Nine Dimensions of Esports Analysis: The Discipline of Foundation Before Interpretation

That night in Busan, the screen returned a blank page.

I had just run the two-stage pipeline on an esports file. Stage one was the extraction stage: find the title, the source, the article type, the core viewpoints, the list of information points, and the entities mentioned — the game title, the teams, the players, the tournament. Stage two is where I dissect across nine dimensions. But stage one came back empty. No title. No source. The article type was unclassified. The information-point list was blank.

The person next to me asked: “So what do we write now?”

I said: “Nothing.”

The room went quiet for about ten seconds. In those ten seconds I counted three ways to fill the blank page. The first was to guess. The second was to borrow an old file and swap the team names. The third was to write about the gap itself.

I chose the third, but not to save the piece. I chose it because after eleven years watching this industry, I learned something more valuable than any analytical skill: when the data does not exist, “insufficient information” is a complete conclusion. It is not a hiding place. It is a result.

The foundation is built first; the interpretation floor comes after

I was born in Vietnam, I work in Busan, and I make a living retelling matches through data. People usually call me a data journalist. I do not object, but I always think my trade has a more accurate name: the person who verifies the foundation before allowing anyone to build a floor on top.

That foundation began one night in Russia.

“On a Russian night, for the first time I saw a number that could hurt.”

In 2026 I was nineteen, a sophomore in Busan. On a World Cup night, I fed all twenty-three shots by Germany against South Korea into an expected-goals model I had written in Python. The screen showed a cold result: Germany generated 1.32 expected goals but scored none, and lost 0-2. I checked it against the highlights, and realized the naked eye had been fooled by the feel of the game. Eighteen of the twenty-three shots — about seventy-eight percent — came from outside the box. The defending champions were eliminated not by a miracle. They were eliminated by an unwise tactical decision, and the model had seen it before the match ended.

Two years later, empty stands taught me another lesson.

In the 2026 season, K League 1 became the first football league in the world to resume in front of no crowd. My 2026 expected-goals model started to drift. I collected 152 matches and found the home win rate had fallen from 46.2% in the 2026 season to 31.6%. I wrote a forty-page report concluding that every 10,000 spectators was worth roughly 0.08 expected goals for the home side.

“The 0.08 coefficient does not measure the silence; it measures what we lost.”

No one had asked for that report. But I knew that without fixing the foundation, every later analysis would be wrong.

In December 2026 I was twenty-three, a new hire at a newsroom. Thanks to the 2026 report, I was assigned to analyze an African national team that had reached the World Cup semi-finals for the first time. I compiled their three knockout matches: they conceded possession at 71.6% but shipped only one goal, while opponents generated 4.02 expected goals in total. The number that made me stop was a PPDA of 25.1 — nearly double the tournament average of 13.2.

“PPDA 25.1 — dropping deep is not a concession, it is stretching the field.”

Local media called it luck. My piece said the opposite: defending does not mean being passive. I replaced the phrase “pinned back” with “choosing to drop deep,” and I learned to defend a controversial argument with sources, sample sizes, and a published list of the model's limitations.

In 2026, that piece connected me with a sports-data company in Lisbon. From that source I found a Korean midfielder at a mid-table club who had played only 564 minutes the previous season, far below the 1,200 minutes written into his contract. I sent his agent a six-page metrics report. On June 8, 2026, I was the first to reveal the loan deal with a 2.8-million-euro buy option.

“A transfer fee does not measure talent; it measures the buyer's desire.”

The agent said they trusted me because I brought numeric evidence, not emotional judgments. From then on my transfer reporting had its own logic: hypothesis, data, source, probability. I dropped vague phrases like “a dip in form” and replaced them with “minutes played down 41% on last season.”

When I moved to covering esports for the Korean market, I carried that entire discipline with me. But I quickly realized football and esports do not share a frame of reference. In football, the rules have stood still for nearly a century. In esports, the rules change every few weeks through patches, and a publisher can rewrite the competitive environment with a single line in the notes. An expected-goals model cannot be applied directly to a match in which the strength of each character also depends on the server version.

That is why I built a nine-dimension framework. It is not a formula for producing answers. It is a checklist for knowing what I am missing before I dare say anything.

Dimension one: patch and meta

Every esports analysis must start with a question that sounds trivial: which game is this. Without identifying the game, you cannot choose the right analytical lens. League of Legends, Dota 2, CS2, Valorant, Honor of Kings — each runs on its own logic, and a metric that is good in one can be meaningless in another.

Once you know the game, you need the version. A patch may only tweak a few numbers, or it may change a mechanic outright, or it may rework a character from scratch. Those three magnitudes do not carry the same weight. A small numeric tweak usually shifts win rates by a few percent; a mechanic rework can invert the entire priority order in the draft.

“Every meta update is a confession by the publisher.”

When a publisher nerfs a dominant character, they are admitting the previous environment was unbalanced. When they buff a forgotten one, they are admitting that dead choices existed and that they created them. A patch is not just a bug fix. It is a memo about design mistakes.

In this dimension I always set four groups of numbers side by side. Win rate tells you whether a choice is effective, but it is noisy because it is affected by player skill. Pick-ban rate tells you what the professionals fear, and it is usually the most patch-sensitive indicator. Average match duration tells you whether the meta's tempo is speeding up or slowing down. And win rate by game phase tells you whether that strength arrives early or late.

A patch that only changes numbers makes win rate move slowly while pick-ban rate moves first. A patch that changes mechanics makes both move almost at once, and usually brings in strategies no one has practiced. This is where I part ways with most mainstream commentary: the crowd looks at win rate after the tournament ends, while I look at pick-ban rate before it begins.

The last part of this dimension is the fit between a patch and each team. A team with a wide character pool absorbs a patch better than one that is only good at a few choices. But a wide pool does not automatically convert into points. It only lowers the probability of collapse when the patch pivots. I have seen teams keep the same roster across many patches and fall behind, not because anyone played worse, but because the whole group lost the one thing it was best at, all at once.

Dimension two: tournament system and format

Format is the most underrated variable in all of esports analysis. People remember who won; few remember the format they passed through.

The format type determines the probability of upsets. A single-elimination bracket pushes the odds of a weaker team advancing very high, because all it takes is one well-timed match. A double-elimination bracket lowers those odds, because the stronger team gets another chance to correct itself. A Swiss format produces more evenly matched pairings but also rewards consistency across many rounds.

Series length is also a variable. A best-of-three is very different from a best-of-five. The longer the series, the greater the edge of a team with tactical depth, and the smaller the edge of a team with a single elite strategy. This is why many teams that look excellent in a short-format group stage collapse in a long-format knockout stage.

The qualification path also belongs on the table. A team that enters directly through a slot has a different practice window from one that has just finished a regional qualifier. That gap does not show up in the standings, but it shows up in the players' legs.

Schedule density is the last variable, and the one I watch most closely. A team playing three matches in four days enters the next one with different reflexes from a team that rested a week. I do not measure this by a feeling of tiredness. I measure it by decision time in late-game teamfights, and by the rate of positioning errors after the thirtieth minute. Both indicators worsen when the schedule is dense, and both worsen faster for teams with thin rosters.

When a tournament reforms its system — for example, moving to a franchise model or changing how slots are allocated — the consequences do not stop at the bracket. They change the financial incentives of an entire region, and therefore change how teams build their rosters. Format analysis that ignores money is half-finished analysis.

Dimension three: teams and players

This is the dimension closest to readers, and the one most easily led astray by emotion.

I split it into four layers. The first is paper strength: total individual skill if everyone peaks at the same time. This almost never materializes, but it is a benchmark. The second is role fit: a good player placed in the wrong role performs below their ability, and that is not their fault.

The third is chemistry. This is the hardest layer to measure. I do not measure it through player statements. I measure it through the speed of information flow in teamfights: the time from when an objective is attacked to when the whole team responds. A well-drilled team responds almost instantly; a newly assembled team responds slowly and often in the wrong direction.

The fourth is bench depth. A team with six capable players is different from one with five capable players and one who has to carry. Depth does not help you win ordinary matches, but it helps you survive a dense run of games.

For each player, I track the form curve over time, not over one match. This curve usually has three phases: a fast rise when a player enters a favorable meta, a plateau when opponents learn to counter, and a fall when the patch pivots. Knowing which phase you are in matters more than judging who is playing well.

Based on my experience watching matches, the most common media mistake is attributing a losing streak to mentality. Mentality cannot be observed directly. What can be observed is the number of lost fights at pivotal moments, the rate of wrong objective calls, and the moment a team loses map control. Those things have numbers, and those numbers usually point to a tactical problem, not a spiritual one.

Coaching also belongs to this dimension. A complete coaching staff includes someone for strategy, someone for opponent analysis, and someone for fitness and psychology. A team missing one link tends to expose it in the knockout stage, when opponents have studied them thoroughly and there is no room left for improvisation.

Dimension four: the regional landscape

Esports is a sport in which geography still shapes playstyle, even as international competition flattens differences.

I divide regions into tiers. Tier one is regions with stable international results and a closed talent pipeline. Tier two is regions with a few teams strong enough to compete but lacking overall depth. The remaining tier lives on wildcard slots and occasional individual explosions.

When assessing a region, I look at four indicators. International results are the clearest but also the most misleading, because they depend on how many teams the region sends. The talent pool is the base indicator: how many capable players exist across age brackets. Academy output is the long-term indicator: whether the region develops its own talent or imports it. And ecosystem health is the composite indicator: how many teams pay wages on time, how many tier-two events are still alive.

Talent flow between regions is a signal I track over the long term. When a region starts importing many foreign players, that is usually a sign its domestic development is stalling. When a region starts exporting many young players, that is usually a sign its internal strength is high but domestic opportunity is narrow.

The playstyle gap between regions is narrowing, but it has not vanished. Different regions still have different habits of tempo, of map-resource usage, and of risk tolerance. These differences decide international matchups more than people think, especially in long series.

Dimension five: club finance and business

This is the dimension fans read least and misunderstand most.

A team's revenue structure usually has four sources: sponsorship, distributions from the publisher or tournament organizer, media rights and merchandise, and capital from the owners. These four have very different stability. Sponsorship depends on results and media heat. Publisher distributions depend on policy. Merchandise depends on fan scale. Owner capital depends on the confidence of a few individuals.

When a team spends beyond its means, the signs usually appear in small places before they appear in large ones. Late wages. Cut analysis staff. Cancelled bootcamps. These signs rarely make the news, but they are earlier and more accurate indicators than any transfer story.

For every deal, I separate two questions. First, does the contract figure match competitive value. Second, does the contract structure create risk, such as a mandatory buy clause or an excessively long term. An expensive contract is not necessarily a wrong contract. A cheap contract is not necessarily a smart one.

What interests me most in this dimension is the relationship between the signing race and long-term financial health. Big teams spend to protect brand position, and that spending is sometimes aimed at maximizing attention rather than points. Small teams, by contrast, are forced to spend to maximize points, because they have no brand to sell. The genuinely valuable contracts usually sit in the second group, where every dollar spent must be repaid in results.

Dimension six: rules and governance

Esports has a feature traditional sports do not: the publisher is at once the referee, the stadium owner, and the legislator.

The rules bearing on a team usually stack in three layers. The first is the game publisher's rules. The second is the tournament organizer's rules. The third is the national law of the host country. These three layers do not always align, and when they conflict, the team that suffers is usually the smallest one.

In this dimension I check five groups of issues. Competitive integrity, including match-fixing and betting. Transfer and registration rules, including deadlines and conditions. Contract compliance, including disputes between players and teams. Protection of minors, an increasingly hot issue as professional debut ages fall. And controversies over publisher governance.

For each suspicion, I build three scenarios: worst case, middle case, and optimistic case. The worst case is usually not a penalty but silence. An unexplained decision does more damage than a wrong decision that is clearly explained, because it destroys faith in the consistency of the rules.

The biggest blind spot of this dimension is the tendency to change rules mid-season. In traditional sports, the rules are nearly fixed within a season. In esports, a publisher can change the rules with a patch, and sometimes the change is announced only weeks before a major event. That is systemic risk, and it appears in no standings table.

Dimension seven: risk profile

Only after passing through the first six dimensions do I have enough material to build a risk profile. I divide risk into six groups: competitive, financial, personnel, rules, public opinion, and systemic.

Competitive risk includes dependence on one player, insufficient chemistry, and vulnerability to a specific strategy. Financial risk includes a broken funding chain, sponsor withdrawal, and falling slot value. Personnel risk includes injury, internal conflict, and losing a head coach mid-season. Rules risk includes penalties and sudden rule changes. Public-opinion risk includes fan pressure after a losing streak. Systemic risk includes the lifecycle of the game and the publisher's pivot strategy.

For each risk I record four things: level, probability, impact, and mitigation. Mitigation matters more than the rest, because it turns a list of worries into a plan. A team with bench depth is mitigating injury risk. A team with diversified revenue is mitigating financial risk. A team holding many strategies is mitigating the risk of being countered.

What I always remind myself in this dimension is not to turn risk into destiny. Risk is probability, not prophecy. A high-risk team can still win it all. A low-risk team can still be eliminated. My job is to describe the probability distribution, not to pre-select the outcome.

Dimension eight: public narrative and expectation

Every team, every tournament, has a story woven around it. That story has its own power, and the analyst must read it without being led by it.

Common stories include: the new king, the long dynasty, the all-domestic roster, the revenge run, and the last dance of a generation. Each story is anchored to a real event but inflates the meaning of that event.

I judge a story's durability with three questions. Does it rest on a real foundation, or only on one unusual match. How large is the sample behind it. And how long can it sustain itself before it needs a new event to keep existing.

The most important part of this dimension is the gap between market expectation and objective assessment. When a team is expected too much of, even a good result can be read as a failure. When a team is underrated, even an ordinary result can be read as a miracle. That gap does not appear in the standings, but it decides how a result is retold.

I regularly cross-check social-media heat against the data foundation. When the two diverge too far, it is a sign of an expectation bubble. A bubble does not burst because it is wrong. It bursts because an ordinary event happens at a moment when many people are watching.

Dimension nine: industry transmission

This last dimension places everything on a larger map.

Upstream are the game publishers: they decide patches, event calendars, and event licensing. Midstream are the clubs, the tournament organizers, and the streaming platforms. Downstream are sponsorship, derivative products, and esports' entry into the mainstream.

A change upstream propagates downward very quickly. A patch that makes a playstyle less appealing can make viewers watch less, which drags ad revenue down, which shrinks team budgets, which slows transfers. The chain is long, but it is real, and it runs from a line in a patch note to a layoff decision in an office half a world away.

I also track the gray zone. Betting and event-adjacent businesses grow faster than regulation. This is where competitive-integrity risk concentrates most, and also where data is hardest to obtain. I do not conclude without data, but I flag it for tracking.

Industry-level events — for example, a regional multi-sport games adding esports to its official program, or a large capital inflow into a region — all leave transmission traces. Those traces usually appear first in slot values, and only later in roster quality.

The contrarian angle: correlation is not causation, and the trap of ritual

The nine-dimension framework has a trap, and I have to be blunt about it.

That trap is ritualization. When a framework is used enough, its users start to believe that simply filling in the boxes produces a conclusion. That is not so. A framework full of data but with the wrong interpretation is more dangerous than an empty one, because it gives a weak conclusion the appearance of certainty.

“Before arguing about wins and losses, I have to question the numbers first.”

But after questioning them, you still have to answer another question: is this number causal, or merely correlational. A character's high win rate does not prove the character is strong. It may simply reflect that the best players choose it. This is a form of selection bias, and it appears everywhere in esports analysis.

Another trap is small samples. A team winning three in a row does not form a trend. A player posting high numbers in two matches does not prove improvement. I have set myself a rule: no trend conclusion under ten matches, and no player conclusion under five hundred minutes played.

The third trap concerns my own trade. Data analysts are moving deeper and deeper into the locker room. Teams hire opponent analysts, model builders, and player-metric specialists. But there is a gap between the model and the actual rhythm of a match. A model says to take the big objective now, while the player in the headset is hearing the enemy gank incoming and must decide within two seconds. The model does not hear that.

A data analyst's conclusion is often detached from the actual rhythm. This is what I keep in mind whenever I present a report to a team. I do not issue directives. I present a probability distribution, along with published error bars and limitations. The decision-maker is still the person inside the match.

And the last trap, the one I met again that night in Busan, is the temptation to fill the blank. When a data cell is empty, the writer's instinct is to fill it. Fill it with inference. Fill it with memory. Fill it with something that sounds reasonable. Every time you do that, you trade long-term credibility for short-term completeness.

I chose a minimum-viability gate. Before the deep-analysis stage is allowed to run, the file must contain at least one game title, one entity, and one information point. Without all three, the process stops. Stopping is an action, not a surrender. It keeps later analyses from being contaminated by entities I invented myself.

People ask whether I fear that this discipline makes me slower than others. I do not. I fear the opposite: faster, but wrong, and wrong with full confidence.

Conclusion: signals for the next cycle

The blank page that night in Busan turned out to be the most useful document I kept that month.

It was a negative control. A sample showing that my process works correctly when there is nothing to analyze. It proved that those nine dimensions are not a fabrication machine, but a machine that refuses to fabricate.

Looking toward the next competitive cycle, there are three signals I will watch closely. First, the speed at which patches are adopted across regions. The gap in when teams grasp a new meta often decides international matchups more than differences in individual skill. Second, talent flow between regions, especially young players going abroad early. Third, the financial health of mid-table teams, where early signs of budget cuts appear before they surface in the transfer market.

And one more signal, less related to numbers. More readers are asking me why I do not issue predictions right after each match. I answer that I need time to build the foundation before building the floor. If readers start to grow comfortable with gaps, it means they have accepted that the completeness of an analysis lies not in answering every question, but in answering only the questions it has enough evidence to answer.

“I do not write about football. I write about the light that data illuminates.”

On some nights, that light shines on an empty space. My job is to tell readers that there is nothing there, and to explain clearly why I know. That is the hardest part, and also the most worth doing.