The Spreadsheet Doesn't Know How to Bow: The Paradox of Esports Data Analysis

I still remember a night in the middle of the transfer window in Incheon. On the table lay a three-page spreadsheet: KDA, damage per minute, gold difference at minute 15, win rate when laning alone. An analyst tapped his finger on the last column and said: "This is the optimal option, statistically." The head coach stayed silent for a long time. He pushed the papers back and said only one sentence: "This number doesn't tell me whether, when we're down 0-2, he'll still dare to look his teammate in the eye." That night I stayed until nearly dawn, not to recheck the spreadsheet, but to understand why two intelligent people looking at the same data see two different stories. Twelve years of hosting broadcasts and writing about esports taught me one thing: our industry owns more data than ever before, yet understands people less than we think. For nearly a decade now, the data-analysis department has become an inseparable part of every professional esports organization. In South Korea, where I live and work, almost no team in the top league enters a match without at least one analyst sitting backstage. Seven or eight years ago, that role was a luxury; today it is the minimum standard. That professionalization brings undeniable benefits. Thanks to data, teams detect holes in the draft phase early, track the trends of each patch, and quantify what used to be the coach's intuition alone. During the transfer window, data helps clubs filter out the noise: a player with pretty numbers in a second-tier league but who falls out of rhythm against stronger opponents will gradually reveal himself across a large enough sample. But precisely because data has become widespread, it has also become dangerous in a new way. When everyone has numbers, people begin to believe that having numbers is enough. Let us start from the most basic thing an esports analyst must do before opening any spreadsheet: clearly identify which game is being discussed. It sounds obvious, but it is the point that countless social-media debates skip. You cannot compare the metrics of a tactical fighting game with those of a first-person shooter. Each game has its own metric system, its own tournament system, and, most importantly, its own governance regime on the publisher's side. Skip this step, and every comparison that follows is meaningless. A metric only means something when it is placed correctly within its frame of reference. KDA is the easiest measure to read and also the easiest to be misled by. A mid-laner with a high KDA may simply be playing safe, ceding every key teamfight to his teammates and then collecting kills at the end. Conversely, a player with a low KDA may be the one who opened three decisive fights with dives that no statistics table records. Data is not wrong; data simply answers exactly the question that is put to it. This is why I always remind myself: I do not write KDA, I write the heartbeat behind the numbers. The same holds for every more complex measure. Damage per minute does not say when the player dealt that damage — before his teammates arrived, or after the match was already decided. Win rate does not say how weak the opponent was. And above all, every metric is born inside a system: a patch, a meta, a schedule, a specific tournament format. Remove the system from the picture, and the number becomes a floating value — technically correct but tactically meaningless. Even the performances considered canonical for Faker cannot be reproduced merely by copying the columns of statistics. One curious thing is that the very differences between games create very different transfer ecosystems. In games operated under a closed-league model, the flow of players is governed by franchise slots and salary budgets. In games with an open circuit, players have more freedom to move, but also more precariousness. A good analyst is one who knows which ecosystem he is standing in before comparing anything at all. And behind it all lies the publisher's governance. Each major publisher has its own set of rules on transfers, contracts, and player age. A deal valid in one system may violate another. Ignore that layer of rules, and any financial analysis can collapse after a single announcement from the organizers. Based on my experience watching matches, I once saw a very elaborate analysis, presented neatly with full charts and comparison tables, that nevertheless failed to specify which game was being discussed. As a result, the entire analytical framework collapsed at the very first step: no game, then no correct metric system, no correct tournament system, no correct governance regime, and finally no trustworthy conclusion. That analysis was full of words yet empty of substance — a reminder that formal rigor never substitutes for a solid foundation of information. There is one memory I keep as an example. In the LCK final on April 25, 2026, when the league had to play without spectators amid the pandemic, I organized an online watch party for students with about one thousand two hundred participants. T1 beat Damwon Kia three to zero, but what kept me thinking was not the scoreline. In a match whose statistics table records only kills, I noticed how ShowMaker of Damwon Kia used vision to control the rhythm of the game — something almost no metric captures in full. Read only the three-to-zero result, and you miss the entire tactical story behind it. But there is one thing I increasingly distrust, even as it is celebrated everywhere: the belief that data can replace human judgment in the locker room. In recent years, analysts have pushed ever deeper into territory that once belonged to coaches and players. They bring charts, predictive models, and conclusions presented as if they were truth. But I have seen far too many data conclusions that were statistically correct yet humanly wrong. The problem is that data is generated on the back end, while the match is decided on the front end — by things that never enter the spreadsheet: a bowed head, a second of hesitation before diving into a fight, a glance at a teammate after losing a skirmish. Those signals appear in no statistical column, yet they are often the decisive variable. A team can win every metric and still lose the match, because no spreadsheet measures a mental collapse at the thirtieth minute. There are applause no one hears that ring louder than a stadium. Those are the moments that, if you read only the data, you would call an ordinary match. I once wrote about a young player having the worst performance of his career, harshly criticized online, and instead of exploiting the failure, I looked into the pressure he had to carry. Had I looked only at the metrics, I would have missed the person. Had I written only about the numbers, I would have written a report as dry as a market report. I am not saying to throw data away. I am saying that data is a poor servant when it is placed as the decision-maker. The industry's mistake is not in using data, but in forgetting that behind every play is a person carrying an entire world of his own. This transfer window will again be full of numbers. There will be contracts praised for pretty stats, and deals criticized from day one. But when the spreadsheet is folded and the meeting-room lights are off, what remains on the stage is not columns of numbers, but people. What is worth keeping is not what the data says, but whether that data is seeing the person.