The laws of trading technology: Patents define the field

April 1, 2015 12:00 PM

The application interface structure was a generic interface between the processes, which included updating and tying into the natural order recalculation between spreadsheets and the neural network. That interface was patentable. This interface also did not depend on the spreadsheet in which it was embedded (see “Spreadsheet integration,” right).

This interface also could be updated using data from the neural network from an event-driven model. The original version was embedded into a Lotus 1-2-3 clone written in C. We had full source code for this project. We then embedded it into Microsoft Excel, to which we did not have the source code. We linked between the interface and the spreadsheet using Dynamic Data Exchange (DDE), the grandfather of today’s  Object Linking and Embedding (OLE) technology. This generic interface between the application and neural network meant that only some support routines needed to be rewritten. The interface was the key to the patent.

This neural network patent passed the data transform test. It took data from a given input (a spreadsheet), transformed that input into different forms and converted it into models, predictions, forecasts, etc., and then transformed those back into input.

However, the machine transformation test is not enough to make something patentable. The transformation must constitute a practical application of a mathematical algorithm, formula or calculation because it produces a useful, concrete and tangible result. 

During this period, Braincel, the product created with this patent, had several competitors but none of them could update the output of the neural network when dependent cells changed in real time. This was at the heart of this invention, and why it was a proper patent.

Page 2 of 5
About the Author

Murray A. Ruggiero Jr. is the author of "Cybernetic Trading Strategies" (Wiley). E-mail him at