Is there an error with the "Inverse Fisher" TASC indicator? Specifically, if I load any security with a value greater than about 6, I don't get any indicator values, but it appears to work when the data series values are less than 5. For example, load "AAPL" and notice that when the stock price gets above ~$6.60 (~8/23/2005) the indicator values stabilize at 1.00, and for larger values (i.e. .SPX), > ~100, the indicator values are NaN. I think for data series greater than 5 or 100 you could divide the data series by 100 or 1000 respectively, to get the indicator to return values -- albeit I'm not sure if the return series would have any predictive worth.

There's no error. It's just the data you feed the InverseFisher with has to be in the -5 to +5 ballpark. It makes sense to apply it to an oscillator rather than price:

InverseFisherRSI
InverseFisherStoch

In fact, when introducing the InverseFisher its author John Ehlers showed how to do it:

The Inverse Fisher Transform (PDF, external website)

Eugene,

Thanks for the link to the pdf... that certainly explains the results I was getting when I applied it to the price data.

Darrell, glad to help.

Happy New Year.