Betting on someone who has done nothing yet: two methods

Peter Thiel is not in Locarno. This article compares two institutions facing the same problem: deciding whom to trust when there is not yet a result to assess.

The festival tackles it at scale. The 79th edition received 7,759 entries and accepted 233, a little over three per cent. The Cineasti del presente section is reserved for first and second features and lines up sixteen of them. A separate prize goes to the best first feature, entrusted to a jury made up of the Frenchman Matthieu Darras, the South Korean Sung Moon and the Swiss-Peruvian director Klaudia Reynicke.

The main competition also includes "A Margem do Rio" by the Brazilians Matheus Farias and Enock Carvalho, admitted even though it is a directorial debut.

Thiel faces the same problem through Founders Fund, launched in 2005. In 2004 he had invested 500,000 dollars in a company run by a twenty-year-old, and that bet became the most cited case of his career.

Where the arithmetic diverges

Venture capital rests on a heavily asymmetric distribution. The fund expects most bets to return nothing, because the yield comes from a very small number of cases. Losses on twenty investments do not matter if the twenty-first makes up for all of them.

A festival cannot work that way. If only one of the seventeen competition films held up on screen, the selection would lose credibility and the following year the audience and the buyers would not show up. The festival has to be right in many cases, not in one.

From this follows a different tolerance for error. The fund looks for the extreme case and accepts being wrong almost every time. The event looks for a high average and cannot afford a comparable failure rate.

There is nonetheless a point where the two methods resemble each other. Both select on indirect signals, because the product does not yet exist in its final form. An investor assesses people, a programmer assesses a film no audience has yet seen.

The festival's advantage lies in the low cost of error. A misjudged film in the programme occupies a cinema for two hours, and the following year's selection starts from scratch. In venture capital the mistake stays in the portfolio for years.