If the recent market action has left you swinging, you should know we are living interesting times. The January performance that went by was the best U.S. market performance since 1987 [1]. The quarter that passed was the best since 2009 [2] and the recession we might or might not have is the unpredictable future. Between the two managers, the active and passive, the only thing constant was the swing. The former came with a performance fee and the option of extreme luck, i.e. you sold on October'18 top and bought at December’18 low. Congratulations! your active manager deserves the performance fee [3] for 2018. The latter is nearing zero fees and is indifferent to the swing or your swinging moods. He is willing to pay you for the swing. Where does this leave you; happier or riskier? It does not matter if you feel happier (markets recovered) and not riskier as AI or new tech is not there yet to make you feel confident about a bot managing your money or risk. The bot could flare up (large drawdown) before it delivers.

AI should be able to understand risk, understand human fallibility, understand the difference between robust behavior and fleeting patterns (garbage in garbage out [4]) before it can engender comfort. Where does AI for investment management stand today? A Strength-Weakness-Opportunity-Threat analysis of AI vs. existing solutions brings out the key differentiators. AI should seek robust statistical behavior. AI should be performance attributed systematically (validated) if it has to overcome its weakness emanating from lack of track record. AI creates a diversification opportunity. Overuse of AI is the biggest threat; if it’s transparent.


AlphaBots US NASDAQ portfolio (Symbol: RMIVG20) is validated by Nasdaq, is based on factor agnostic mathematical rules, has zero leverage, a low turnover and is driven by a systematic and replicable process. Despite the recent market volatility and decade beating recovery, the model outshined its peers by 2% in Q4’18 (the way down). This was followed by a sharp recovery into 2019 and a 4% outperformance for Q1’19 (The way up). Recovery ratios (model recovery days/benchmark recovery days) is a great metric to understand model stability. The model is currently testing new highs. The tables below list the other live AlphaBots.

Monthly Performance Returns for RMIVG20 and other AlphaBots

Source: AlphaBlock.org 29 March, 2019

Source: AlphaBlock.org 29 March, 2019

Oct-Dec’18, YTD, Oct’18 - Mar’19 Performance Returns for AlphaBots vs. Benchmarks

Source: AlphaBlock.org 29 March, 2019

Source: AlphaBlock.org 29 March, 2019

AI will not solve your problem until you understand what it can or can not do. Machines can not eliminate risk but they can manage it better. Managers vs. machines is not a contest. Machines are here and they are learning and they don’t enjoy swings.


[1] E. McCormick, S&P 500 posts best January since 1987, Yahoo Finance, 31 Jan 2019
[2] 1Q Market Review: Great returns, but with a twist at the end. The Canadian Press, 30 March 2019
[3] C. McGrath, Hedge funds see fee increases in 2018, Pension and Investments, 30 Jan 2019
[4] C. Thomson, How to Teach Artificial Intelligence Some Common Sense, Wired, December 2018
[5] W. D. Bondt, R Thaler, “Does the stock market overreact?”, The Journal of Finance, 1985
[6] H. E. Stanley, “Phase transitions and critical phenomena”, The Journal of Finance, 1971
[7] A Goyal, S Wahal, “The Selection and Termination of Investment Management Firms by Plan Sponsors”, The Journal of Finance, 2008.
[8] H. M. Kat, F Menexe, “Persistence in Hedge Fund Performance: The True Value of a Track Record”, SSRN, 2002.
[9] J Keppo, A Petajisto, “What is the True Cost of Active Management? A Comparison of Hedge Funds and Mutual Funds”, SSRN, March 2014.
[10] A Lo,”Where Do Alphas Come From?: A New Measure of the Value of Active Investment Management” SSRN, 2007.
[11]  S. Goldwasser, S. Micali, C. Rackoff, “The Knowledge Complexity of Interactive Proof-Systems, SIAM Journal on Computing”, 1989.