The Prediction

On 9 September, the Centre opened The Experience¹. Since the following day, the colour has been drawn and sealed on the server before anyone chooses among four: the target already exists; it is only hidden. The Prediction asks the next question, a more demanding one: can one designate what does not yet exist? Each day, before noon, everyone touches a spot in a square; at twelve twelve, the server places a point in it, at random. This article describes how it works, and why it was built this way.

1.Two experiments, two questions

In The Experience, the colour is drawn before the choice. What one seeks to perceive is already there, locked in the server, and only the revelation is missing. This is the situation of remote viewing, where the target exists and waits to be described.

In The Prediction, the order is reversed. The point is designated first, sometimes almost a day in advance, and only drawn afterwards. At the moment each person places a finger in the square, there is nothing to perceive in the ordinary sense: the information does not yet exist anywhere. If the designated points come closer to the drawn point than chance allows, this will not be the reading of hidden information, but access to information that did not exist at the moment of choosing. The literature calls this precognition; the Centre sees in it perceptibility, observed across time.

The question is not new. In 1927, the engineer John William Dunne recorded his dreams in order to compare them with the events of the following days². In 1969, the physicist Helmut Schmidt asked volunteers to predict which of four lamps would light up, the lamp being chosen afterwards by the decay of a radioactive nucleus³. Keeping the Laboratory’s two experiments separate answers the same concern of method: each asks only one question at a time. Mixing what is hidden with what is yet to come would make any result unreadable.

2.What is measured

The square has one thousand positions on each side. Each designated point is measured in two ways. The first is coarse: is the point in the right quarter of the square? Chance lands there one time in four. The second is fine: the distance between the designated point and the drawn point, expressed out of a thousand.

The second counts most, because it keeps everything. A point that narrowly misses the right quarter and a point that lands at the opposite side are not worth the same, and the quarter confuses them; the distance tells them apart.

One still needs to know what chance gives. Two points placed at random in a square are, on average, a little more than half a side apart: 0.5214 times the side, a value established in 1978 by the mathematician David Robbins⁴. On this grid, that makes 521 out of a thousand. The Centre published it in the protocol⁵ before the first draw, and checked it by simulation over three hundred thousand draws. This number holds for two points drawn at random. The designated point, however, is not drawn: a person places it, and the place chosen changes what chance gives. From the centre of the square, a draw lands on average 382 out of a thousand away; from a corner, 764. Comparing with 521 the distance of someone who always aims at the middle would credit them with an accuracy that is mere geometry. Each prediction is therefore read against chance seen from the exact place where it was put, a value computed in advance that knows nothing of the point drawn; the group is read against the mean of those values. Over the whole square, that mean gives back Robbins’s 521.

3.A point sealed at the instant it comes into being

An experiment of this kind is worth something only if nobody, not even the Centre, can know the point before the hour. That is why it is prepared nowhere: neither drawn the day before and kept in reserve, nor calculated in advance. At twelve twelve, a server task draws two independent numbers, one for the width and the other for the height, then at once seals the pair with a thirty-two-byte secret and its SHA-256 fingerprint. Everything is written in a single line, which the database then refuses to modify or erase. Submissions close at noon, twelve minutes before the draw: no prediction can be placed once the point exists.

The seal lets anyone check that the revealed point is indeed the one drawn at twelve twelve. By recomputing the fingerprint from the point and the secret, one finds the one recorded at the instant of the draw. The same rule protects each prediction: once confirmed, a point no longer moves, and a measurement is never redone.

4.The contrary hypothesis

One must examine what a sceptical mind would say, because that is how the Centre works. The premonitions that strike us are almost always told after the fact: we remember the dream that came true, and forget all those that came true in nothing. In an experiment, the same trap takes other forms: counting only the good days, stopping when the curve is favourable, changing the measure along the way, slicing the data until some subgroup looks remarkable.

Recent history offers a serious example. In 2011, the psychologist Daryl Bem published, in a leading journal, nine experiments that seemed to show an influence of the future on present responses⁶. The following year, three laboratories repeated one of them exactly and found nothing⁷. The debate that followed was not only about precognition; it was about how to analyse data when one already knows what one hopes to find.

The Prediction was built against these traps. Everything that decides the result was written before the first data point, and published: what is being looked for, the distance of chance, the group threshold, the number of predictions to reach before any assessment, and the deviation from which a result will be spoken of. All predictions count, the good ones and the others; none can be withdrawn, moved or redone. And the point is drawn by the server, at random, without anyone being able to choose it or know it.

If, in the end, the points come no closer to the drawn point than chance, this will be published in the same way. An experiment that can say nothing against what it hopes for says nothing.

5.The group, and one minute before placing

A single prediction says nothing: a close point may be a coincidence, and so may a distant one. That is why The Prediction looks first at the group. From thirty points placed for the same draw, the observations page shows the map of the day: everyone’s points, the drawn point, and the group’s centre of gravity. This centre falls mechanically closer to the middle of the square than most of the points; its distance to the drawn point must therefore be read with caution, and the protocol does not rely on it. What counts is the group’s average distance, against chance seen from the points placed that day.

The protocol also looks for a second thing. Before placing a point, each person states the condition they are in: calm, restless, tired or distracted. Then, one time in two, drawn by lot, a minute of preparation is offered: stillness, slow breathing, attention resting on the body. The other time, the square opens directly. The lot makes the comparison honest: those who prepare are not more motivated people than the others; they are the same people, one day in two.

Here one finds again the three conditions described in the article of 9 September¹: a body in tune, a mind that does not scatter, emotions that do not distort what passes through. If perceptibility obeys these conditions, predictions made after the minute of preparation, in a calm state, should come closer to the drawn point. This is what The Prediction will make it possible to observe, over time.

6.What each person sees, and what will be published

Each person keeps a logbook: the list of their predictions, the right quarter or not, the distance, the declared state, preparation or its absence, and a day-by-day curve against the line of chance. A single day below that line proves nothing; a curve that stays there, over dozens of days, deserves attention. This logbook belongs only to the person who keeps it, and no individual figure will ever be published.

The observations page shows what the group did, draw after draw, without any identity. The cumulative assessment will wait: nothing will be published before ten thousand predictions⁵. At that point, the Centre will say what has been observed, with the number of predictions and the confidence interval, and will speak of a result only if the deviation from chance exceeds three standard deviations, a threshold set before the first data point. Until then, no good day will be presented as proof, and no bad day as a refutation.

7.Placing a point

The Prediction is open to everyone, with the account of the Laboratory, with no other registration and no starting condition. One point may be placed each day, between twelve twelve and noon the following day. Those who wish receive the map of the draw by email every day at twelve twelve, and can stop it with one click.

Nobody will tell you whether you predicted correctly; you will see it for yourself, over time, and the group will see it with you.

It is open to everyone, on this site: The Prediction.

Centre EPÉU, 16 September 2026

Centre EPÉU — Château de Villereglan, Saint-Martin-de-Villereglan (Aude) · the site chat

← Research and observations — all texts