When Data Lies: Lessons from a Power Outage Notice Mislabeled as Tennis

A power outage notice from the Islamabad Electric Supply Company (IESCO) about scheduled grid maintenance in Islamabad and Rawalpindi on Monday, March 17, 2026, suddenly appeared in a tennis analysis system with the label 'Domain Label: tennis'. No players, no tournaments, no statistical indicators. Only substation names and outage windows from 9 AM to 5 PM. But the automated system mislabeled it. And that was the moment I realized: in the era of big data, classification errors are more dangerous than missing data. I have spent nearly three decades tracking matches from the A-League to the World Cup, from silent stadiums during COVID to the roaring stands of Melbourne Park. I learned that data never lies – but it took me ten years to know when it tells half-truths. An article about power outages is not a half-truth; it is a completely different truth placed in the wrong context. When the whole world looks at the goal, I look at the off-ball run. But when an entire system looks at an administrative notice and thinks it is tennis, I must look at the classification system itself. Let me take you inside the process. An article about IESCO – the electricity provider for Pakistan's capital – was automatically assigned to the tennis category. Perhaps due to the keyword 'match' appearing in 'maintenance schedule', or perhaps due to a machine learning algorithm confusing 'court' with 'tennis court'. Whatever the reason, the consequence is clear: without a cross-check step, this article would enter the tennis database, polluting analyses of player form, match schedules, and tournament trends. PPDA does not decode Croatia. It decodes the football Croatia hides within its patient shell. But a mislabeling algorithm decodes nothing except its own laziness. I remember 2026, when the A-League paused due to COVID and I lost full access to stadiums. While colleagues shifted to social commentary, I launched the 'ghost home stadium project': collecting data from 37 behind-closed-doors matches. I found that home win rates dropped from 49.2% to 41.3% in empty stadiums. The pandemic did not erase data. It stripped away the glossy paint and revealed the skeleton of the game. But that skeleton only has value if the data is verified. A power outage notice is not tennis data, no matter how it is labeled. The problem does not stop at one article. Imagine if 1% of articles in a system were mislabeled like this. In a season with roughly 3,000 ATP and WTA matches, that equals 30 'ghost' matches appearing in analysis. An analyst could miscalculate home win rates, misjudge serving trends, and even mispredict outcomes. I witnessed this in a small league: in 2026, while reviewing A-League GPS data, I noticed 18-year-old Daniel Arzani of Melbourne City averaging 4.6 successful dribbles per match – double the league average. I did not wait for rumors; I called the coaching staff directly and requested his full movement data across 12 rounds. When Celtic signed him in August 2026, I had a complete data profile from before his Melbourne departure. But had I relied solely on automated classification labels, I might have missed Arzani – or worse, analyzed an electricity article as if it were a match. The IESCO story teaches me a deeper lesson: data has no emotions, but the people who create data do. An algorithm does not feel shame when mislabeling. A system does not tire of repeating errors thousands of times. Only humans – data journalists – bear the responsibility to ask questions. I do not need to see how many matches they play. I need to see how many meters they run in a situation no one notices. Similarly, I do not need an algorithm to tell me which category an article belongs to. I need to read it myself, verify it myself, and decide myself what deserves to enter the analysis. Look at the numbers in the IESCO article: 9 AM to 5 PM, 11 substations, 3 affected areas. These numbers are perfectly accurate – for an administrative purpose. But when placed in a tennis context, they become garbage. This reveals a counterintuitive truth: accurate data in the wrong context is more dangerous than inaccurate data in the right context. A wrong xG figure can be detected and corrected. But an accurate figure about power outage schedules, mistaken for a form indicator, will silently corrupt every downstream analysis without anyone noticing. I have learned this through years of working with data. In 2026, at the World Cup, I calculated Croatia's PPDA against Argentina at 7.9 – meaning they allowed opponents fewer than 8 passes before engaging. My analysis proved Croatia reached the final through a deep-lying midfield system that shielded space, not through inspiration. The article sparked controversy, but weeks later UEFA's analysis department confirmed the numbers. I became the only pressing specialist in the Asia-Pacific region. But I never forgot that every number I publish must pass a test: where does it come from? How was it measured? What does it mean in the actual match context? Without clear answers, I discard the number – no matter how attractive it seems. The IESCO lesson is not just for automated classification systems. It is for all of us – data consumers. In a world overflowing with information, the most important skill is not finding data, but knowing how to doubt data. When a tennis article appears without player names, without scores, without tournaments – stop. When a number is too perfect, too aligned with the story you want to tell – doubt it. I have tracked the careers of hundreds of players, from Arzani to Pedri, and I realize that the most valuable discoveries often come from unexpected places. But they only have value if I rigorously verify every data source. The empty stadium of 2026 did not weaken players. It exposed the fake statistics once shielded by crowds. Similarly, a mislabeled power outage notice does not weaken tennis analysis systems. It exposes the verification gaps we often overlook. Data is cleaner than any interview – but only when placed correctly. When data is misplaced, it ceases to be data; it becomes noise. And noise, no matter how beautifully formatted, never becomes signal. So what must we do? I propose a simple but effective cross-check process: before feeding any article into analysis, ask three questions. One: does this article mention at least one specific tennis entity – a player, tournament, match, or organization? Two: are the numbers in this article directly related to tennis competition activity? Three: if I remove this article from the data, would my analysis change? If the answer to the third question is 'no', then the article does not belong in the system. It is that simple. I have applied this process since 2026, when I began publishing raw data with download links in every article. It drove away some sources, but I accept that. Because I know: data never lies – but it took me ten years to know when it tells half-truths. And a half-truth, like this IESCO article, is more dangerous than an outright lie.