The Joke That Kept Being Right

The Super Bowl does not forecast the stock market.

That claim has to get past a stubborn number. From 1967 through 1997, one rule was right 28 years out of 31, scored on the Dow. An original National Football League franchise wins the game, and the Dow finishes the year higher. An original American Football League franchise wins, and the Dow finishes lower.

Twenty-eight hits. About 90 percent.

A fair objection is that stocks rise most years, so a rule that usually says “up” will look clever. In those same 31 years the Dow finished higher 23 times and lower 8. Someone who ignored the game and predicted “up” every January would have been right 23 times. The indicator beat that person by five years. It matched several real down years after wins by the Jets, the Dolphins and the Raiders. I would not call it a bookkeeping trick. It beat the market’s plain habit of rising.

What are those five extra hits allowed to prove?

For three decades, the match was real.

I still would not trade it.

There is no path from the winning locker room to the earnings behind the Dow, or to the rate at which investors price those earnings. Which league a club belonged to in the 1960s changes neither. A mood story fails a simpler test. The Pittsburgh Steelers, the Baltimore Colts and the Cleveland Browns were old NFL clubs that moved into the American Football Conference in the 1970 merger. Under the rule, a win by any of them counted as bullish. If you remember the indicator as “AFC means down,” you remember a different rule from the one that posted the famous record.

Much of the late run was a football dynasty sitting on top of a bull market. From the 1985 game through the 1997 game, the NFC won 13 Super Bowls in a row, and every winner was an old NFL club. The rule said “up” 13 times. The Dow agreed 12 times, and fell only in 1990. That is overlap, not a finding about risk.

The earlier years are what made careful people repeat it. Old AFL teams won several of the years the Dow actually fell. Once someone had noticed the pattern, it kept agreeing.

Leonard Koppett, a sportswriter, wrote it down in February 1978, in The Sporting News, after the pattern was already visible. A follow-up ran in Sports Illustrated in April 1979. The version that spread said the column had been in The New York Times. Asked years later, Koppett could not point to one.

I wrote earlier this month about how to read a market statistic with only a handful of cases. Six rows feel thin. Thirty-one years feels like the point where you are allowed to be impressed. I don’t think the extra years did that job. The rule was chosen because it had already fit. The next decade mostly repeated the fit, while old NFL teams kept winning and the Dow kept rising. It never asked whether a football score belonged in a forecast.

Then the champions changed.

The Broncos, an original AFL team, won in January 1998 and again in 1999. The rule called for down years. The Dow rose about 16 percent, then about 25 percent. In February 2008 the Giants, an old NFL team, won, a bullish call. The Dow fell about 34 percent. The Chiefs, an original AFL club, won in 2020, 2023 and 2024. The Dow finished higher in each of those years. A worn-out rule can still be right by accident. An AFL win in a real bear year counts as a hit, and those years still occur. What failed was the claim of unusual accuracy.

I may be wrong about how much of the early alignment was luck. Eight down years make a short list, and a few of them can land on one league without a cause. On the mechanism, I have no such doubt.

There isn’t one.

On December 7, 2001, Jason Zweig called Koppett and asked about the indicator. The long streak was already over. Koppett did not try to salvage the years that still matched. “It’s a joke!” he said. He had meant the whole thing as a satire on the fallibility of human statistical reasoning. “It’s too stupid to believe.”

A quarter-century on, he was still trying to get that heard.