Skip to content

Latest commit

 

History

History
25 lines (18 loc) · 1.39 KB

release-notes.md

File metadata and controls

25 lines (18 loc) · 1.39 KB
Raw

Grademark release notes

v0.2.1

BREAKING CHANGES

  • Rebuilt the optimization function:
    • You now have the option of two algorithms: grid and hill-climb.
    • The grid search algorithm is exhaustive but slow and gets slower the more parameters you are trying to optimize.
    • The hill-climb algorithm is non-exhausive but much faster, especially as the number of parameters increases. Use the option numStartingPoints to choose the number of random starting points that are used to seed the algorithm.
    • Restructured the output of optimization for memory efficiency and easier visualization.
    • You can now set the seed used for random number generation.
  • Reduced the dependency of Grademark on Data-Forge:
    • Most results from and inputs to Grademark no longer use DataFrame. They are just plain old JavaScript arrays.
    • The inputSeries parameter still does use a DataFrame.
    • To convert your old Grademark code:
      • Instead of inputing a DataFrame call toArray on it instead to convert it to an array.
      • If you want output as a DataFrame just wrap the returned array in a new one, e.g. const trades = new DataFrame(backtest(...));

v0.0.1

BREAKING CHANGES

  • Arguments to functions for strategy rules have changed. Instead of having individual arguments to each function, arguments are now bundled in objects for future expandability and better auto-competion.