Missing Data: When a Sports Analyst Must Learn to Say 'I Don't Know'

The first number in every report I write is usually the comfortable one. But there was one time when the only number I had was zero. No article title, no source, no data point, and the verification-timestamp field read 'not assessed.' A nine-dimension analysis with a complete template, yet not a single cell held real information. For years I had told readers that every number knows how to speak if we are willing to listen. That evening taught me the opposite: when the number goes silent, the analyst must go silent too, or say plainly that he does not know. The esports analytics industry is growing faster than the discipline of verification. Every streaming platform, every regional tournament, every new patch generates thousands of data points: pick rates, side-based win rates, creep score per minute, objective control time. Behind those numbers sits a question rarely asked aloud: if the data point does not exist, what will the analyst do? In my working environment, the bridge between the Vietnamese and Malaysian markets, missing data is not the exception; it is the norm. A match at a tier-two regional event may not be fully captured by any commercial system. A champion balance update may ship without official patch notes. A team may change head coaches without a press release. In those situations, the data gap becomes part of the picture, and how an analyst handles that gap determines the entire value of the conclusion. I call the three degrees of data absence the three layers of silence. The first layer is missing data: there is a subject, there is a question, but no figure yet. The second layer is unverified data: there is a figure, but its provenance is unclear. The third layer is unassessable data: the subject is so vague that even asking the question is meaningless. Most mistakes in sports analysis do not come from the first or second layer; they come from quietly filling the third with guesswork. Imagine a match where you have the pick rates of both teams but no data on each player's time in game. You can build a very persuasive story about Team A winning through a meta-appropriate lineup. But that story rests on an unverified assumption: that Team A won because of strategy, not because Team B lost a key player in the tenth minute. If you have no data on the lost player, you are selling a conclusion you have no right to own. This is where the concept of null-value handling becomes important. In data analysis, an empty cell is not zero. Zero is an observation: 'this team did not score.' An empty cell is an admission: 'I do not know whether this team scored.' These two lead to completely different conclusions. A serious analyst must distinguish between them and state clearly which case applies. When I built my xG model across five Bundesliga seasons from 2026 to 2026, I calculated from more than twelve thousand shots. Some matches had incomplete shot-location data. I had two choices: interpolate from similar matches, or mark the cell empty and drop that match from the sample. I chose the second. The result was a smaller sample, but every remaining figure stood firm. Had I chosen the first, I could have had a larger sample and a prettier conclusion, but that conclusion would have been built on sand. The same holds for esports. When a patch changes a champion's power, the first question is not whether that champion is now strong or weak, but whether I have enough data to answer. If the patch is three days old, any win-rate conclusion sits in the first or second layer of silence. The analyst has a duty to say so, rather than offer a number that looks certain. This duty is not timidity. It is a form of professional courage. Drawing a conclusion without data is easy; saying 'I don't know' when a conclusion is demanded is hard. In a market where every article can become a regulation order, saying 'insufficient information' is not a failure; it is the only way to protect readers from a wrong decision. But there is a flip side I must admit. Some people turn the admission of gaps into a defensive profession: every question answered with 'we need more data,' every conclusion postponed, and in the end no judgment ever delivered. That is another failure, no less serious. Missing data does not mean data cannot be analyzed. It means the analysis must be placed in a clear context, with the confidence level stated transparently. A good analyst can say: with the data available, I lean toward this possibility, at a low confidence level, and here is what I need to raise that confidence. That is a responsible judgment, entirely different from avoiding judgment. The paradox is that the very ability to make a judgment under imperfect data is the most valuable skill. Anyone can read a complete table of numbers. But the ability to see where the data is still missing, and to reach a conclusion while remaining aware of its limits, is what separates an analyst from a copyist of statistics. Before believing your eyes, check what your eyes have already believed. I wrote that line for matches, but it holds for data tables too. A beautiful column of numbers can make us forget that behind it lies an unverified source. A smooth chart can hide a sample too small to conclude from. Numbers never panic; people panic, and they are the variable. In sports analytics, that panic usually shows up as a conclusion delivered too early. There are two things that never lie: data and time. But both only refrain from lying when the analyst is honest about what he does not know. A report that admits its gaps is not a failed report; it is a report that has fulfilled its function. The next matchday will bring new tables, new patches, new teams. The question I ask myself each morning is not whether I have enough data, but whether I have been honest with the gaps. In an industry where trust is the greatest asset, the only thing worse than having no answer is a wrong answer presented as truth.