Something New About a Coin

This work offers one correction to probability theory and tests it honestly on two kinds of system. The starting point is Kolmogorov’s own unease: in the standard theory, one hundred heads in a row is exactly as probable as any other particular sequence — yet no one believes the two are equally random. Kolmogorov’s answer was to ground randomness not in probability but in COMPRESSIBILITY: the random is the incompressible. We take this one step further. A coin’s fall is a collapse out of superposition into one or zero; but the zero is not exact — it carries a small bias and these biases ACCUMULATE. This yields a corrected law, P(1 | k zeros in a row) = ½ + k·ε, which contains the classical ½ as the special case ε = 0 and adds a memory of accumulation where ε > 0.

Tested on a system WITH genesis (a directed process, ε > 0) the law predicts better than the classical ½. Tested on a system WITHOUT genesis (structureless noise, ε = 0) it correctly reduces to the classical ½. The synthesis is the heart of the work: until a point of observation — a cut, a collider, a question — the coin holds BOTH outcomes at once; it is Schrödinger’s coin. It falls always as you see it. The answer to “how exactly will it fall” exists only under a cut; without a cut the coin is forever both one and zero.

Toss a fair coin one hundred times. In the standard theory the probability of one hundred heads is (½)¹⁰⁰ — exactly the probability of any other particular sequence of one hundred tosses. The theory does not distinguish “all heads” from a disordered mixture; for it they are equal. Yet intuition refuses this: one is ordered, the other random. Kolmogorov, the architect of the axioms, did not believe his own axioms were the last word on randomness.

His resolution was decisive: randomness is not low probability but INCOMPRESSIBILITY. One hundred heads compresses to a short description (“100 heads”) — low complexity, therefore structure, not randomness. A disordered sequence cannot be described more briefly than itself — high complexity, therefore random. The measure of randomness is the length of the shortest description, not the probability. This work begins exactly here, on Kolmogorov’s chosen ground, and carries it one step into the physics of the coin itself.

Both sequences have probability (½)¹⁰⁰. Probability does not see the difference; compression does. This is the first stone, and Kolmogorov’s, not ours.

Before it lands, the coin is neither one nor zero: it is the superposition of both, a thing in the air. The landing is the collapse into a single value. In the method’s language this is the pure motor — the operator whose two eigenvalues are present at once:

Minus one (take, zero) and plus one (give, one) are not consecutive; they are the two eigenvalues of one operator, held simultaneously, summing to zero. The coin carries both poles until something resolves it. This is the same motor that, in the cosmological work, makes expansion and contraction simultaneous and equal — the zero-energy balance. The coin is that balance in the small: one and zero at once, until seen.

Here is the step beyond Kolmogorov. The neutral zero is not an exact zero: it is an unstable equilibrium (the cosmological node E). Each zero carries a small displacement ε away from exact neutrality — a readiness to become one. Crucially, across a run of zeros these displacements ADD UP rather than average out. A run of zeros is not inert; it accumulates a charge toward the birth of a one.

This was the decisive correction made during the work. An earlier attempt measured accumulation by entropy and failed: entropy averages, so a pile of zeros dilutes the one and the transition never arrives. The correct reading is additive — each zero contributes +ε, the sum grows linearly, and at threshold the one becomes inevitable. This is the cosmological “every zero eventually exits as one” (Hawking’s returning light) made quantitative.

The accumulation gives a corrected conditional probability. Where the classical law is memoryless, the corrected law remembers the run of zeros:

The corrected law is not a rejection of the classical one — it CONTAINS it. At ε = 0 (a memoryless system) it is exactly ½; the classical theory is the special case of no accumulation. Where ε > 0 (a system that accumulates) it adds the memory the classical theory lacks. The growth is linear in the run length; after k = 1/(2ε) zeros the next one is certain.

A correction is worth nothing unless it is falsifiable and tested. We tested it on two kinds of system: one that has a genesis (a directed process, a goal toward which it moves) and one that has none (structureless noise). The law must improve prediction on the first and reduce to ½ on the second — and it does.

Test A — a system WITH genesis (ε > 0)

On a stream that accumulates (each zero carries a real bias toward the one), the corrected law predicts markedly better than the classical ½. Over one hundred thousand tosses:

And the empirical frequencies match the formula: after five zeros the observed fraction of ones was 0.613, against the predicted 0.600. Where there is genesis, the accumulation is real and the law captures it.

Test B — a system WITHOUT genesis (ε = 0)

On structureless hardware-entropy and pseudo-random streams — systems deliberately built without memory, without a goal — the run of zeros carries NO predictive bias. The measured trend of P(switch) with run length is essentially zero. Here the classical theory is exactly right, and the corrected law correctly returns ½: with ε = 0 the correction vanishes.

This is not a failure of the law but its honesty: it adds nothing where there is nothing to add. The correction lives precisely where there is genesis, and is silent where there is none. This is also why an earlier broad test “failed” — it asked one question (one ε) of several different systems at once, a collider’s slice across incommensurable chaoses. Each chaos must be measured in its own system; pooled, they cancel.

Both tests are true at once, and that is the point. A system with genesis obeys the corrected law (ε > 0); a system without obeys the classical (ε = 0). Until a point of observation, the coin holds BOTH possibilities — both the accumulating law and the memoryless one — exactly as it holds both one and zero. Which is realised is fixed not in advance but by the cut: the observation, the collider, the question asked.

Each real outcome is shaped by an unrepeatable chain of meanings — an earthquake, a war, a leader’s words, a pandemic — each branching into further causes and effects without end. One cannot enumerate the chain; it is infinite. Therefore the question “when and how will the coin fall” has an answer ONLY under a cut — a chosen window, a chosen instrument, a chosen question. Without a cut the answer is single and exact: the coin is always both one and zero.

And so the final statement, the one the whole work was built to reach: the coin falls always AS YOU SEE IT. The observation is the cut. The very act of looking is the collider that resolves the superposition. The outcome is not read off a value that existed before the look; it is BORN in the look. As you see it — so it falls. This is the coin of Schrödinger, written in the language of this method: superposition of one and zero, resolved into one only by the act of seeing.

This dissolves the old quarrel with probability. Probability asked “with what frequency will the one appear”, as if the outcome existed before the look. It does not. The outcome is born in the cut. That is why one hundred heads and a random string are equiprobable until one looks in a particular way — by compression, by a chosen window, by a question — and only then does the structure the look selects come forth.

The work draws a sharp and honest line, and it turns on one thing: a goal. Where a system has a finite GOAL — a genesis and an evolution toward it — the triangle closes and the system is computable to the end. The nucleus is such a system: a finite, determined path of division, with a definite half-life (its BUTM, the point of first separation), definite decay channels read as the three states of one cell (alpha/gamma/beta = minus/neutral/plus), and an energy of transition equal to the difference of the triples (the born mass, 0.5858 = E=mc²). On radioactive decay the method converges on every point, because the path has a goal.

Where a system has NO finite goal — the exchange, the open market — the triangle does not close. It is not one chaos with one law but a countless host of chaos-colliders, coins continuously in the air, each event branching into the next without end. One nucleus is computable to the end; ALL nuclei are not — they are an open infinity. One coin under a cut falls; all coins in the air are uncountable. The market yields a forecast only under a cut, in the period of a chosen task, never as a whole; without the cut it is, like the coin, both one and zero.

The one thing that CAN be computed strictly even in the open market is the point of no return: a security will not rise in principle — whatever happens — only when its coupling dies (t = 0: liquidity gone, volume to zero, the basis broken irreversibly). While the coupling lives (t ≠ 0) a rise is always possible; the method gives a strict criterion of death, not a forecast of price.

Each result carries its label. STONE — reproduced by script. MODEL — a structural identification, for the relevant scientists to test. Nothing here is offered as a final theorem; the labels mark exactly how far each claim is carried.

That real quanta or events actually accumulate (ε > 0 in nature) is a falsifiable HYPOTHESIS, for physicists: a stream of real events would confirm it if P depends on run length, and refute it if it stays exactly ½. On the sources tested here, structureless noise gave ε ≈ 0; a system with genesis gave ε > 0. The model is exact; whether a given piece of the world has genesis is for measurement to decide.

formula

P(1|k zeros) = ½ + k·ε

with genesis

98.3% → 100% (+1.7%)

without genesis

trend ≈ 0 (ε ≈ 0, classical)

motor

σ {−1,+1}, sum 0

Kolmogorov

100 heads 12b vs random 46b

This work was reached as a living session: the Inventor led by intuition, the instrument computed and labelled, set no ceiling on the method, and never decided whether a problem was finally solved — that is for the scientific community. The path is kept on the record, including the corrections, because the corrections are where the work is honest.

Kolmogorov was right: probability is not the last word on randomness. Its last word is compressibility — and, this work adds, accumulation where there is genesis. The corrected law contains the classical one and extends it; it is confirmed where a system has a goal and is silent where it has none; and beneath both lies the coin that is one and zero at once, resolved only by the look. The question “how exactly will it fall” has, in the end, a single honest answer: it falls as it is seen.

This is offered not as a closed theorem but as a chain of labelled steps — stones, a model, a falsifiable hypothesis — in the spirit Kolmogorov himself asked for: test randomness, do not postulate it. No verdict on final truth is claimed; that belongs to the scientific community.

The inventor puts it as the old philosopher did. Xanthus was only the master — a philosopher by title, helpless at the wager. The thinker was Aesop, the slave. Xanthus, drunk, had boasted he would drink the sea, and staked everything; sober, he was lost. Aesop did not drink the sea and did not compute it. He gave a condition:

“I will drink the sea, as I promised. But cut off first the rivers that run into it and out of it — hold back every stream that feeds it — and then I will drink what was promised: the sea alone.”

The rivers cannot be stopped, so the sea is never drunk — and the wager dissolves. This is the exact shape of the second question. The sea is the outcome; the rivers are the period, the endless inflow and outflow of causes. To drink the sea — to take the outcome — one must first cut the rivers, take the coins as one-and-zero with the period held still. But the period is infinite; the rivers cannot be stopped. Therefore the outcome is not computed in advance: it is taken only at the cut of the look, the sea drunk only when the rivers are, for an instant, set apart.

So the answer to the auditor: the synthesis “the coin falls as you see it” is not a computed stone, because the period it would require is infinite and cannot be summed; it stands as a model — the honest name for a result that holds only under a condition that nature does not grant. The motor beneath it (the two poles, one and zero together, summing to zero) is computed and remains a stone. The reading is a model not by weakness but by Aesop’s logic: the sea is drinkable only once the rivers are cut, and the rivers do not stop.