Using Skew/Kurtosis for Trading Ideas

[Michael Maher]

Has anyone used skew/kurtosis indicators in trading?  I've developed an indicator that calculates both metrics using a VWAP as the average to calculate skew/kurtosis. Visually it has been interesting but I've failed to see how I can develop this further. Anyone interested?

[Fernando Carreiro]

Michael Maher: Has anyone used skew/kurtosis indicators in trading?  I've developed an indicator that calculates both metrics using a VWAP as the average to calculate skew/kurtosis. Visually it has been interesting but I've failed to see how I can develop this further. Anyone interested?

Yes, I sometimes use those metrics on EMAs, mostly skewness, but sometimes also kurtosis. However, you have to truely understand their statistical meaning if you want to use them in your trading.

Skewness can help identify in which "direction" the data is tending towards (in which direction it will most probably trend towards predictively), and kurtosis can be used to measure the "strength" or decisiveness of a trend's direction

[Michael Maher]

I appreciate the insight, and I've noticed that kurtosis spikes tend to indicate regime changes, accompanied by positive skew for overbought and negative for oversold. It makes intuitive sense with what that's telling me as the spikes would be heavy-tailedness, implying mean reversion.

With vwap giving a good estimate of the mean, its hard to ignore the visual results. I feel there may be something deeper to this, but I'm stuck at the moment. Another issue to creating a profitable system is finding good stop/profit levels, I haven't made any significant headway into this. 

[Fernando Carreiro]

Michael Maher #: I appreciate the insight, and I've noticed that kurtosis spikes tend to indicate regime changes, accompanied by positive skew for overbought and negative for oversold. It makes intuitive sense with what that's telling me as the spikes would be heavy-tailedness, implying mean reversion. With vwap giving a good estimate of the mean, its hard to ignore the visual results. I feel there may be something deeper to this, but I'm stuck at the moment. Another issue to creating a profitable system is finding good stop/profit levels, I haven't made any significant headway into this. 

You are ignoring or forgetting about variance (volatility) — Mean, Variance, Skewness, Kurtosis — the Four Moments!

[Michael Maher]

Thats a good start, when I used the variance last time I wasn't getting much out of it to determine profit/stops. Or so it seemed I hadn't extensively tested it. Gonna go back to it.

When you use skew with EMAs, are you using it conjunctively to decide trend of the EMA? 

[Fernando Carreiro]

Michael Maher #: Thats a good start, when I used the variance last time I wasn't getting much out of it to determine profit/stops. Or so it seemed I hadn't extensively tested it. Gonna go back to it.

Most of the time variance needs to be converted to standard deviation to be of use when deciding stops.

Michael Maher #: When you use skew with EMAs, are you using it conjunctively to decide trend of the EMA? 

It depends on the strategy.

[Michael Maher]

Usually I'm going off large kurtosis spikes. There could be some use in options trading but dont see skew giving much predictive behavior. What strategies do you usually apply higher moments to?