Sending an alert during the North Korea-Trump events made me reflect on probability. Here are my thoughts plus a performance report.
Recent performance. Since early July we’re had 3 signals. The table shows 0.43% profit from trading at 11:50 and 0.32% profit from trading at 12:00.
One reaction has been, “Wow such small profits.” Well that’s normal. The following histogram shows most returns since Jan 2013 were between -0.1% and 0.4%. And it’s pleasing to see only 9 returns less than -0.1%.
But most importantly, a statistical system can’t be evaluated in 1.5 months.
Economic Probabilities versus Certain Probabilites. I realize the phrase “Certain Probabilities” is an oxymoron but it supports a distinction I’d like to make. When one flips a coin there’s a 50% chance it’s heads. When one draws a card there’s a 25% it’s a diamond. After recording millions of draws from a jar with red and blue jellybeans one can calculate the probability of a red draw. Exogenous events in the future don’t change these “certain probabilities” in closed systems.
Economic probabilities, however, are estimated from history and continuously change – with some stability we hope. For instance, signals from my model found SPY rising 78% of the afternoons over the last four years but it’s unlikely 78% will hold over the next 4 years. In fact, probabilities in each year were: 74%, 91%, 80% and 69%.
Since no afternoon in the prior 4 years had a war scare I issued the warning: “North Korea and Trump willing” in my August 9th alert.