First: what the company does and how it makes money, in plain language. Second: why it will earn more in five years — the specific engine, not "the market is growing". Third: what would prove me wrong — which event or number would tell me the idea failed. If you can't write those three lines, there is no thesis, only a feeling that went looking for supporting data.

The third line is the most important and the rarest to actually write. It's what turns selling into a planned decision rather than an emotional reaction to a decline.

Between 10 and 20 positions. Fewer than eight and you're exposed to a single event at a single company; more than 25 and you hold an index with commissions and effort attached. Entry is usually made in three tranches rather than one — not to "time" anything, but to reduce the cost of a timing error.

The add rule that separates investors: add to winners whose thesis strengthened, not to losers whose thesis weakened. The natural instinct is precisely the opposite, which is precisely why it costs money.

Sell when the thesis is disproven, when margins erode for three straight quarters without explanation, when management loses credibility, when valuation reaches a historical extreme requiring implausible assumptions, or when there is a clearly better use for the money. Don't sell because the stock fell 20% — a decline is not information about the business.

And the most important thing over the long run: compounding needs time. High turnover erodes returns before tax even enters, and in Israel real capital gains are taxed — so every early sale charges twice. Fewer, better-founded decisions are the private investor's genuine advantage over institutions.