Pricing an Esports Player: Four Data Layers and One Empty Report

A perfect skeleton wrapped around nothing

In August, a twelve-page analysis file landed on my desk in Chicago. Its structure was beautiful in a way that makes people in my trade jealous: patch and meta analysis, tournament format analysis, roster and player analysis, regional landscape, club finance, rules compliance, risk profile, public narrative, and finally industry transmission. Nine sections. Each one had a table, an assessment column, and a confidence line.

I read it top to bottom. Every field said the same thing: insufficient information. No patch name, no tournament name, no team, no timestamp, not a single measurement. A skeleton carefully engineered to hold up something that does not exist.

I should have sent it back with a complaint. I did not, because in fourteen years of tracking transfer markets I have never seen a document describe this market so honestly: plenty of skeleton, very little data, and no shortage of people willing to read conclusions out of empty cells.

The week before, I had sat down to rewatch four group-stage matches from a North American regional league I follow closely. My habit is to take notes by the minute, not by highlight. After four matches I had three pages, and one line made me stop: at minute fifteen, the team rated weaker led on gold, yet lost on bot-lane resource share and lost the major objective count as well. A week later that same team was eliminated. Nobody wrote about that line, because it was not loud. The crowd leaves, but the numbers stay — and for the first time I saw them as empty, in a different sense altogether.

Context: a market with no shared window

Football has two transfer windows fixed by national and continental federations. Esports does not have that structure. Each publisher runs its own calendar. Riot Games publishes off-season periods and roster-lock dates for League of Legends by region; the VALORANT Champions Tour has its own transfer cycle and roster lock; Counter-Strike 2 operates through community norms and third-party events. There is no shared window, no single registry, and almost no cross-publisher enforcement mechanism.

The first consequence is that rumours multiply faster than data. A social media account posts three words, a livestream gets cut into ten clips, a blurred wallpaper appears in an agent's screenshot — all of it becomes evidence in the reader's eyes. But evidence is not data. A photograph can prove two people met. It cannot prove a contract was signed, let alone what the contract is worth.

My readers in the US market consume all of it at once: rumour rankings, projected rosters, salary estimates built by outsiders. They are flooded, and when people are flooded they filter by belief rather than by structure. So my job in every transfer piece is not to add another list. My job is to rebuild the four data layers a real deal passes through, so readers know which layer they are standing on and which layers are still empty.

Every number is a story waiting to be verified. And in a transfer market, every story gets told before the number arrives.

Layer one: contract structure, not the figure in the headline

Most transfer stories end with a sentence shaped like: the deal is worth X. That is the crudest and most error-prone layer, because X is usually a figure leaked on purpose by someone inside, or added up by outsiders from several sources without stating a method.

A professional esports player contract, in its full form, usually contains: base salary per year or per month; a signing bonus paid once or in tranches; performance bonuses with very specific triggers such as reaching a regional final or qualifying for an international event; image rights and a share of revenue from jerseys and personal streams; contract length with a one-way or two-way extension option; and most importantly, the release clause — often called the buyout.

The release clause is where real value sits. A player can earn a modest base salary while holding a low buyout, meaning the owning team is holding an asset that is easy to take away. Conversely, a player on a high salary with a buyout set three times higher means the team has locked itself into a relationship it cannot easily exit, and in the market that player barely moves until the contract drops below twelve months.

Across years of data work for sports organisations, I always ask the same question before reading any transfer figure: what is this number the sum of, over how many years, and who has the authority to confirm it. If the reporter cannot answer those three, I file it as low-grade rumour, even when it comes from a well-known outlet.

There is another category of transfer news that is far more reliable and far less noticed: early terminations, loan moves, and extensions. All three leave administrative traces. Early termination usually carries a settlement, and a settlement has to appear in the organisation's books. Teams that publish those traces late are usually teams hiding a payroll problem.

Layer two: the payroll and the allocation problem

Payroll is the second layer and the least discussed in transfer writing. A big deal says nothing unless it sits next to the team's total payroll and the years remaining on existing contracts.

According to documents Riot Games has published for the North American league across several seasons, the league's salary system includes a floor and a mechanism controlling spending, under which teams exceeding a threshold pay an additional contribution redistributed to the rest. I deliberately do not quote a specific figure here, because those thresholds have changed season by season and I have only read the published documents, not internal books. Quoting a number without naming the document version and the year it applied is the fastest way to turn a valid fact into an invalid argument.

What is certain: once a spending threshold exists, a player's value stops being absolute and becomes relative to the payroll space remaining. A team that has already committed most of its payroll to two stars must choose between a third player of average quality on a low salary, or an unproven young player. The market calls that a small move. In reality it is a move forced by structure.

My self-assigned task every transfer window is a four-column table: position, years remaining on contract, estimated salary band, and the depth shortage for that position in the region. That table does not need a complex model. It needs patience and consistency. At Northampton, we did not have technology, we had patience and a spreadsheet.

Once those four columns are filled, most rumours fall away by themselves. A player with two years left, in a position where the entire region has only four qualified names, does not move unless a release clause is triggered. No amount of noise changes that.

Layer three: performance data and the definition trap

This is the most seductive and most dangerous layer, because it feels precise. I paid to learn that.

In June 2026, during the World Cup in Russia, I published an expected-goals model of my own. In the match where Germany lost to Mexico, my model said Germany created about 2.1 expected goals and deserved to win. The next day a veteran analyst pointed out a methodological error: I had not subtracted the shot-angle coefficient and had not accounted for defender pressure, inflating the output by roughly thirty-four percent. I spent the next six weeks, the rest of the tournament, rewatching all sixty-four matches and recalibrating the model with tracking data from every phase of play. When Germany went out in the group stage, I wrote a piece arguing against myself, calling the first analysis a rushed conclusion from raw data.

The lesson was not that I was wrong. The lesson was that a metric's definition determines which conclusions it can support. Data never lies, but the person who defines it can.

In esports this problem is worse than in football, because most metrics come from match logs published by the publisher, and the labelling rules are set by the publisher. Gold difference at minute fifteen is a simple measurement, but it lumps together gold from minions, from objectives, from kills and from towers. Two teams with the same gold lead at minute fifteen can be in completely different states: one controlling the map, the other having taken one big kill and retreated. One number, two stories.

The same applies to damage per gold, kill participation, and vision score per minute. None of them says anything on its own. They speak only when placed inside the spatial structure of the match: who stood where, at what moment, with how much resource.

Two older examples help me explain this to American readers. The first is PPDA, the number of passes a team allows per defensive action, where lower means more aggressive pressing. In March 2026, while a master's student in sociology, I volunteered as a data analyst for Northampton Town in League One. I found the team had a PPDA of just 8.7, lowest in the league, yet an unusually high chance conversion rate of 14.2 percent. I wrote a forty-page report arguing that the team's high press was in fact active defending rather than disorganised attack. The manager at the time waved it off. After a run of five straight defeats, he dropped the pressing line eight metres deeper. Northampton stayed up with two points more than the relegation places.

The second example comes from Euro 2026. My model then, built on expected goals and PPDA, predicted Italy would be eliminated in the quarter-finals, because they averaged only 1.2 expected goals per match, roughly twenty-five percent below their direct rivals. Italy won the tournament with only the seventh-highest total expected goals. Rewatching the footage, I found a variable the model had never included: the average distance between the two centre-backs was just 21.4 metres, the smallest in the tournament. That distance produced tempo control and killed counter-attacks before they became shots. I wrote a self-critique titled Italy does not need expected goals, they need position, and it drew twelve thousand reads in twenty-four hours.

In esports there is an equivalent that few track: the gap between the two mid-laners during the laning phase, the rotation speed of the jungler, and the moment a lineup compresses into a block. None of those three appears on the stat sheet handed to viewers. They only appear when you map ten players' positions minute by minute, and then you realise a winning team does not win because it is stronger on every metric, but because it holds a spatial structure that prevents the opponent from generating a favourable one.

A wrong measurement is more dangerous than no measurement at all.

That is why in any report I produce, the limitations section comes before the conclusions. Readers need to know which variables I could not control before they trust what I assert. When four data fields all read insufficient information, that is not the analyst's helplessness. That is the result.

Layer four: injury, return, and career length

The final layer is the most tightly controlled, and the one the transfer market misreads most.

I hold a professional position shaped over many years: a player's return timetable is controlled by the team's communications department. When a statement says the player will be reassessed at the end of the week, the most accurate reading is that the injury has not healed. That phrase exists to protect the player's market value while the real condition remains unconfirmed. In fourteen years I have never seen a case where an end-of-week reassessment led to the player returning the following week.

In esports the most common injuries sit in the wrist, elbow and shoulder, alongside vision and sleep problems. But injury data is almost never published. Football has public injury databases and associations that require reporting. Esports has no such structure. A player absent for three weeks might be out with a wrist injury, for personal reasons, for an internal suspension, or because he is negotiating a transfer. All four are announced with the same sentence: personal reasons.

That makes pricing a player in treatment an almost unverifiable calculation. You do not know severity, you do not know history, you do not know whether the injury recurs cyclically. You only know that the owning team is discounting or holding firm, and that choice is itself a signal.

Career-length data sits in the same layer. A top-level esports career is typically shorter than a football career, while the youth pipeline and post-retirement support are close to non-existent. This is a point I always raise in client reports. A team buying a twenty-two-year-old is not buying ten years of service. They are buying a short window, and the value of that window depends on whether the team has a transition system.

The case of Tyson Ngo, known as TenZ, is the clearest example of how this layer behaves. In September 2026, TenZ announced he was leaving professional VALORANT to move into a content-creation role. That was a rational decision for a player with a large following. The problem is this: very few players can move into content creation with comparable income. The rest leave quietly, and I have never seen a database that tracks them.

Around the same period, the story of Oleksandr Kostyliev, known as s1mple, in Counter-Strike showed another form of movement: after years with Natus Vincere, he moved to compete on loan at another organisation. Loans are under-analysed in esports even though they are the most efficient risk-sharing tool. The owning team keeps the contract and the right to reprice later; the borrowing team gets a player for a short period at lower cost. Late-window markets drift toward loans, and that is what deserves tracking more than the headline figures.

The contrarian angle: the gap is also a signal

At this point I have to argue against myself, because that part cannot be skipped in any report I write.

The case for cautious reading is simple: without data, every conclusion is inference. But the counter-case is just as strong. A gap is not meaningless. When an organisation does not publish a contract value, does not state a duration, does not confirm an injury status — that is a behavioural pattern, and patterns can be measured.

I started tracking the publication frequency of organisations over the last two seasons with a simple sheet: every transfer announcement logged against three boxes, whether it stated contract length, whether it stated contract type, and whether it stated the position alongside the previous season's minutes. The early results, on the small sample I collected, are that organisations publishing contract length also tend to publish contract type, and organisations hiding both tend to be those changing head coaches or facing payroll turbulence.

I deliberately name no organisation, because my sample is too small to assert causation. Correlation is not causation, and in an industry where data is published at the owner's discretion, the risk of reading a rule out of coincidence is enormous.

But it taught me something about my own trade. When a report arrives with nine sections and every data field empty, my reaction should not be to complain. The appropriate reaction is to publish the gap itself, with reasons and with the milestones to track next. Every match is a data sample, but belief is the only variable that cannot be entered into the sheet.

I do not trust intuition, I trust data — and data itself taught me to trust no one.

What to track in the next cycle

For readers following the current transfer window, I suggest four milestones instead of four predictions.

Roster-lock dates for each regional league, published on the publisher's calendar. Before that date everything can flip; after it, the roster is evidence.

Years remaining on the contracts of players in positions with depth shortages. This is the one variable an owning team cannot fully hide, because it appears in every extension or parting announcement.

Changes in coaching staff. In esports the head coach is usually the first to be replaced and the first replacement to affect transfers, because the new coach's philosophy determines which positions get bought.

The patch. Every time the meta shifts, the relative value of positions shifts with it, and that is when repricing old performance becomes most urgent. A player with strong metrics in one version does not automatically keep them in the next, and anyone ignoring this variable is pricing a season that no longer exists.

Those four milestones do not need a model. They need a spreadsheet, patience, and one small habit I have practised for years: before believing anything, ask where its definition comes from. In a transfer window where noise drowns signal, keeping one honest empty cell is sometimes the most valuable analytical act available to you.