369 Paper Entries: 361 Closed, 8 Open, $0 Realized

The May 2026 Distressed paper-test aggregate contains 369 entries: 361 closed at the simulated 1¢ entry price, eight remained open, no closed position hit the target, and realized paper P&L was $0. That is a failed test of the intended exit behavior, but it is not 369 realized losses.

Correction, September 9, 2026: the original headline described 369 losses, and later prose still inferred intraperiod prices, inevitable losses, and a successful replacement strategy from an aggregate that cannot establish them. This revision removes those conclusions. The historical counts remain unchanged. We do not have a new validated outcome for the proposed replacement.

What the retained record supports

The numerical account uses the retained aggregate CSV, whose basis column points back to the previously published operator summary. It preserves that account; it does not independently corroborate the original database. This was a separate historical strategy simulator, not the current Bot Builder Paper Account, and the export is not a trade-by-trade execution ledger.

The aggregate labels zero trades as moving off 1¢. Without the underlying tick history, we interpret that narrowly: the export does not establish an exit above the entry price. It cannot prove that no executable quote ever changed between entry and close.

What the strategy was trying to test

The historical configuration described a scanner for cheap Bitcoin event contracts, with a 1¢–10¢ ask filter, a time-to-close window, five contracts per entry, up to eight concurrent positions, and a 90¢ exit target. These are the operator's recorded configuration notes. The aggregate itself does not independently reproduce the scanner, selection logic, or point-in-time inputs.

The hypothesis was that some selected cheap contracts would move enough to reach that target. The closed sample did not show that behavior. Price alone does not measure mispricing, and choosing the cheapest candidate does not establish that its probability was underestimated.

We cannot infer from this sample that every cheap contract is untradeable, that another threshold would have earned money, or that a particular volatility filter would have fixed the strategy. Each proposed change is a new hypothesis with a new test requirement.

What the aggregate cannot answer

• Which source response and market rules were available when each candidate was selected.
• Whether a quoted price had enough executable depth for the requested size.
• Why each simulated close used the entry price, and how missing quotes were handled.
• The price path between evaluations, queue position, partial fills, rejections, and latency.
• A complete live-equivalent fee, spread, and settlement reconciliation.

These are missing observations, not zero-valued metrics. The old article's precise volatility and price-path claims exceeded this evidence. More explanation cannot repair missing records; the next test must capture them.

The review guard we needed

Before another run, preserve a rule version, exact market identity, input timestamps, selection and skip reasons, order intent, simulated fill assumptions, closed/open counts, and reconciliation status. Set a review checkpoint in advance. Reaching a checkpoint should prompt inspection of the sample and failure modes; it should not automatically loosen the rule or authorize live operation.

A correct skip is a useful result. If an entry ceiling rejects the current ask, inspect that receipt before changing the ceiling. An error or missing source is different from a successfully evaluated false condition. Our Operator Note on reading a Paper decision walks through that distinction.

The earlier article described a replacement with a higher entry floor, a strike-distance filter, a volatility gate, and tiered exits. Those remain unvalidated ideas in this record. This revision makes no claim that a replacement is currently running or has improved results.

Inspect a smaller rule in Paper

The Paper limit-entry and exit-target recipe is a separate, authored Builder exercise. It does not recreate the historical exchange-wide Distressed scanner. Select one real market, review one contract at a 50¢ entry limit and an absolute 60¢ sell target, and inspect the supported order semantics before saving the stopped draft. The target is an order instruction, not a promised fill or a stop-loss. It keeps one open position at a time; after an exit, it may enter again if the same conditions hold and the five-minute action cooldown has elapsed.

For a reproducible gate check, we supplied the registered evaluator a synthetic market named KXRECIPE-26SEP09-DEMO with a YES ask of 55¢. The export labels that synthetic observation September 9, 2026 at 12:00 UTC; the price-only evaluator does not read that clock or check the close window. Against the 50¢ entry condition, the observed result is false. Download the observed condition output and compiler fingerprint. The file labels synthetic inputs separately from its actual generation timestamp.

This check covers the price gate only. It does not reproduce the old test, evaluate the complete runtime, demonstrate funded Paper execution, or establish profitable live performance. In your own connected account, review the full draft and its first explained Paper evaluation before drawing conclusions about how it behaves.

Methodology: historical numerical statements above come from the retained aggregate; historical configuration details come from the operator notes; the new condition check uses explicitly synthetic inputs. The correction preserves those boundaries so later work can build on what was observed without inheriting unsupported conclusions.