Skip To The Main Content

Browse these ideas available for licensing that have the potential to improve lives while creating jobs to support and grow the region.

All ideas
are listed as non-confidential technical summaries.

Ideas were generated by academic institutions, entrepreneurs, and corporations.

Ideas from our research institution partners can be reviewed by following the web links to their database.

Highlighted ideas from our partners are featured separately and organized by category. All listings are updated as new technology opportunities become available..

For assistance in connecting with, or attaining further information on any of these technologies, please contact us.

Finding Signals Hidden in a Financial Index Time Series

The world of finance has an old adage when it comes to the Dow Jones Industrial Average: Sell in May and buy at Halloween. This implies that DJIA price time series has a yearly periodic component of effectively buying and selling stocks. In response to the lack of scientific validity of predicting DJIA and other market volatility, researchers at Kansas State University have developed a method to discern a yearly cycle or periodicity hidden in a larger non-periodic and random variation of a stock price time series. This method reveals a yearly periodicity in the DJIA and gives validity to the idea of selling in May and buying in late October. The average shape of this periodicity is straightforward, has persisted since 1950, and is not determined by outliers.

Applying this method to the 30 components of the DJIA we find that 13 of them clearly show a yearly periodicity. To determine useful trading information each stock time series is divided into two equal blocks of time and yearly buy and sell dates are extracted from the first (earlier) block of data. These earlier buy and sell dates are then applied to the second (later) block of the time series to determine yearly percent profit as compared to a hold strategy. The difference in the ‘buy and sell’ and ‘hold’ averages from this method is a noteworthy 10.4% - 5.6% = 4.8%/year.

-Creates more predictability in buying and selling stocks
-Investment returns can be distributed over many years
-Reveals certain stocks for which the method will work

-Developed as software for all levels of investing
-Applicable to other price indices such as the S&P500, CAC, DAX and Nikkei
-Potential use in other fields containing large data in time series

-Provisional patent application filed in March 2015.

Additional Details


Kansas State University

Intellectual Property Protection

Pending Patent

Interested? Request More Information