true friends, rather than eat salmon topped with gold leaf and truffles with a group of folks who couldn't care less if I fell off the face of the earth.

Better stuff than you apparently. The algebra hasn't changed, but the concepts of digital filters, group delay, nyquist sampling rates, etc. etc. didn't exist 100 years ago. Certainly traders were not using anything more advanced than SMAs. Clerks were not calculating Inverse Chebysev...

Whoever reads this thread and intends to use MAs, please use some of the newer low-lag MAs, and not the traditional ones. There are free ones by John Ehlers, or paid ones by Jurik. Lag, or group delay, will kill your performance with the traditional MAs. Also, please take the time to learn how...

Certainly not perfectly Gaussian, as all financial distribution have some amount of skew and kurtosis. However, it's random enough that this simple MA crossover scheme will not produce a positive expectancy system. Also, R/S analysis has shown time and again that daily data is very nearly...

That's the game isn't it? Finding when one market regime ends and another begins, in real time. The problem is that regime changes are randomly distributed through time. However, not perfectly random, so perhaps there is your opportunity.

Me either, just taking a light hearted spin on the opening posts lament of why their method suddenly stopped working. Then Gaussian gets bent out of shape at the meer hint of EMH and random walk. Guess what, that is still the best explanation in a broad and general sense of how prices evolve...

I'm sure with your intellect, you have proved both of these points already. And of course on this anonymous public forum you do intend share those proofs, right? You truly must be the wealthiest person in the world!

Sort of as if prices randomly changed their state. As if prices were randomly walking around. Maybe one could model such price behavior with geometric Brownian motion (waiting for Mr. Mcginnis to chime in.) St = St-1* exp((μ-(σ2/2))*t + σWt) Where, St is stock price at time t St-1 is stock...

Beowulf

Sounds like momentum is a bit like fractal long-term memory in the markets. Lots of studies over the years have tended to show that the phenomena exists, but it does not appear that anyone can capitalize on the effect either.

People sell OTM options because of the higher probabilities of a winning trade. However, people selling options should look at maximizing payoff instead of just probability. Payoff calculations will sometimes lead to selling ITM or ATM options. payoff = probability*(reward/risk)

Try using two different EMAs with different alphas of the form: y[0] = alpha*(x[0]+0.5*(x[0]-y[-1]))+(1-alpha)*y[-1] Not that I think it will ultimately result a profitable trading system for you, but if you insist on using MAs, at least use ones that have minimized lag.

There are techniques based on linear prediction that project values into the future based on past data. However, I doubt in trading that they would give anything more than 50% accuracy in results, but I haven't studied the technique too extensively.

It's been a truism in financial markets forever that most traders are net losers. It's because the methods they use are too simple and cannot effectively deal with the randomness of price movements over time. You trash random walk, and its' associated skew and kurtosis, but never explain let...

Absolutely that's why it is so hard. When I suggest that Random Walk is still the best description of price movement, even with it's well known flaws, people get pissy. I do concede that there is just enough non-randomness in markets that it might be possible to exploit with TA methods. However...

That's kind of my point. The underlying science people are using to gain their "edge" is based on assumptions of the way in which price moves. Those assumptions are incomplete descriptions of reality. That's why all models are wrong to one degree or another. That doesn't mean you can't make a...