The Empty Spreadsheet: When an Esports Analyst Has to Say 'Insufficient Information'

At three in the morning in Los Angeles, I opened the output file my analysis pipeline had just returned. Dozens of fields had been defined in advance: game title, patch number, tournament name, format, team, player, coaching staff, time sensitivity, source quality. All of them blank. Not a single information point extracted, not a single entity named.

Six years of working with sports data had taught me to live with bad spreadsheets: missing columns, timezone drift, duplicate records, inconsistent units. A completely empty spreadsheet is a different animal. An empty spreadsheet is data in its own right.

My first xG spreadsheet taught me that every goal has a hidden story. It also taught me the reverse: a claim with no measurement behind it is nothing but a guess. That night I had two options. Write the full nine-dimension analysis and fill every blank with reasoning that sounded entirely plausible. Or write only what could be written: that there was, for now, nothing to say.

The esports analysis industry picks the first option far too often.

Why a blank cell is worth more than a good paragraph

A serious esports analysis workflow runs in two stages. Stage one decomposes the raw article into discrete information points: game title, patch, tournament, team, player, timestamp, source. Stage two takes that input and only then runs the nine-dimension deep analysis — patch and meta, tournament format, roster and form, regional landscape, club finance, rules and governance, risk profile, public narrative, and industry transmission.

Those nine dimensions do not stand alone. Every one of them is anchored to stage-one information points. Without an anchor, stage two is an empty frame shaped like professionalism.

The first thing an empty frame cannot answer is the simplest question of all: which game is this. It sounds like a formality, but it determines the entire metric system behind everything else. League of Legends runs on a patch cadence of roughly two weeks, where pick-ban rates and champion win rates move fast. Dota 2 moves by season, with large patches that can upend how the game is played. CS2 ships smaller updates that feed directly into in-game economy and weapon value. Valorant runs by act alongside regional competition cycles. Applying the wrong metric system to a title means building a correct equation around the wrong variable.

Vietnam is a clear illustration of how metrics depend on the title. In League of Legends, the VCS was for years one of the regions with a stable World Championship slot, with teams such as GAM Esports and individuals such as Levi (Đỗ Duy Khánh) and SofM (Lê Quang Duy), who reached the 2026 World Championship final in Shanghai with Suning and lost 1-3 to Damwon Gaming. In Arena of Valor and CrossFire, the meta instead revolves around domestic publisher update cycles and the Southeast Asian competition calendar. Three titles, three metric sets, three entirely different ways of reading a match.

Nine dimensions, and what each costs when data is missing

Start with the patch, because it is the root variable. A patch is only analytically useful when it answers three questions: which mechanic changed, who benefits, who suffers. Answering them requires numbers. Pick rate, ban rate, win rate, average match duration, share of matches ending before the twentieth minute. Without those series, the sentence 'this patch favours early aggression' is just a sentence.

The second dimension is tournament format. It is the most underrated and also the most verifiable. A BO1 bracket carries a far higher upset rate than BO3, and BO5 all but eliminates the value of a single explosive performance. Bracket shape, seeding, rest days between rounds, and the moment a new patch is introduced between groups and playoffs — all of these are quantifiable variables. But to quantify them, you first need to know the tournament's name.

The third dimension is team and player. This is where data discipline breaks down most easily. Paper strength, role fit, chemistry, bench depth, week-by-week individual form. In football I measure defence through PPDA and line height. In esports I measure it through opening-kill creation, kill participation, gold-to-damage conversion, and stability across knockout matches. Both have one thing in common: a three-match sample says nothing.

The fourth dimension is the regional landscape. The LCK, LPL, LEC, LCS and the regions behind them differ sharply in roster depth, academy output, and import flows. Saying one region is stronger than another without numbers is a meaningless sentence. The strength of a region lies in whether its fourth- and fifth-best teams can still beat another region's champion. Depth is the index, not the peak.

The fifth dimension is finance. Sponsorship revenue, publisher distributions, salary spend, capital injections. Salary-to-revenue ratio is one of the earliest indicators of a club heading the wrong way, and it usually shows up before the standings do. Since 2026, when major leagues moved to franchising with buy-in fees reported in the tens of millions of dollars, an esports team's financial problem stopped being the problem of a group of friends who happen to be very good at a game.

The sixth dimension is rules and governance. This is the dimension where data cannot save anyone if the underlying event is missing. Esports history carries deep scars. In 2026, South Korean police arrested a group of StarCraft: Brood War players, including Ma Jae-yoon, on match-fixing charges — a case that shook the credibility of an entire national esports scene. In 2026, a match involving team iBUYPOWER at a CS:GO event became the centre of allegations of deliberately losing for betting purposes, and in January 2026 Valve banned the players involved from all Valve-run events. Events like these can only be analysed with names, dates, and disciplinary documents attached. Without them, every comment on competitive integrity is speculation.

The seventh dimension is the risk profile. Competitive risk, financial risk, personnel risk, rules risk, public-opinion risk, systemic risk. Each needs a probability and an impact level. A risk matrix with no subject to attach it to is an empty table with a header.

The eighth dimension is public narrative and the expectation gap. This is my favourite, because it is pure arithmetic. A team that wins three straight matches through three late-game comebacks will be described by the media as 'finding form'. A model says something else. Crowd expectations rise faster than actual quality, and that gap is the thing worth measuring. When the narrative runs hotter than the fundamentals, that is a signal to sell, not to buy.

The ninth dimension is industry transmission, running from publishers upstream, through clubs and streaming platforms midstream, down to sponsorship, derivative products and mainstream reach downstream. A policy change upstream can take several seasons to reach a mid-table team. That chain can only be drawn when at least one named link exists.

Nine dimensions. Every one has a frame. Not one of them has content.

When honesty looks like incompetence

There is a professional pressure few people talk about. A report that comes back with every cell blank reads as a sign that the analyst is not good enough. A report full of prose, full of terminology, full of team names and champion names, reads as a sign that the analyst is a professional. Between the two, the second always wins with a skimming reader.

The second report has a problem. If stage one extracted nothing, then every entity appearing in stage two was generated by the model itself. Team names, player names, patch numbers, win rates — all of them can be produced with very convincing confidence. And one fabricated entity at stage two contaminates all nine dimensions above it, because every dimension is anchored to stage one.

I have been on the other side of this problem. In 2026, assessing a target striker for a mid-table club, my model showed he had scored 4.5 goals fewer than expected across a season. The cause lay in bad luck, not in a decline in form. The club signed him and he scored on the opening matchday. The model was right. That same period, I missed a deadline on a set-piece report because I wanted to push accuracy to a perfect level. A colleague told me something I still use: a model that is eighty percent right and delivered on time beats a perfect model delivered after the match has finished.

That is true about speed. It is not true about foundations. Delivering a data-backed model early is discipline. Delivering a data-free model early is fabrication.

I do not predict the future by intuition; I only read the traces numbers leave behind. The problem is that when no traces exist, reading slowly is the only meaningful action left. Every dataset is a scripture, and I am a slow reader — including when the scripture is blank.

There is a subtler temptation. Taking a model that worked in one field and applying it directly to another. I come from football, where xG and PPDA gave me two consecutive correct calls: the home-advantage model in 2026, when I gathered data from more than three thousand matches across Europe's five major leagues and calculated that home teams were handed roughly 0.38 goals per match by crowds — a figure that fell away sharply when the Bundesliga returned to empty stadiums; and Morocco at the 2026 World Cup, where defensive data spoke before the world was willing to listen.

When home stops being home, every assumption has to be rewritten. But that lesson does not transfer to esports automatically. Football and esports differ on the surface, yet the same layer of data sits underneath — and that layer only works when the context is correctly defined. Applying a football model to an Arena of Valor match without checking the similarity assumptions is another form of fabrication, politer but still fabrication.

The only thing that can be concluded with certainty from an empty result concerns the process, not the industry. An empty stage one usually comes from one of two causes: a broken extraction pipeline, or a source article that never contained substantive professional content — a paywalled stub, a link aggregator, or a brief too short to carry anything. Both are input problems, not market problems.

That night I did not write the nine-dimension analysis. I wrote a checklist: game title, entity name, information point with source, patch number, tournament format, source quality, time sensitivity. Seven lines. An esports article can only begin once at least the first line is filled.

If next season you read an esports analysis so smooth that nothing in it needs verifying, try one thing: count how many data points in it can be traced back to an origin. The result of that count will tell you whether the analysis is about the match, or about the person who wrote it.