When a Ninth-Tier Team Sheet and a Champions League Winners' List Carry the Same Name

The Tiefenbach team sheet for a Kreisklasse fixture listed a name in the goalkeeper position. Not a specialist goalkeeper. A centre-back, borrowed shirt, gloves on for two consecutive matches. First match lost 1-2. Second match won 3-2. Four goals conceded across two games, no save percentage, no shot map, no defensive metric recorded anywhere. All that survives is one line in the match report: handled the role pretty well.

In the same week, another database carried the same string of characters with 49 caps for Germany, one international goal, five Bundesliga titles, a place in the Bayern Munich squad that won the 2026 Champions League, and a live professional contract at Borussia Dortmund.

Both records cannot be true. A player registered and fielded professionally cannot simultaneously appear on the team sheet of a ninth-tier German fixture. A man who announced retirement at 31 cannot also be playing a European cup tie. I read reports like this more slowly than usual, because the contradiction sits in the raw data layer, not the interpretive one. When the raw layer is wrong, everything built on top is decoration.

My daily work in the transfer market is reading rumours, cross-checking sources, and deciding what deserves a line in the tracking sheet. During a transfer window the volume of information grows exponentially while the volume of real signal stays flat. Most of the job is deletion, not collection. A report earns its place only if it answers three things: who the primary source is, which key field confirms identity, and which fact can be verified independently.

In 2026 I published an analysis of Marseille against PSG. PSG won 3-0. My expected-goals data showed Marseille created the more dangerous chances. I received hundreds of dismissive comments, most aimed at my gender rather than my method. I rebuilt the framework across 23 Ligue 1 matches and showed that PSG were winning heavily on an abnormal conversion rate, not on control of the game. Three months later PSG lost 1-2 to Lyon and the underlying numbers fell away exactly as projected. PSG won that year, but I chose to believe in the shots that did not go in.

That lesson shaped how I read every report, amateur ones included. Without a table there is no conclusion. When two tables contradict each other, the first task is to establish whether they describe the same subject at all.

Watching lower-division matches in southern France gave me one concrete detail: at amateur level the match report is written by one person, usually a volunteer, usually within thirty minutes of the final whistle. A misspelled name is routine. Attaching an entire professional CV to an amateur name is not a typo. That is an editorial decision, and editorial decisions have motives.

Three data layers need separating.

The technical layer. A centre-back putting on gloves is not a tactical choice; it is the consequence of a personnel event, namely the first-choice goalkeeper breaking his hand. At Kreisklasse level almost no club registers a third goalkeeper. The operating model runs on member fees and local sponsorship, with no room for a dedicated reserve. When the event happens, the club does not spend money, it improvises with people.

A centre-back is the least disruptive outfield profile to convert, because three skill sets overlap: aerial duels, reading of the game, and distribution with the feet. The modern ball-playing goalkeeper archetype overlaps substantially with a centre-back skill set in a possession system. The overlap stops precisely there.

The technical blind spot of a centre-back in gloves lies in reflex saves, low handling, catching under contact, and one-on-one positioning. None of those four skill groups is forged by hundreds of matches at centre-back. With no save data to measure, any assessment of the handled-it-well variety is a subjective observation by the report writer, not a measurement.

On results: four goals conceded in two matches. At this level expected goals do not exist, defensive passing volume does not exist, pressure indices do not exist. The figure of four, standing alone, proves neither competence nor incompetence. A two-match sample sits below any meaningful statistical threshold. The only honest conclusion is that there is no conclusion. Croatia in 2026 taught me that heroes have biological limits, and those limits only surface once the sample is large enough. Two matches is not a sample. Two matches is an anecdote.

The financial layer holds one datum worth discussing, and it depends on an identity assumption. The report states the player was not offered a contract extension by Borussia Dortmund, with injuries cited as the reason. Read as squad-cost management, this is a contract-expiry exit.

The mechanism: a high-earning player with an unstable fitness record enters the final year of his deal. The club has two options. Sell at a depreciated price, accepting a book loss and losing a squad place mid-season. Or hold until expiry, collecting no fee but incurring no new cost and leaving no residual amortisation for the buying club. Dortmund are said to have taken the second route.

For the buying club, a free agent arriving after expiry carries a clear accounting advantage: no transfer fee to amortise, a starting book value of zero, and no sell-on clause payable to the previous club. For the selling club, it is an acknowledgement that the asset has lost its liquidation value. The transfer market does not buy players, it buys stories.

But this financial layer rests on an unverified assumption. If the name at Tiefenbach and the name in the Bundesliga record belong to two different people, the entire layer collapses. That is why the verification layer matters most.

Lay the contradiction on the table. Side one: 49 caps for Germany, one international goal, TSG Hoffenheim then Bayern Munich, the 2026 Champions League title in the final in Lisbon on 23 August 2026 when Bayern beat PSG 1-0, five Bundesliga titles, and a professional contract at Borussia Dortmund signed as a free transfer in the summer of 2026. Side two: 31 years old, retired, not extended, currently playing Kreisklasse for Tiefenbach, occasionally as an emergency goalkeeper.

The two sides are mutually exclusive at the registration level. The player named Niklas Süle in the professional record is an active top-flight centre-back, not a retiree. Ranking the hypotheses: a namesake collision is the most likely, because Niklas Süle is not a unique string and the most common way such pieces are built is to take a genuine amateur story and attach a professional CV as the hook. A conflation error during aggregation is the middle case, where two subjects were merged in editing. A wholesale fabrication is the least likely, because the amateur portion contains specific injury and scoreline detail that is hard to invent from nothing.

People see a comeback; I see a chart that is breaking. The same applies here. Readers see a curious story. I see two datasets tied together by a thread that does not exist.

One point needs stating clearly, because data analysis is often misread. Rejecting the hook is not rejecting the whole event. The amateur story stands on its own: a ninth-tier side loses its goalkeeper to a broken hand, a centre-back borrows the gloves for two matches, one win one loss. That is good material. The problem is that it was packaged with a CV that does not belong to it.

The mechanism producing this error deserves more attention than the error itself. In sports media economics, a purely amateur report has close to zero distribution value. Attach a top-level name and distribution value spikes, because algorithms prioritise names people already search for. The motive need not be deliberate deception. It only takes one aggregator noticing the name match, finding it interesting, and pushing it. From there, every share replicates the contradiction one step further.

This is why I never conclude from a single source, even one that looks reputable. Correlation is not causation, and a shared name is not a shared person. Two identical strings on two different tables amount to a correlation. Turning that into causation requires a common key field: a registration number, a date of birth, or direct club confirmation. Without a common key, every inference is a probability, and a probability must be declared as such. A risk model saves nobody, but it gives them a chance.

There is one noise variable I always interrogate before closing: what if my own table is wrong. Here the only available check is the next round team sheet. The signal to watch is not the scoreline. It is the registration list. If the name at Tiefenbach reappears in goal within the next two fixtures, he is a real amateur and his story deserves telling without borrowing anyone biography. If the name vanishes, that is evidence the story existed only to serve a headline.

Data is the only thing I trust after watching too many promises break. In this case the data says something simple: before asking whether someone played well, establish who that someone is.