We tested the McRib indicator: bitcoin did worse than random

The McRib came back to McDonald's yesterday, and with it the claim that bitcoin doubles afterwards. Luke Nolan of CoinShares checked it against ten returns and found nothing. We ran the test ourselves on daily closes, because a debunk is only worth as much as the numbers behind it, and ours came out slightly worse than his.

“There is no correlation at all. People omit McRib returns when it doesn't fit the narrative.”

— Luke Nolan, CoinShares, https://decrypt.co/381909/mcdonalds-mcrib-doesnt-actually-boost-bitcoins-price-sorry-everyone

We took the six McRib national returns we could date to a day or close to it, and measured what bitcoin did over the following 30, 60 and 90 days.

Thirty days later, three of the six were positive. The median was +0.2% and the mean +7.4%. Ninety days later, again three of six, with a median of −3.4% and a mean of +16.3%. Both means are carried entirely by one date: 2 December 2020, when bitcoin went on to add 152% in three months. Strip that out and the pattern is a coin flip with a negative tilt.

Worse than a randomly chosen day

Now the comparison that settles it. Over the same decade of daily data, a randomly chosen day gave a positive 30-day return 56.1% of the time with a median of +2.5%, and a positive 90-day return 58.2% of the time with a median of +8.4%. The McRib dates hit 50% at both horizons. Buying because the sandwich returned did worse than buying on no reason at all.

That is the shape of every indicator built this way. Pick enough start dates and some of them will precede a rally; the ones that preceded November 2021 or the first quarter of 2022 quietly fall out of the thread. Nolan put it plainly: people omit the returns that do not fit. Carlos Guzman of GSR Markets added the obvious missing piece, which is that there is no mechanism connecting a menu decision to a futures basis.

The same lesson as the golden cross

We have run this exercise before on signals people take seriously. The golden cross turned out to be 0.133% deep across eight precedents, which is a technically real event with no size to it. The McRib is the same lesson with better jokes: the test is cheap, the dataset is public, and the answer is almost always that the baseline was doing the work.

This piece is informational, not a recommendation to buy, sell or hold any asset. We measured six McRib national return dates, two of them accurate to the week rather than the day, against BTC-USD daily closes on 27 October 2026; six observations cannot prove or disprove anything on their own.