What actually happens inside an algorithm-driven portfolio?

Take away the jargon and an algorithm-driven portfolio is just a checklist. A computer works through it on a schedule, the same way every time.

One cycle

1. Say, the 50 or 100 biggest companies in India. The list is fixed in advance.
2. 2. Score every stock
3. Each one gets a score from the same rules, such as how much it has risen in the last few months.
4. 3. Pick and size
5. Keep the best scorers and decide how much goes into each. A cap stops any stock from getting too big.
6. 4. Place the orders
7. In a typical setup, orders go to the investor’s own broker account and the shares sit there.
8. 5. Check again
9. A few weeks later, score everything again. Sell what dropped out. Buy what moved in.

Then it starts over. If a scary headline turns up in between, nothing changes, because the rules stay the same and only the scores move.

It’s worth knowing what the algorithm is not doing. It isn’t guessing the future, and it has no idea which stock will rise next month. It simply bets that stocks scoring well on a sensible rule will usually beat the ones that score badly. Usually, not always.

Why wait a few weeks between checks? Because every trade costs money. If you check every day, you pay a lot to react to noise. If you check too rarely, the portfolio drifts away from the plan. A few weeks sits in the middle.

In short, an algorithm-driven portfolio is a routine you can read and question, not a prediction you have to trust.