My road of trials, 1983-1994

When someone on the team saw something I had missed, the goal became to extract exactly where our views diverged, test that specific difference against the evidence, and learn from whichever side turned out to be wrong. Over years of doing this deliberately, I learned to separate a person from their idea and judge the argument on its own merits — a habit that sounds obvious in the abstract and is genuinely difficult to practice when the disagreement is personal and the stakes are real money.

This period was also when the firm's approach to hiring, training, and setting standards started to firm up into something closer to a system. The underlying work was never only about investing well in the moment; it was about designing, deliberately, how decisions actually get made when the stakes are real and emotions are running hot enough to distort judgment — because that's precisely the condition under which a good process earns its value.

One specific innovation from this period was writing investment decision rules down as explicit, testable criteria — closer to algorithms than to gut feel — so that a rule could be checked against history, refined, and eventually even encoded into early computer models rather than living only in one person's head. The trials of this decade kept repeating the same underlying message in different forms: strong results come from strong processes, reinforced daily, especially — maybe only — on the exact days you least feel like following them.

A second discovery from these years reshaped how the firm invested. I learned that a portfolio built from many good return streams that don't move together is far more reliable than any single brilliant bet. Combining enough of them cut risk sharply without cutting expected return, which I later called the Holy Grail of investing. That insight, plus the discipline of separating our active bets from plain market exposure, became the backbone of the flagship Pure Alpha approach — and it meant a single bad call could no longer threaten the firm the way 1982 had.