ML Model #2 Is Probably Done
I added a fifth model to the launch lineup on a whim back in March. It won the first iteration, so I kept it. It is now last of the four, and it is probably getting retired in October.
Read more →Updates, insights, and lessons from the experiment.
I added a fifth model to the launch lineup on a whim back in March. It won the first iteration, so I kept it. It is now last of the four, and it is probably getting retired in October.
Read more →The third iteration was ugly. Every portfolio lost money and every portfolio lost to the S&P 500, by 10 to 20 points. I said last time I was bracing for a rough stretch, and here it is. However, we learn much more in failure than we do with success. I do not want to say I was hoping for this to happen, but in some way I was. The honest story is not just that the AI trade cooled off, it is that a proper audit of my own data turned up gaps I did not know were there, and a setting I thought was a tiebreaker turned out to be a lock.
Read more →The second 42-trading-day iteration is in the books, and it was a monster — every portfolio beat the S&P 500 by 16 to 33 points as the memory and semiconductor names went vertical. I am pleasantly surprised by where the first two iterations landed, and a little nervous about how concentrated that success has become.
Read more →The first 42-trading-day iteration ended with all four portfolios ahead of the S&P 500. That is a great start, but the index had a strong run too, so the wins say less about the models than they do about the tape.
Read more →The S&P 500 barely moved (+0.54%) but every portfolio kept extending its lead, with ML Model #2 now up 24.49% since launch. One Wednesday rally did most of the work, and the first scheduled rebalance is now five trading days away.
Read more →Earnings season is here, the S&P 500 had its best week in a while at +4.52%, and every portfolio managed to do even better. Growth Evaluator quietly led the pack at +7.42% for the week.
Read more →Six weeks in, every portfolio is finally positive since launch. ML Model #2 leads at +10.46%, with the S&P 500 barely hanging onto a positive cumulative return.
Read more →A four-day week courtesy of Good Friday, and all four portfolios made the most of it. Every model beat the S&P 500, and ML Model #2 crossed back into positive territory since launch.
Read more →After three straight weeks of beating the S&P 500, all four portfolios took it on the chin. A brutal Thursday sell-off did most of the damage and erased ML Model #2's lead since launch.
Read more →All four portfolios outperformed the S&P 500 for a third straight week, but a sharp Friday sell-off erased most of the week's gains. ML Model #2 continues to lead with +1.83% since launch.
Read more →Another volatile week driven by Middle East oil fears and a jittery VIX, but all four portfolios outpaced the S&P 500. ML Model #2 even squeezed out a positive gain.
Read more →Our first week coincided with one of the largest global conflicts in recent history. Here's how the portfolios performed and what we learned from our biggest losers.
Read more →After over a year of building, tuning, and backtesting, TheStockExperiment.com is finally live. Here's the story behind the launch.
Read more →