Search results

For oscillator indicators, look at the work of John Ehlers. Specifically, Automatic Gain Control, the Fisher Transform, and Inverse Fisher Transform.

The attached articles by John Ehlers give you an idea what can be done with adapative EMAs. Btw, moving averages are lowpass filters in digital signals processing terminology. EMAs are infinite impulse response (IIR) lowpass filters. An SMA is an example of a finite impulse response (FIR)...

What you are thinking about is called an Adaptive Moving Average(AMA). Search for the AMA of Perry Kaufman, Tushar Chande, and John Ehlers. There are many statistics that you could use to make the EMA adaptive. I like Ehlers' work the best.

Definition of group delay (lag): The negative derivative of phase with respect to frequency. Again, this implies that price is a complex wave with many different frequency component waves. It's just a model remember and thus the caveat.

Here is the 411 on moving averages and lag: Moving averages are lowpass digital filters. The two types are Finite Impulse Response (FIR) and Infinite Impulse Response IIR). An SMA is a type of FIR filter, and and exponential MA is a type of IIR filter. Lag is the common name for group delay...

All the white splotches gives Taleb's picture a certain Einsteinesque look. In any case, we need tail risk funds for the masses. Something from Fidelity, T Rowe Price, or Vanguard, with expense ratios of 0.5 or less.

Try this EMA for yourself. Technically, it's a one pole infinite impulse response (IIR) filter, and considered Kalmanesque. alpha*(price[today]+gain*(price[today]-EMA[1dayago]))+(1-alpha)*EMA[1dayago] alpha = 0.2 gain = 0.5 Play around with values for alpha and gain variables.

While this is a true statement, it's far too simplistic to develop a trading strategy around. Specific rules need to be applied to this macro view. In fact, algorithmic rules generally these days. Intensive analysis is needed in order to develop such a positive expectancy method. That is where...

Price action is fractal in nature however. If the OP is not inclined to sit in front of a computer screen all day, then they could use the same thought processes on daily or weekly data. It's tempting for a person to think of day trading as a faster way to grow wealth than swing trading, but...

Because if you have truly made an advancement in derivatives pricing, you are wasting your time and ours by posting on an anonymous trading forum. You are not even posting your results at Wilmott. Submitting a paper and getting it accepted at an established peer reviewed journal will get you...

There are no technical indicators or price action patterns that give you indications with anything better than approximately 50% precision. You would be damn lucky to find something that gave a correct indication 55% of the time.

If you are going to use MAs, one should learn to view them as lowpass digital filters so that you can better engineer them to meet your needs, which typically means less lag and better attenuation of the "noise." John Ehlers is the man to study for digital filtering in finance. None of the more...

A reasonable model of price behavior is pink noise. Therefore, a trend is just waves with a longer periods (lower frequency) as the dominant market frequency, and mean reversion is catching waves with smaller periods (higher frequency) as the dominant market frequency. When prices transition...

The problem is, that however you define a trend, when they start and when they end are randomly distributed in time. You need to backtest your definition of a trend in order to get an idea if it has a statistical edge and thus lead to a positive expectancy method. Otherwise you are shooting in...

The biggest problem you have is that you never backtested your system, and try to determine if it has positive expectancy before trading it. When markets transition from trending to cycling modes is highly randomly distributed in time. No statistical edge in your method means no positive...

If not using a random sample of the population, then you could bias your selection for any given set of tasks. Not very smart of Eckhardt to allow such a bias.

It all depends on if you can define what a breakout is to you, so that you can put it into algorithmic form and backtest it. What you want is to determine if your breakout trading method will stand a chance of being positive expectancy when it comes to live trading. Don't know how to define...

Why would people want to trade your instrument versus existing structured products? How would dealers make money from its' trading activity, from the spread?

An important question to ask is how is the sponsor or dealer of the new product able to make money from the products' trading activity? For example, sponsors of ETFs typically make money from fees. If what you have is a model and not a tradable product it will be hard to get it adopted. If it...

On a more philosophical note, this thread begs the questions, can a publicly disclosed trading system that had positive expectancy, still retain positive expectancy? Also, can anyone be taught to develop a positive expectancy trading system? I would say no to both questions.