Written by: Matheus Tavares

You’ve probably lived this scene: it’s a national-team match day (in my case, Brazil), traffic thins out, the streets empty, and the whole country seems to hold its breath at the same time. Offices go quiet, conversations narrow to a single subject, and for ninety minutes a nation of millions does more or less the same thing, in the same rhythm, all at once.

That is what makes a World Cup match so unusual. It isn’t only that a lot of people are watching, but that everyday life bends around a single event, with a level of synchrony that almost nothing else on the calendar can match. A moment like this is, almost by definition, extraordinary: it sits far outside the routine of an ordinary afternoon. And extraordinary moments, it turns out, are exactly what this text is about.

When a whole country changes its behavior together, that change doesn’t stay on the streets or in front of the TV. It leaves a fingerprint in data, and one of the clearest places to see it is in how people move money. During a national-team match, the volume of instant transfers drops to a fraction of what you’d expect for that time of day. At the final whistle it bounces back as if nothing had happened, and in between you can even spot half-time. 

But how can we tell that this is truly unusual, rather than just a normal variation? That’s where a simple statistical idea comes in: the outlier, a value that sits far outside the usual pattern.

So, what exactly is an outlier?

Think of the temperature on an October afternoon in your city, sitting around the same comfortable mark day after day, until one afternoon it spikes far above anything the season usually brings. Or that electricity bill that always looks about the same and, in one particular month, arrives frighteningly high. Those are outliers: values that stray far from what you’d expect in that context.

The keyword is context. A burst of transfers at three in the morning would be strange; at noon, it’s expected. An outlier is a value that is unusually far from what normally happens in that specific situation.

Check our job opportunies

First, you have to know what “normal” looks like

To recognize the extraordinary, you need to know the ordinary really well. And the ordinary, in the world of transactions, has rhythm: activity follows predictable patterns by day of the week and by hour of the day. There are peak hours, there’s the calm of the early morning, there’s the difference between a Monday and a Sunday.

That’s why, to find out whether a moment was truly out of the curve, the comparison has to be fair: compare a Monday with other Mondays, and 2p.m. with other 2 p.m. periods. Comparing match time with the dead of night would tell us nothing. Comparing it with the same time on similar days does.

How do you measure an outlier?

To turn that intuition into something a computer, or an analyst, can actually use, statistics offers three simple ingredients.

The first is the average: the value we’d normally expect in that context. The second is the standard deviation. Put simply, it tells us how much values usually move above or below that average. A time slot with very consistent activity will have a small standard deviation; one that naturally varies a lot will have a larger one.

Then comes the z-score. It tells us how far a specific moment is from the average, measured in standard deviations. A z-score of 0 means the value is right at the average. A z-score of −1 means it is one standard deviation below the average. A z-score of −4 means it is four standard deviations below what we’d normally expect.

A useful rule of thumb is that about 99.7% of values fall within three standard deviations of the average. So once a value goes beyond three, we’re looking at something extremely unusual. That’s why “three standard deviations” is often used as a practical dividing line for an outlier.

Hold on to that “three”: it’ll come in handy in a moment.

The World Cup: an outlier with a known cause

A national-team match is the perfect example, for two reasons. First, the cause is obvious: everyone stopped to watch. Second, it’s on the calendar, so we know exactly where to look.

And the data doesn’t disappoint. During the Brazilian team’s match, the volume of instant transfers plunged to just over half of what you’d expect for that time of day, roughly four to five standard deviations below normal. Remember that going past three is already extremely rare? This is a textbook outlier.

And then there’s the detail that makes the chart especially fun to read: you can see half-time. Activity briefly rebounds in the middle of the match, as people take care of things before attention shifts back to the game in the second half. In its own quite way, the data tell the story of the match minute by minute.

Does it only happen in Brazil?

No. The same basic pattern appears when other national teams take the field, but with some interesting local variants.

In Mexico, the World Cup host, the drop was even sharper: activity fell more than twenty standard deviations below expected. If three is already extremely rare, twenty is an extraordinary result: the kind of reading that would be all but impossible without a very strong cause behind it. There, too, you can see the half-time breather and, as a bonus, a spike just before kickoff: that last-minute transfer to settle up the barbecue before the ball starts rolling.

In Colombia, activity dropped to less than half of normal during the match, and then surged well above average the moment the game ended, almost as if the whole country started breathing (and transacting) again at once.


Three countries, three different payment systems, the same human behavior. That gives us more confidence that the pattern is real rather than a coincidence: when something similar appears across independent contexts, it becomes harder to dismiss as a fluke.

OK, but what’s the point of spotting outliers?

Beyond making for a fun chart, spotting outliers is one of the most important jobs for anyone who works with data, and it connects to several things that affect our daily lives.

•  Security: the same basic reasoning can be applied to the behavior of a single customer: what normally looks typical for this person? If a transaction is very different from someone’s usual pattern, it can raise a flag for a closer look. Technology and teams of people can then step in to help protect the customer. Detecting what falls outside the usual pattern is a core part of fraud prevention.

•  Keeping systems running: understanding when activity rises or falls helps size the infrastructure so everything keeps working, including that instant after the goal, when lots of people go back to transacting at once.

•  Data quality: sometimes an outlier isn’t a real event at all, but a measurement error. Spotting it helps prevent that bad measurement from leading to the wrong conclusion.

And there’s a subtle point: not every outlier is a problem. The match is an outlier for a clear, explainable reason. The job of the people who look after data is precisely to separate what’s expected and explainable, like the World Cup, from what deserves a closer look, like an unusual transaction with no obvious explanation. 

The statistical tools we’ve seen here — the average, standard deviation, and z-score — provide a simple first way to make that distinction.

What the World Cup example shows

A country pausing to cheer gives us a surprisingly clear way to see a serious idea in action: first understand what “normal” looks like, then notice when something falls far outside it, and finally ask why.

Sometimes the answer is a World Cup match. In other situations, the same principle can help protect accounts, keep systems reliable, and make sure decisions are based on trustworthy data.

That may be the most interesting thing about an outlier: spotting one is only the beginning, and what really matters is understanding what caused it.

How we did this analysis

Every figure in this article uses aggregated and anonymized data. No individual customer information is used or exposed. By design, the analysis is presented only in relative, statistical terms, never in absolute values.

Check our job opportunies