Fund results (blue) vs. S&P Index, Dec 2019 to Feb 2021
Stock Prediction System
We wrote some months ago about how we had built a social media meme based prediction system which we have been using for stock prediction – and set up a fund structure to test it. Our first trading year has been “interesting”, to put it mildly. Not only were we learning how to use a meme based predictive system to trade on the markets (very different from paper trading) but no sooner had we started than the markets crashed as Covid lockdowns loomed. An unexpected benefit of a meme based system was its early exit from the February-March crash as the memes went negative and it exited.
53% Profit in 15 months – Year One Results
The results have been very encouraging (and enriching) – a 53% profit in 15 months. As can be seen from the graph above the DataSwarm Fund soundly and consistently beat the S&P.
Our Background – IARPA Challenge Award
We built a predictive analytics system and in 2019 won an Award in the IARPA 2nd Geopolitical Forecasting Challenge. If you ever read the book “Superforecasters”, this Challenge is described. We learnt from this that our approach was very good at financial forecasting, so we built a specialist system to predict asset prices, initially testing it on paper trading, then small sample live trading, then from December 2019 moved to live trading under formal fairly standard Fund rules.
Zeitgeist Signals – Risk and Reward
Because the system uses predictive algorithms, one can select the signal for required risk and reward. We call our signal the “Zeitgeist” – the ghost of the times – as it aggregates a very large number of opinions from a variety of sources. The chart below shows various options for the Zeitgeist output, of increasing levels of possible returns. It shows that the maximum potential return we could have obtained this year was a c 80% return.
That’s an amazing outcome, sure – but there is a catch. It comes with far higher risk – it has a Beta (measure of risk) of 1.1, ie: a 10% higher risk than putting your money in an Index tracker like QQQ for the NASDAQ and get a 52% return (the main Indices always have a Beta of c 1.o). You may decide it is a risk worth taking, you may not….
Variable signal reward levels with their associated risks (Beta)
Achieved High Returns with Low Risk – 0.4 Beta
We obtained a 53 % return with the actual traded fund, and it roughly tracked the NASDAQ as you can see – but our Beta was 0.4 – that’s only 40% of the risk of actually investing *in* the NASDAQ, so we got the same benefit at 2/5 the risk. This was year 1, using a new technology so we had opted for a low risk / (very) acceptable return rather than going for the stars – or broke.
We were advised by many finance people we spoke to that the bigger market is actually for people who want a good return at lower risk and sleep at night, compared to those with an appetite for high reward, high risk approaches.
Anyway, we’ve learned a lot about the technology and done a lot of tuning in year 1. We may just run it a little more hot in year 2 – but we will see what happens. Will Year 2 be as “exciting” as year 1? Who knows – but if you look at the top right of the charts we have just seen an “interesting” jump and correction…..