Sample First, Cheque Later: The Jepchirchir Case and the Marathon Ledger With No Auditor

Thirtieth of August, 2026, Sydney. The women's marathon result sheet sat in front of me in two lines — Peres Jepchirchir, two hours eighteen minutes thirty-one seconds; Irene Chepet Cheptai, more than three minutes back. I was revising my women's lines across all seven World Marathon Majors that week, and the note I wrote in the margin of that page is still in the folder: "Don't price the time. Price the margin."

Some days later came the notice from the Athletics Integrity Unit. An out-of-competition sample, collected in Kenya before the race, had tested positive for EPO. Provisional suspension. Look at the order of the paperwork — sample first, victory second, prize money cleared after that, and the notice last of all. In ledger language, the sequence is inverted. We call it settlement before audit: the money has been distributed, the ranking points allocated, the trophy has gone home, and only then does the examiner knock on the door.

I trust the receipt more than the highlight reel. In this case the receipt was written first. The question is not how good a runner Jepchirchir is — that ledger is separate and it is full. The question is why it took until 2026 to catch an Olympic champion, a world champion, a winner of seven majors, and whether the catch was a design feature or a lottery ticket that happened to open.

A thirty-two-year-old note before anything else. Athens 2026, men's long jump pricing. I was building lines for a small syndicate off IAAF season lists and wind readings, and I noticed the London books were paying for reputation rather than for the measurement environment. The Athens outlier was not noise; it was an invoice. Since that week I have held one rule: conditions before numbers — wind, altitude, sample size, standard of competition. The day I wrote that rule for myself, the marathon ledger changed for me too.

So here I will first write down what cannot be written without its conditions.

Jepchirchir's ledger first, because it is not on trial

In 2026 she broke the women's half-marathon world record. In 2026 she broke it three more times — resetting her own ceiling three times inside one year, which in half-marathon terms is an extraordinary document. Gold in the women's marathon at the Tokyo 2026 Olympic Games, run in Sapporo. New York City Marathon 2026, Boston Marathon 2026, London Marathon 2026 — the London win being the one my own sheet records as the best women-only performance, and I keep that in a separate column from the men-paced lines. In September 2026 she returned to Tokyo and won the World Championships. On 30 August 2026 she won Sydney in 2:18:31, by more than three minutes.

This is not a list of suspicion. It is part of the inventory without which any discussion of this case becomes dishonest. When a runner writes a half-marathon world record four times in seven years and wins majors on four different continents, her name carries a price — appearance fees, broadcast rights, sponsor contracts, series points. So when the AIU notice arrives, the market's first question is not what is true. It is which ledger this case will be written into.

The AIU process: a suspension is a step, not a finding

The Athletics Integrity Unit is an independent body operating under World Athletics, and its remit is not championships but a year-round out-of-competition testing programme. Athletes must file whereabouts; the testing agency can arrive anywhere, at any hour, on the strength of that filing. A positive result brings a provisional suspension, which means the athlete cannot compete until the case is resolved.

One thing should be said plainly, because in television language these three things sound identical: an A-sample positive, a B-sample confirmation, and a final finding of an anti-doping rule violation. They are three separate documents, each with its own date. The B-sample analysis, the laboratory documentation, the chain of custody, then the hearing, then — if it goes that far — the Court of Arbitration for Sport. A metric dictionary is a promise you can audit, and so is an anti-doping code: every stage has its definition, its deadline, its appeal route. People who deliver a verdict in a day find that document trail inconvenient.

What EPO is, and why it is hard to catch

Erythropoietin, EPO, is a natural substance produced within the kidneys that stimulates the creation of new red blood cells. Injected, it is detectable in urine or blood only for a short window, and it increases the volume of red blood cells — the vehicles that carry oxygen to the muscles.

Its popularity and its forensic difficulty come from the same property: the window is short. In plain terms, this is a substance whose effect lasts for weeks and whose trace lasts for days. That is why the international system built a second instrument — the athlete biological passport, which reads change across years of blood data rather than a single day's sample. That instrument works only as often as blood is drawn. In a country where blood is drawn twice a year, a passport is not a trend. It is a smudge.

Sydney became the seventh major in 2026, and that is the real structural change

That Sydney became the seventh World Marathon Major in 2026 is not at the centre of this story, but it sits at the centre of this story's economics. Going from five majors to seven is not a calendar adjustment. It changes four things: the number of large paydays available to the top twenty, the market rate for appearance fees, the politics of series points, and — least discussed — recovery time.

Marathon is a sport in which the calendar decides a large share of the performance. Eight weeks between majors makes a race; twelve weeks makes a project. As the number of majors rises, the athlete gains a new freedom — which race to skip — and a new pressure: three cheques in three months. Put seven large cheques on the top table and the part of preparation people call unseen work compresses. The day repair time shrinks, the market for external assistance in recovery grows. That is not a moral observation. It is calendar arithmetic.

The four-number ledger, of which only two are public

I keep a column for what the market refuses to price. Here it is a column of four questions, and these four will decide whether this is a case about an athlete or a case about a system.

The first: the exact collection date and the exact analysis date. All that is public is that the sample was taken in Kenya before the race of 30 August 2026. "Before" is a word, not a date. Four weeks before or four days before? It matters, because the gap between the EPO window and the race determines whether the question is one of performance or one of process.

The second: the status of the B sample. Without it, every analysis is incomplete, whichever way it leans.

The third: the athlete biological passport history. This is the most important number and the least public. If a flag was already raised in the 2026-25 readings, this is not a single-night accident but a pattern, and the question becomes one of administrative failure. If the passport showed nothing, then this catch was a lottery ticket that opened — and the conclusion is that testing volume is so low that success depends on luck.

The fourth: the Sydney appearance fee and its contract structure — who paid, when, which part carries a morality clause, and whether the race organiser has any route to recovering its money.

Not one of these four is fully public. When the sample is small, the process must be loud. Here the process is still whispering.

The margin, not the time: what three minutes tells us

One number is public and has not been read properly: the winning margin. More than three minutes. In a major marathon that is not a race, it is an accounting entry. In my own archive of women's major winning margins from 2026 to 2026, a gap of three minutes is rare, and when it appears the cause is usually not the athlete but the field — how strong the second-best runner in the race was. I am not writing that as a public claim; it is my sheet, my column, and the reader deserves to know where it came from.

Which does not mean three minutes is evidence of chemical assistance. It means the Sydney race was not a contest, it was an exhibition — and in an exhibition the loser occupies the space of a supporting act. So when the question of annulment arrives, the largest financial loss is not Jepchirchir's. It is Cheptai's. She lost by three minutes, but if the win is voided she becomes a first-time major winner — the biggest cheque of a career, stopped by somebody else's error.

On the other side is the time: 2:18:31. I did not record the Sydney course measurement and that day's temperature in my sheet, so I will not file this mark as a record or a non-record. I will say it sits inside her established band — the band she built in London and Tokyo. And that is precisely what makes this case hard. If the time had been aberrant, the story would be easy. This time is normal, and normal times do not come with a sellable suspicion.

Why marathoners get caught: the geometry of testing

Across two decades of building lines on the majors, I have noted something rarely written down: marathon runners are simultaneously the easiest and the hardest out-of-competition targets in the sport.

Easy, because their pre-race block is long, dull and geographically fixed. Iten, Kaptagat, Ngong — these are camp names, and a camp means an address. A sprinter moves through three continents; a marathoner runs the same route at the same altitude for ten weeks. For a testing agency there is nothing better.

Hard, because the same stability fixes the evidence. When the sample is collected at the camp rather than at the race, the question becomes how clean the sample's journey was — from Kenya to the laboratory, how many hours, through whose hands, at what temperature. The one thing that is public in this case is that the sample was collected in Kenya, not in Australia. That is the biggest clue in the whole story, and it is a clue about the testing plan, not about the athlete.

That clue says the AIU was working Kenya during the preparation weeks, not Australia during race week. It is the right strategy, because substances are used in preparation, not on race day. It also marks where the system worked and where it did not. If this was a scheduled, planned test written into a calendar, then somebody decided to look at the camps. If it was random, it is a ticket that opened. The distance between those two readings is the distance between two policies.

The performance question: correlation is not causation

Now the part where I have to be severe, because this is where the most lying happens — in both directions.

On one side are those who say: positive sample, therefore 2:18:31 is fake. On the other are those who say: the sample predates the race, therefore the race is unrelated. Both are in a hurry.

Take the physiology. EPO's core function is to raise oxygen-carrying capacity. Its clearest benefit accrues to work sustained around threshold — which is exactly what a marathon is. Imagining that this can improve a performance by a few minutes is not difficult. The conclusion does not follow that this particular time on this particular day came from that. A marathon has a hundred variables: course, wind, humidity, pacing group, aid stations, shoes, and the largest one — the absence of a rival. In a race with a three-minute margin, that last variable is heavy enough to flatten everything else.

There is a technical point here I consider important. The sample predates the race. That moves the centre of gravity of the investigation away from Sydney and back into the block before it. If a substance was used, it was used in preparation, and the output of that block may have shown before Sydney. The correct question is therefore a backward walk through the ledger: the Tokyo World Championships of September 2026, London 2026, Boston 2026, New York 2026. If the inquiry stops at Sydney, the inquiry is incomplete whatever the verdict.

I will put it in one sentence: a positive sample proves the substance was in the body; it does not prove that the time that day was made by the substance. Those who read the first sentence and extract the second conclusion are translating data into punishment. Those who read the second and discard the first are translating process into absolution. In both, the audit is missing.

Who carries the bill: four debtors, and one who pays most

The most useful question in a ledger is not who is guilty but whose name the bill is in. There are four parties here.

The Sydney organisers. They bought a headline and received a test. One year after gaining major status, their race is recorded against a win that sits in a hearing room. The sponsors and broadcasters were selling this race as the seventh major; now they have to sell it with a question attached.

The series arithmetic. If the win is voided, points are redistributed, and that is the work no broadcaster shows — yet it sets next season's appearance fees. The ledger remembers the margin the broadcast forgets.

The Kenyan federation. Its most sellable asset was never medals. It was trust. When the strongest marathon nation in the world appears in the news with another large name, the damage is not to one athlete; it is to a camp system in which thousands of young runners train unpaid, for whom a single foreign major cheque is the only capital investment of a life.

And the fourth — Irene Chepet Cheptai. She speaks least in this story and pays most.

The British control group, with the structure stated first

Writing from Liverpool, the temptation is to measure everything against UK Sport's World Class Programme. I will not, because I have given myself a rule: comparison first, mechanism first, prestige afterwards.

The British model is a small, centrally funded, salaried athlete pool accounted for through the World Class Programme and policed by UK Anti-Doping, which runs the same whereabouts and biological passport machinery. Roughly six dozen athletes, a visible payroll, a visible audit.

The Kenyan model is thousands of self-financing runners in camps, mediated by agents, living off appearance fees and foreign major prize money. No national payroll, no central contracts. The Anti-Doping Agency of Kenya does not have the budget this size of pool requires, which is why the effective regulator in practice is the AIU and not the national agency. That is not a criticism of anyone. It is a financial fact.

Now the part that does not transfer. Britain's salaried model cannot be exported to Kenya, because there is no public payroll for runners there and probably never will be. What can transfer is centralised whereabouts management at camp level and funded, unannounced testing inside camps — on the road, not at the airport. It is not cheap. Cheap remedies are part of this problem.

The contrarian angle: "the system worked" is also the wrong reading

The line I hear most, and disagree with: look, testing worked — a sample was taken, a substance was caught, a suspension followed. Match won.

I call that a late receipt, not a win. How late requires the third number, the biological passport history, which is not public. That is the real weakness of this case: the document that would settle whether this was a system success or a system lapse is not in front of the public.

If the passport caught nothing and the EPO test caught this, then passport-based surveillance was inert in this athlete's case. And if a system's primary instrument is inert and a direct test is doing the work, then the hit rate depends on testing volume, not on design. That is my objection. Calling a single catch a system's victory is like calling a single accurate forecast proof of a model.

The second contrarian point is more uncomfortable. In the rush to void Sydney we forget that a provisional suspension is not a finding. The defence's first attack will not be on the drug but on the sample's journey — chain of custody, transport, laboratory documentation, analytical method. EPO is the substance you cannot accidentally eat; it cannot be hidden in a supplement. That is its strength and that is the weakness of a defence: when contamination is unavailable as an explanation, the attack goes to process. What we will read over the next three months is not physiology. It is forms. The reader who came for a drug story will have to read forms.

The third. Do not turn this into a Kenya story. The cases that surfaced in Kenyan athletics through the 2010s were real and carry their own dates. The more useful lesson here is that direct EPO positives in major marathons are rare — rare because testing concentrates on the top twenty. How often is the blood of the women ranked twentieth to hundredth examined? That cell in my sheet is empty, and an empty column is also a conclusion.

What to watch next

Four things. The date and result of the B sample. The AIU stating the collection date — "before the race" is a phrase, a date is a fact. Whether a biological passport review opens, and whether it reaches back into 2026-25. And the Sydney cheque — whether the organisers ask for their money back, and under which clause.

Behind those four sits a fifth question no code answers. In the marathon economy, appearance fees in the millions are agreed between agent, organiser and sponsor, and that ledger has no independent auditor. Testing happens on bodies, not on deals. If anything is to change from this case, the change will not be in the sanction. It will be in the addition of a column that is currently blank.

I will wait for that column. The rest of it — the receipt is already in.