1 Day Return builds computer trading algorithms for institutional investors and hedge funds. Our programs are built to be self-learning computer algorithms that we develop to run by themselves every day, multiple times a day.
We program them to be self-learning so that they will measure their results on a daily basis against all previous data collected. The programs make weighted purchase decisions based on that activity so that "stronger" recommendations will get more capital while "weaker" recommendations will get less.
Every time an algorithm runs, it scans a list of over 6,000 stocks for indicators. Each algorithm may trade anywhere from zero trades to thousands of trades per day.
After several months of development we launched the basic algorithm program on March 9th, 2009 by giving it $3 million in play money to invest in a stock market simulator. The chart of gains and losses below represents the end of day gain/loss for each trading day. We ran the program every trading day until the 17th of April, after which it was shut down due to achieving a 33% gain.
Algorithm Results for March
This will be replaced by the player.
Algorithm Results for April
This will be replaced by the player.
The front page of this site is an example of a subset of the algorithm's program being used on a stock by stock basis. Because we are using the program on a larger scale, we can't guarantee any of its recommendations in any way. Data in the simulator so far has shown that the "Strong Buy" recommendations are returning 70-80% accuracy, but since the market changes quickly its really impossible to say whether these results will hold up over time.
As each program goes into beta mode, we will put its entire trading functionality on display on 1DayReturn so that you will be able to follow along with its daily activity.
October 25th, 2010 -
Project Kronos is now online
