World Cup model maker says failed picks show limits of forecasting

A World Cup forecasting model from the author of Klement on Investing took a public hit after Japan, Germany and the Netherlands all lost in a tight knockout stretch. For everyday investors, the episode lands beyond football: it is a reminder that models can frame probabilities, but they cannot remove luck from outcomes.

The author, who describes himself as a professional economist and investment strategist, wrote that after 12 years and three successful World Cup runs, his luck finally ran out. He said his aim had been to show that economic models are less precise than many people assume, and joked that the failed calls ultimately supported that point.

In a post before the matches, he had warned that he could arrive at the office with “serious egg on my face” if Japan, Germany and the Netherlands all lost within 12 hours. After they did, he said he was at his desk with a “No Refunds” poster and a Dutch flag at half mast.

Penalty shootouts did the damage

The author pointed to the knockout format as the core reason forecasts can break down. In single-elimination matches, one result ends a team’s tournament, and he said each of the recent games could have gone the other way, especially those decided by penalties.

Penalty shootouts are a clear example of model risk, meaning the chance that a forecast tool misses because real life does not follow tidy assumptions. A model can estimate team strength, form or likely matchups, but it cannot fully account for one saved shot, one miss, or a small bounce at the wrong time.

The author congratulated Brazil and its fans, including a nod to Neymar Jr., and also praised Paraguay and Morocco for what he called major upsets. He said the Japanese national team had shown itself to be a strong collective that achieved things many people would have viewed as impossible.

He also apologized to Dutch supporters if his forecast gave them false hope. The Netherlands’ loss, he wrote, was especially cruel, citing goalkeeper Bart Verbruggen’s near-save in the penalty shootout. He said he still believes the Dutch side can win a major trophy, possibly at the European Championship in two years, while adding that he does not plan to forecast that tournament.

The model is not being retired

The author said his inbox filled after the misses, split between people mocking the failure and others asking whether he would retire the model. His answer was no.

He compared the experience to investing, where forecasts are often wrong even when the process is serious. He cited the saying that to make money, a person needs to be right “six out of eleven times,” and said his own World Cup record stands at three correct calls out of four.

His broader message was that luck plays the largest role and that no model or system can produce perfect forecasts. The best outcome, he wrote, is to be right more often than wrong.

He said he expects to return with World Cup forecasts in four years, as long as readers treat them as fun and entertainment. For investors watching from the sidelines, that may be the useful lesson: even a good model is still a guess with structure around it.

This story draws on original reporting from Klement on Investing.