But that is just it. There's no emotion in an autostrat, right? His model failed, like 90% of all the others?
From the article: "As his losses surpassed 10%, Rachmat was compelled to intervene: He took some decision-making power back from the algorithm."
Exactly. This is why proper professionals use Monte Carlo analysis to establish a statistically valid operational envelope for their model(s).
the problem with most all mechanical trading is the drawdown. because of the value of the contracts i trade now i use a base trading model with good money management and pick the trades using statistics.
%% Depends on the market; 25% dd is much better than a loss, in QQQ, for example. Easy for me to say -not long QQQ now + never long TSLA, which may dd down to zero/goose egg LOL.. ,