The Unicorns
What no theory can owe — and what this one owes instead.
There is a demand this site will never satisfy, and this page exists to say so plainly rather than let you discover it and count it as a defect:
Neither can anyone else, about anything, and the demand itself is the mistake. Internally this project calls such demands unicorns — requests for an animal that doesn't exist — and one gate on the board already carries that verdict openly: SG-1 explicitly does not claim that its thirteen-dimensional shape is the unique geometry nature could have used.
That's not a hedge added by lawyers. It's the honest shape of the problem, and here is the argument in full.
The demand, at full strength
“You've built an elaborate scoreboard, but you can't prove your candidate is the best one, or even a correct one. The space of possible geometries is infinite — you've checked a handful. Until you can show yours is right, why should anyone care?”
Taken seriously, this objection doesn't just sink this project. It sinks physics.
No theory in history has met it. The Standard Model — the most precisely confirmed theory humans have produced — was never proven correct; it was proven consistent with every measurement so far, which is a different thing, and its nineteen-plus free parameters were never proven to be the right ones, only measured. General relativity was never shown to be the unique theory of gravity; it was shown to keep passing tests that killed its rivals.
The reason is structural, and it has been understood for a century: any finite set of observations is compatible with infinitely many theories, so verification of a physical theory — proof that it is true, or best — is not a hard goal. It is not a goal at all. It is a category error, like demanding the largest integer.
An infinite search space makes this worse, not better, for anyone hoping to sneak the claim in through effort: you cannot exhaust what cannot be enumerated. When this site compares its candidate against alternatives, it therefore does the only honest thing available — it declares the exact shortlist being compared, in advance, and when that shortlist is not provably complete, the comparison is labeled exactly that: open, exhaustion not proved.
A comparison that names its shelf is information. A comparison that gestures at “all reasonable alternatives” is theater, and you will not find one here.
What replaces the impossible demand
If proof-of-correctness is off the table for everyone, what distinguishes a serious candidate from an arbitrary one? One asymmetry, and it is the engine of this entire site:
A theory can never be proven right. It can absolutely be proven wrong.
That asymmetry is not a consolation prize — it is the entire mechanism by which physics makes progress through an infinite space it cannot search. You do not find the right theory by verifying it. You prune, and what survives pruning under the harshest tests you can build earns provisional standing — never proof.
The flagship flavor prediction was falsified at 15.9 standard deviations, and the failure is preserved on the board, not repaired away.
The vacuum of the original construction was shown to be unstable — a saddle, not a resting point — and that result stands unretracted.
The first attempt at the three-family count was found to be counting the wrong object entirely, and the failed version is kept alongside the repair.
The project's own review machinery found that the central stabilizing assumption was chosen rather than derived, and recorded that as a failure of the stronger claim.
A framework built to flatter its candidate does not produce that record. This one did, because pruning is the point.
What a theory actually owes is not proof. It owes a stated way to die on every claim; honest labels on everything assumed; comparisons that name their shortlist; and, eventually, at least one number that comes out right without having been put in.
That last debt is real, it is currently unpaid, and the companion page on dissolution says the same about its own axiom. Unpaid debts stated in public are how this site does business.
The objection this position still has to answer
“Then you've excused yourself from every standard”
“How convenient. ‘Nobody can prove correctness’ becomes a universal alibi — every failure gets waved through because hey, no theory can be proven right anyway. You've built an unfalsifiable position and dressed it as humility.”
No. The unprovability of correctness is a statement about which questions exist. It is never usable as evidence about how this candidate performed — and that rule is absolute here. When a calculation fails, it fails; “well, nothing can be proven anyway” has never appeared next to a failure on this board and never will.
Unprovability of correctness is not evidence of adequacy.
The demand this page refuses and the demands this site accepts point in opposite directions. Refusing to claim “best” raises the price of everything that remains checkable, because checkable results are all there is.
That is why every gate carries explicit conditions that would kill it; why every measured input is labeled as measured rather than smuggled in as a success; why negative results are kept and displayed rather than deleted; and why the one comparison that could someday upgrade a candidate's standing — a genuine prediction, confirmed — is named here as the debt it is.
Two different objects, two different standards
One more distinction, because the objection usually blurs it. This site contains two kinds of thing, and “correct” applies to them differently.
The geometry is a candidate
It is a claim about nature. It can never be verified, can always be falsified, and its current honest status is that it reconstructs known physics from chosen ingredients rather than predicting anything new.
Every gate that closes says “given the declared construction” somewhere in its fine print, because that is what is true.
The building blocks are a harness
They are not claims about nature at all — they are the test rig: the list of requirements any candidate must face, with pass conditions, required evidence, and failure triggers.
A harness is not “correct” or “incorrect.” It is sound — does it catch known errors? — which is tested here by deliberately feeding it constructions that must fail and checking that they do.
It is also complete or not — does it test everything that matters? — where the honest answer is no, and it never will be: the requirement list grows only when an actual execution exposes an actual gap, and both times that has happened, the gap and the new requirement it produced are documented.
A harness proves nothing right. Its entire job is to prove things wrong, faster and more reproducibly than an unaided reader could. Demanding that it be provably complete is the same unicorn in a different pasture.
What is actually assumed
If the currency here is honest labels, this is the ledger. These are the assumptions — the things taken rather than derived — across the whole construction, in plain language:
1The finite floor
Physically meaningful distinctions are those recordable at finite resolution. This is the axiom behind every dissolution on the board; it has its own page, its own justification, its own refuter, and its own openly unpaid rent.
2Noncontextuality of measurement
The probability rule of quantum mechanics is recovered on this board given one assumption about how measurement outcomes can depend on context. The framework itself proved this assumption cannot be derived from the geometry — so it is named as a pin, not disguised as a result.
3The frozen internal shape
The candidate's small internal geometry is treated as fixed background structure rather than a dynamical thing — because when it was allowed to move, it was unstable, and that instability is a preserved result. Freezing it is a choice, lawfully implemented and honestly labeled as chosen rather than derived.
4The chosen ingredients
The specific geometry itself, from a declared shortlist; the Higgs field it carries; the structure that yields three families; the rule that protects the proton; and the field that relaxes the strong-CP angle. Each is a construction — selected because it satisfies the requirements, never claimed to be forced by them.
5The measured dials
The Planck mass, the weak scale, and the force strengths at laboratory energies. They are read from experiment, labeled as read from experiment, and never counted as successes.
Everything else on the board — the charge arithmetic, anomaly cancellations, counting results and stability analyses — is derived given the items above, and the “given” is always written out.
Assumptions are named, and consequences are checked. Nothing here is claimed to be the best. Everything here is claimed to be labeled.
What would prove this page wrong
If this site ever claims its geometry is unique, proven, or best; if a comparison ever hides its shortlist or presents an unexhausted search as exhausted; if any assumption listed above is ever presented as a derivation; or if “correctness can't be proven” ever appears as a response to a failed calculation — then this page has failed at its one job, and it says so here, in advance, in writing.