A great chant is supposed to be a country's free, self-renewing marketing — the thing a stadium of strangers will sing without being asked. As the 2026 World Cup approaches on home soil, the open question for any brand, broadcaster, or federation is simple: which chants actually land with an American audience, and how much of that is the hook versus the flag behind it?
We put all 13 tournament chants to a nationally-spread panel of 391 digital twins, twice. First blind — the chant text alone, with an English translation where needed — then with the singing country revealed. Matching the same twin across both passes isolates the pure effect of national identity. For every chant we captured a 1–5 appeal rating and the reasoning behind it.
Brox.AI runs panels of AI digital twins — models of real survey respondents that carry their full background and reasoning. Ask the panel a question and you get an answer in hours, with no recruitment, fieldwork, or waiting. What separates a Brox run from a survey:
| Panel | 391 US adult digital twins, spread across Gen Z → Silent generation, both genders, carrying their full background profiles |
| Stimuli | 13 real World Cup chants, shown as text with an English translation where the original isn't in English |
| Design | Two passes on the same panel — blind (chant only) then country-revealed (singing nation attached). Matching twins across passes isolates the identity effect |
| Appeal metric | How much they like the chant — 5-point scale (1 very poor → 5 excellent) |
| Qualitative layer | Every rating carries the twin's spoken reasoning; ~391 responses per chant per pass (10,163 in total) |
| Run date | 11 June 2026 |
Mexico's "¡Ay, ay, ay!" scores 3.24 / 5, the only chant above the midpoint and #1 in both passes. The field average is just 2.50; 12 of 13 chants sit below 3.0.
81% of ratings are identical once the nation is attached; the overall mean shifts +0.05. National identity nudges a chant by a tenth of a point — it never rescues a weak one or sinks a strong one.
All three USA chants land below average — U-S-A 2.78, "I believe" 2.59, "We love ya" 2.41. The flagship "I believe that we will win" ranks 7th of 13, even after a home-team boost.
Mexico's chant clears 3.0 in every generation and is gender-neutral. "I believe" runs from 2.0 in the Silent generation to 2.9 among Millennials — a 0.9-point age gradient, never above 3.
"Two World Wars" floors at 1.37 (100% rated 1–2) as a "war brag"; Iran's name-chant at 2.02 as empty. The reveal barely moved either — a flag can't redeem a hook the room finds hollow or hostile.
Mean appeal (1–5), blind pass — chant text only, before any country was revealed.
| # | Chant | Country | Mean | Spread (SD) | Rated 4–5 | Rated 1–2 |
|---|---|---|---|---|---|---|
| 1 | ¡Ay, ay, ay! "Sing and don't cry" | Mexico | 3.24 | 0.70 | 38% | 14% |
| 2 | Olé, olé, olé | International | 2.85 | 0.81 | 23% | 34% |
| 3 | U-S-A | USA | 2.78 | 0.74 | 16% | 35% |
| 4 | Vamos Nippon | Japan | 2.74 | 0.61 | 8% | 33% |
| 5 | Vur! Kır! Parçala! | Türkiye | 2.71 | 0.70 | 13% | 40% |
| 6 | Viking thunder-clap row | Norway | 2.61 | 0.74 | 10% | 44% |
| 7 | I believe that we will win | USA | 2.59 | 0.83 | 14% | 48% |
| 8 | Yes sir, I can boogie | Scotland | 2.46 | 0.63 | 4% | 53% |
| 9 | We love ya | USA | 2.41 | 0.58 | 2% | 57% |
| 10 | Stand up for the Socceroos | Australia | 2.40 | 0.60 | 2% | 57% |
| 11 | It's coming home | England | 2.36 | 0.63 | 2% | 59% |
| 12 | Iran, Iran, Iran | Iran | 2.02 | 0.46 | 0% | 88% |
| 13 | Two World Wars and one World Cup | England | 1.37 | 0.49 | 0% | 100% |
This is a muted field, and that's the point: a general US audience is a tough room for soccer chants, so the gap between chants is the whole game. The single chant that breaks away does it on craft — catchy, positive, singable — not on flag-waving. Judge a chant on how far it clears the field, not on whether it clears some absolute bar; almost nothing does.
"¡Ay, ay, ay, ay, canta y no llores!" — "Sing, and don't cry." #1 in both passes.
At 3.24 / 5 it is the only chant above the midpoint, the only one zero twins rated a 1, and the flattest across age and gender in the whole study. The reasons twins give are strikingly consistent: it's catchy, it's positive, and it's easy to sing — the three words that recur in almost every high score.
And it won on the hook alone — it topped the field blind, before anyone knew it was Mexican. Attaching the country didn't add anything; if anything it slightly cooled the score (−0.09). However a US audience reacts to "Mexico," the appeal was already banked in the phrase itself.
A hook that travels is built, not borrowed. Mexico's chant clears the field on pure form — a singable, upbeat phrase that works even to an audience that doesn't speak the language or know the team. That's the asset to engineer for: positive and singable beats clever or loud, and it survives translation.
The three USA chants in the field, blind appeal — none reaches the midpoint.
| USA chant | Mean | Rank of 13 | Rated 4–5 | Change when revealed |
|---|---|---|---|---|
| U-S-A | 2.78 | 3rd | 16% | −0.03 |
| I believe that we will win | 2.59 | 7th | 14% | +0.15 |
| We love ya | 2.41 | 9th | 2% | −0.04 |
The de-facto USA anthem, "I believe that we will win," ranks 7th of 13. Telling twins it was their own team's gave a real +0.15 patriotism bump — the second-largest reveal effect of any chant — but it still couldn't reach 3 / 5, and 48% put it in the bottom box. It was also one of the most divisive chants tested (SD 0.83). The cleaner, barer "U-S-A" actually outscored it, and the reveal moved it essentially zero — because everyone already knew whose it was.
A simple ranking says "the home chant is mid-table." The reasoning says why: there's no home-team premium to lean on. Across all three USA chants the flag adds at most a tenth of a point, and the verbatims reject the content as repetitive or basic — not as un-American. Don't assume a home crowd will carry a weak hook; on this evidence, it won't.
Paired within-twin change from blind to country-revealed. Positive = the flag helped.
The headline is the absence of an effect. Four in five twins rated a chant identically once they knew the country. Where pride helped, it helped a little — and on the two strongest or most-recognisable hooks (Mexico, "Coming home") the reveal slightly cooled the score. No chant was rescued or sunk by its flag.
National identity is a tailwind, not an engine. It can add the last tenth of a point to a chant that already works, but it cannot manufacture appeal that the hook doesn't earn on its own. Spend the effort on the hook; treat the flag as reinforcement, never as the strategy.
Mean appeal by generation — the consensus winner vs the divisive home chant.
The pattern holds on gender too. Men rated chants slightly higher than women overall (2.57 vs 2.47), with the widest gap on Turkey's aggressive "Vur! Kır!" (+0.20 male). Mexico and Scotland's chants were the only two that were perfectly gender-neutral — another marker of the consensus appeal that put Mexico on top.
The winning chant isn't winning by exciting one segment hard — it's winning by leaving no one cold. A consensus hook that every age and gender lands on is worth more than a polarising one with a loud base, because a stadium, a broadcast, and a brand campaign all need the room to sing together. Optimise for the flattest profile, not the highest peak.
The two bottom chants are rejected near-uniformly, and the reveal can't save them.
Two World Wars and one World Cup (1.37) is the most uniformly rejected message in the study — 100% rated it 1–2 and 63% gave it a flat 1. Twins read it as a war brag: rude, not playful. Iran, Iran, Iran (2.02) fails the opposite way — not offensive, just empty, "just saying a country name" with no hook to hold. Crucially, both barely moved when their country was revealed (+0.05, +0.08): they were rejected on what they say, not whose they are.
One theme caps every score in the study: in a meaningful share of responses the twin volunteers, unprompted, that they don't follow soccer at all — and it sits behind most low ratings. No hook converts a disengaged listener, which is why even the winner tops out at 3.24 rather than 4+. Read top-box share (who a chant genuinely energises), not the raw average, which is dragged down by people no message can reach.
Each twin saw the chant text first (blind), then the same chant with the singing country attached. English translations were shown where the original isn't in English.
| # | Chant (as shown) | Country revealed |
|---|---|---|
| 1 | ¡Ay, ay, ay, ay, canta y no llores! ("Sing and don't cry") | Mexico |
| 2 | Olé, olé, olé, olé! Olé! Olé! | International / universal |
| 3 | U - S - A | USA |
| 4 | Vamos Nippon | Japan |
| 5 | Vur! Kır! Parçala! ("Kick! Smash! Shred! Win this match!") | Türkiye |
| 6 | [two drum beats, then everyone "rows" an oar in unison with a low "yoo"] | Norway |
| 7 | I believe that we will win! (×3) | USA |
| 8 | Yes sir, I can boogie | Scotland |
| 9 | We love ya, we love ya, we love ya… | USA |
| 10 | Stand up for the Socceroos, stand up for the Socceroos | Australia |
| 11 | It's coming home, it's coming home, it's coming, football's coming home | England |
| 12 | Iran, Iran, Iran! | Iran |
| 13 | Two World Wars and one World Cup, doo-dah, doo-dah… | England |
