January, 2018 Life Member Affiliate Group Meeting (by CR Kime): At the January 3, 2018 IEEE Madison Life Member Affinity Group meeting, Professor B. Ross Barmish of the Department of Electrical and Computer Engineering, UW- Madison, presented an overview of his collaborative research via a presentation entitled “On Stock Trading Algorithm Research Based on Adaptive Feedback Control Loops.”
In contrast to model-based technical analysis methods which use future stock price predictions from past stock price behavior, his model-free approach employs a feedback control to determine investment level. To motivate his work, Barmish presented conceptual arguments, as well as an example showing how price prediction methods can run amuck in the presence of volatility in the market. Reference was made to model-free approaches by others in the 1990s, with embellishments to address risk containment in 2000. An analogy between coin tossing and stock trading illustrated drawdown of assets as a critical problem, even in the face of a perfectly modelled coin. Beginning with a very elementary example, the coin-tossing problem was modeled by a simple feedback control loop. This simple framework was used to demonstrate that models with initial significant constraints on market behavior can be developed that permit the modeling of a stock transaction, incorporation of non-linearities and modeling for long or short trades. The culmination of this is his Simultaneous Long-Short (SLS) Controller.
Back-testing simulation results presented demonstrate the effectiveness of the SLS Controller on detailed daily price data with realistic practical considerations. For Apple stock, after the Crash of 2008-2009, a simulation using the SLS Controller demonstrated that it properly abandoned short investment in a rising market. Another simulation, again using the Apple pricing behavior, involved active trading by a day-trader. In this case, the assets were volatile but down-trending in spite of the rising closing prices. For this scenario the results indicated that day-trading can be problematic in terms of asset management. A final simulation was described for a very volatile time period for the Euro FXE currency. Using a Buy High – Short Low approach, beginning with a $10,000 initial investment, a gain of 35% over 1000 days with much less volatility than that present in the currency trading prices was accomplished.
The talk concluded with a brief description of five areas for future work followed by a great question and answer session.
February ECN Meeting (Review by Randall Iliff): On February 8th the IEEE-Madison ECN “Meet and Greet” took place at Sector 67. The group was small, (five in attendance), but resulted in a number of useful introductions and connections. Stefan, Jim and Randy were attending the group for the first time. Tom and Chuck shared the history of the group, events from last year and discussed upcoming topics including how the new tax law will impact those in consulting roles. Tom provided an overview of the mission of the group to connect talent with needs and invited input on how that might be accomplished. Suggestions included finding a more “business recognizable” location for some events. The thought was that Sector 67 is very entrepreneur friendly, but may not be seen by corporate innovation leaders as a place to find talented professionals. Suggestions included the MG&E Innovation Center, UW locations, and others. We also discussed how to reach the levels and roles in organizations that would represent good connections. Because the group was small, we also had time to engage in interesting conversations around current topics.