Paper trading involves no real capital and therefore no real execution or real psychology. This article is precise about which claims a public paper record supports and which it never can, including for our own future records. Nothing here is investment advice. Decisions are yours.
Paper trading has a terrible reputation, and it has earned it. Private paper records are worthless, because nothing stops their owner from remembering the wins and forgetting the rest. But add one ingredient, cryptographic attestation of every decision before its outcome, and paper trading becomes genuinely useful: the cheapest honest test that a process exists and is being followed. The trick is knowing exactly what that test does and does not prove.
The evidence ladder
Rank the kinds of evidence a strategy can offer, weakest to strongest. A backtest proves the rules would have worked on history, subject to every bias this cluster documents, survivorship, overfitting, fill assumptions, as our own audit found the hard way. An attested paper record proves something a backtest never can: the rules produce their outputs in real time, before outcomes are knowable, with no lookahead possible even in principle, because the timestamps are anchored outside the publisher's control. Small live capital adds real fills, spreads, and slippage. Audited live size adds capacity and the psychology of consequential money.
Each rung answers an objection the rung below cannot, and an honest label states its rung. The industry's chronic sin is selling rung-one evidence in rung-four language. The opposite sin, dismissing rung-two evidence because it is not rung four, throws away the one test that is nearly free.
What attested paper publication actually proves
Be precise here, because this is the rung our derivation program will occupy first. A paper record whose every daily output is hash-committed at publication, by the mechanics described here, proves four things. The process exists: outputs arrive on schedule, every day, including the boring ones. It is mechanical: signals follow the pre-registered rules, which an outsider can check once those rules are revealed, the same discipline as walk-forward validation. There is no lookahead: each output demonstrably predates the prices that judge it, which kills the backdating and curation exploits that make private records worthless. And the record is complete: the chain shows every day, so losing stretches cannot be quietly omitted.
Those four properties are exactly the ones the signal industry's records lack. Which is why an attested paper record, humble as it sounds, already clears a bar that almost no paid product in this market clears.
What it cannot prove, stated adversarially
Now the other side, argued the way an opponent would argue it. Paper fills are fictions: marking a signal at the daily close assumes liquidity thin alts may not offer, ignores spread and slippage, and never meets a fast market's reality, so paper performance is an upper bound on live performance, always. Paper positions carry no capacity information, since a process that works at notebook scale may move its own prices at real scale. Paper psychology is not psychology: nobody has ever panic-closed a spreadsheet cell, and the discipline a paper record demonstrates is the software's rather than the operator's, though for a fully mechanical system that objection loses most of its force. And a short paper record proves nothing about regimes it has not met, because eight weeks of risk-on tape tests nothing that matters most. Any presentation of paper results that does not volunteer these limits is wearing the record as a costume.
Why we require eight attested weeks anyway
Given those limits, why does our pre-registered gate demand a minimum of eight weeks of attested paper publication before any strategy claim? Because the gate is not trying to prove profitability. It is trying to prove operational honesty before money or marketing enters.
Eight weeks forces the pipeline to survive contact with reality's calendar: data outages, weekend cascades, boundary conditions, the temptation to fix something quietly, all with the chain recording whether the rules held. It generates a public, complete, real-time sample to set against the validation backtest, the first genuine out-of-sample evidence that money cannot buy retroactively. And it establishes the habit that matters most, publishing outputs before outcomes, in the same lane that will carry live results later, which is the whole measurement posture. Eight weeks is a floor, not a target. The record simply never stops, and the label upgrades only when the rung does.
How to read anyone's paper record, including ours
The checklist compresses to five questions. Is every entry attested, or merely claimed? Is the record complete, or a highlight reel? Are the fills marked conservatively, at prices the stated size could plausibly have achieved, with costs deducted? Does the presentation state its rung on the evidence ladder in plain language, paper meaning paper? And what regimes has the record actually met?
A publisher who answers all five without being asked is treating you as the auditor you should be. When our own derivation candidates enter their paper period, this article is the standard their presentation has to clear, and you are invited to hold the two against each other line by line. Paper trading in public is not proof of profit. It is proof of process, and in a market drowning in claimed profits and absent process, that is the scarce thing. Decisions are yours.