This model is designed to be used with one of the time series smoothing functions (for example, the Line Wavelet Filter, the exponential moving average, or the MACD Enhanced signal).

What factors should be used to choose a tick data time series filter?

  1. As little lag as possible. When a decision is made on filter data, it will reflect the market better.
  2. Removal of small bumps that lead to the model making the wrong decision.

Model Set-up

  • Read the last five days of trading data. For each day, find the downward "runs" in the smoothed data. A downward run is a drop with an upward turn around at the end. Find the largest N (where N=4) drops for each day. Average the drop size over the five days to find the drop size to be used by the model.

Model Outline

  1. When there is a drop that is the size of the drop calculated in the set-up, wait for a "turn around" in the smoothed data. When the signal has turned (appears to be heading up), buy the stock.
  2. if the stock falls sufficiently below the buy value, sell it.
  3. if the stock has risen beyond a certain point and there is a downward turn, sell it.
  4. Sell if the stock has been held for some amount of time, to free up capital.
  5. Sell at the end of the day