Nine Layers of Esports Analysis: A Data Framework for Practitioners

Three in the morning in Seoul. The live stats panel on my second monitor suddenly went blank — the tournament's data server crashed right in the deciding game, moments before a major objective fight. On the main screen, the two teams were still throwing themselves at each other, but every number I relied on — stage win rates, resources per minute, expected kills, objective timings — vanished in a blink. What I saw over the next ten minutes turned out to be a bigger lesson than the match itself. The arena began to fill the gap. One caster shouted that Team A had transformed. Another insisted Team B had lost its nerve. On social media, thousands of people told the same story about willpower, fate, and a moment of transcendence. Not one of them was holding a single number. The moment the data disappears is the moment people start inventing mythology. When there is nothing left to count, people count with emotion. I have counted every gap on the pitch when the crowd fell silent — and in my profession, a gap in the data table is also data, except it is data about absence. A few years ago, I moved my entire workflow onto a two-stage structure. Stage one is extraction. Stage two is analysis. It sounds simple, but most of the mistakes I have seen in this field come from merging the two: people read data and draw conclusions at the same time, and the conclusion arrives before the data has even formed. Stage one has a single rule: no commentary. It records only what exists — match name, team names, patch version, timestamps, numbers, statements, actions. If a field is empty, I leave it empty and mark it clearly. I do not fill it with guesswork. This is what I learned from nights spent reading statistics after a big match: data is only trustworthy when the person recording it does not add a little drama of their own. Stage two is where I ask questions. But it can only begin once stage one has returned at least one concrete data point. Without a game title, without a team, without a patch version, every analytical framework is meaningless — you cannot pick the right frame of reference when you do not know which discipline you are talking about. League of Legends, Dota 2, CS2, VALORANT and Honor of Kings operate on radically different logics. Analyzing a Dota 2 match in the language of League of Legends is shooting yourself in the foot from the first sentence. Below are the nine layers that every deep analysis of mine must pass through. The order matters less than the completeness. Miss one layer and the conclusion can still be right — but it is right by luck, and in my world, luck is only the unexplained residual. PATCH AND META A patch does not merely change a few numbers. It changes the value of every decision in the match. A champion gaining a few percent damage can turn a forbidden strategy into a mandatory one; an item losing power can push an entire playstyle into a museum within a week. When I assess a patch, I ask five questions. Where is the meta heading — early skirmishes or long-term resource control? How large is the change — a small tweak or an upheaval? Who benefits and who suffers? What do win rates and pick-ban rates say, and at which rank bracket? And finally: does the team I am analyzing have a champion pool that fits the new meta? The trap is that patches always come with lag. On the tournament server, people often play an older version than the practice server. A team can look terrifying in scrims and collapse on stage, simply because they practiced on a different meta than the one that will be played. That is why I always separate the question of the tournament version from the practice version, instead of merging them into a single belief. TOURNAMENT FORMAT Format is the most underrated variable in the entire industry. Single elimination differs from round robin; the Swiss system differs from a classic group stage; a best-of-three series is entirely different from a best-of-seven. Every format choice is a choice about which kind of team survives. A short series rewards careful preparation and teams able to unleash a surprise strategy. A long series rewards tactical depth, the ability to read an opponent across multiple games, and mental endurance. A team strong in a narrow champion pool but excellent at one composition can win a short tournament, then exit early from a long one. The same team, the same roster, two opposite results — purely because of format. The schedule is also part of the format. Match density decides whether a team has time to adapt to a new meta. Three matches in four days is entirely different from three matches spread across three weeks. I always log match density into the file before saying anything about form. TEAM AND PLAYER At this layer, I measure four things: paper strength, role fit, chemistry level, and bench depth. Paper strength is the summed individual value of the players. But the strongest roster on paper has never automatically been the strongest team on stage — a lesson anyone who has watched enough big matches knows by heart. Role fit matters more than reputation. An individually brilliant player who does not suit the team's tempo drags the whole group down. Chemistry is measured by shared hours and the stability of the starting lineup. Bench depth determines whether a team can survive a long season. For each key player, I draw a form curve over time rather than judging from one match. An individual posting high scores in three straight games may just be the effect of a weak opponent. An individual in decline may simply be playing under a meta unfavorable to their skills. I do not believe in inspiration — I believe in standard deviation. In other words, I only trust a trend once it has cleared the random fluctuation band of its own noise. REGIONAL LANDSCAPE A team does not exist in a vacuum. It exists within a region, and that region has a rank. I compare regions on four indicators: international results, talent pool, academy output, and ecosystem health. International results are the crudest but most honest measure. The talent pool tells you a region's future, not just its present. Academy output shows whether a region is feeding itself or buying in. And ecosystem health — number of stable teams, number of tournaments, level of investment — shows whether that region is sustainable. The flow of transfers between regions is the earliest signal. When a region's young players start being bought up en masse by other regions, that is not good news about development quality — it is bad news about the ability to retain talent. Conversely, when a region keeps an entire generation of talent intact, it is usually preparing for a cycle of dominance. CLUB FINANCE AND BUSINESS This is the layer the crowd ignores most, and the one that decides long-term survival. I look at four lines: sponsorship revenue, publisher distributions, salary expenses, and capital injections. A team can win on stage and go bankrupt in the books. A salary bill growing faster than revenue is the sign of a balloon about to burst. Publisher distributions depend on the health of the entire league, not on that team's results alone. And capital injection from owners — if it is the only source of life — is a gamble, not a model. I pay particular attention to bad signals: unpaid wages, dissolution, team sales, sudden withdrawals. These events often surface late in the press but show early signs in transfer data. A team selling a core player mid-season without a clear competitive reason usually has a cash-flow problem. The bubble in young-player prices, with enormous fees for people who have never played enough top-level matches, is a form of naked gambling — and gambles of this kind usually end in a purge. RULES AND GOVERNANCE This layer checks five items: competitive integrity, transfer and registration rules, contract compliance, minor protection, and disputes with publishers. A case involving competitive integrity can wipe out an entire season for a team, no matter how strong it is. Transfer rules change with each game and each region, so a deal valid in one place can be a violation in another. Contract compliance is where quiet disputes happen. Minor protection is where ethics meets business reality. And publisher disputes are where real power is exposed. When analyzing an event at this layer, I always build three scenarios: worst case, middle case, and optimistic case. Not to predict accurately, but to know which direction I will have to update my model if things develop. RISK PROFILE I split risk into six groups: competitive, financial, personnel, rules, public opinion, and systemic. Each has its own probability, impact level, and mitigation. Competitive risk is a champion pool that does not match the meta. Financial risk is a salary bill exceeding revenue. Personnel risk is a key player losing form or motivation. Rules risk is an undisclosed violation. Public opinion risk is a wave of criticism that paralyzes team morale. And systemic risk is a change from the publisher that can reverse a team's entire advantage in a single update. The key point of this layer is not predicting which risk will occur, but knowing which risk would collapse my model if it did. Those are the risks I track most closely. PUBLIC NARRATIVE AND EXPECTATION Every team, every player, carries a story. That story may be true or false, but it always carries weight. My job is to measure the gap between the story and reality. I check three things: whether the story has a fundamental basis, whether the sample size is large enough, and how long the story is expected to last. A three-match win streak can create a comeback narrative, but a sample of three matches is not enough to conclude anything. When social-media coverage far exceeds the underlying data, I know I am looking at an expectation gap — and that gap usually closes in a painful way. This is where I stand against the crowd the most. Not because I enjoy confrontation, but because the numbers have not confirmed the story the crowd already believes. When the data table does not lie, my heart begins to listen. Before that, I stay silent. INDUSTRY TRANSMISSION The final layer places an event within the larger flow of the whole industry. The transmission map runs from the upstream — publishers, who decide patches, licenses and schedules — through the midstream of clubs, tournaments and streaming platforms, down to the downstream of sponsorship, derivatives and mainstreaming. A small decision upstream can create a large wave downstream. A schedule change that reduces viewership lowers sponsorship value, which cuts salary budgets. A controversial patch can push players away, eroding the talent pool. Conversely, a successful international tournament can lift an entire region to a new tier. I also watch the gray zone: betting and activities not tightly regulated. This is where systemic risk accumulates most quietly, and where fewest people are willing to look. WHAT I LEARNED FROM A BLANK DATA TABLE Back to that night in Seoul. When the data server crashed, I did not try to reconstruct the match from memory or feeling. I opened a blank file and wrote one line into it: not enough data to conclude. That was the most honest conclusion I could offer in those ten minutes. The irony is that very moment taught me more than any complete analysis. It showed me that the hardest part of the job is not reading data — it is keeping discipline when there is no data. The pressure to say something, to have a take, to appear knowledgeable, is the strongest and most dangerous pressure of all. I have seen analyses built out of thin air. They always look convincing, because the writer has filled every gap with plausible-sounding assumptions. A game title guessed at random. A team assigned generic traits. A conclusion drawn from a premise that does not exist. And because no one checks the premise, the conclusion spreads. The honesty of zero is the hardest thing to teach. It requires accepting that some questions cannot yet be answered, and that admitting this is not weakness but professionalism. In an industry where everyone wants to have an opinion first, the person brave enough to say 'not enough data' is often the most trustworthy over the long run. I built these nine layers of analysis not so that I always have an answer. I built them so that I know exactly what I am missing. When stage one is empty, I do not fill it with imagination. I keep the structure intact, mark every gap, and wait for real data to arrive. The structure still stands, ready to be filled. That is the value of a framework: it does not create truth, but it shows you where the truth is missing. WHAT TO WATCH NEXT With the regular season underway, the signal I watch most closely is not the standings, but the gap between expectation and data. When a team is celebrated after three matches, I reopen the file and check the sample size. When a team is buried after one loss, I check whether that loss truly reflects anything, or is just an unexplained residual. If you work in this field, try building the nine layers yourself for an event you care about. Fill in every layer. Then circle the cells you had to leave blank for lack of data. Those circled cells are your to-do list. In analytical work, what you do not know matters no less than what you know — and sometimes it matters more. Every match is a puzzle piece; I do not watch it, I decode it. But a missing piece cannot be replaced by an imagined one. You leave the space empty, note it down, and move on. That is the only way the final picture stays correct.