Attempt cash ledger
Reconcile entered posted and pending cash events by attempt and category, retaining earlier shortfalls.
One unit and one cutoff
The starting value belongs to an external cash journal, not the advertised evaluation account. Its as-of time bounds the supplied observations. A later event requires a different scenario and cannot silently enter the earlier balance. The authored packet uses fictional XCU units; a real currency label would still not authenticate a payment. Currency conversion, taxes, opportunity cost and unentered obligations are outside this calculation. Keep the population and status source with the interpretation. A real product card may explain whether a charge applies, but that is different from a supplied cash row stating whether it has been posted.
Four status-direction buckets
Each non-void row contributes to one of posted debits, posted credits, pending debits or pending credits. Fee, reset, activation and data are debit categories; refund and payout are credit categories; other permits either direction. These conventions do not establish that a programme charges a particular fee or grants a refund. A void row remains in the trace without changing totals. A category name is not evidence of settlement. A pending payout is not current posted cash, and a posted credit remains a supplied assertion rather than an authenticated receipt. The category relationship prevents contradictory input without filling missing payment evidence.
A posted path with retained shortfalls
Only posted rows change the running balance. The minimum includes starting cash and every subsequent posted balance. Negative posted observations retain their event identities after later recovery. Tied times preserve entered order, which is a scenario convention rather than proof of subsecond sequence. The low-start teaching case demonstrates that positive ending cash does not erase earlier shortfalls. Pending and void rows show the current balance in the trace but cannot create new posted-shortfall identifiers. This is an external-payment path, not trading equity, buying power or a reconstruction of exposure. A real chronology needs authoritative payment timestamps and statuses.
Pending commitments are a separate snapshot
Cash after pending debits equals current posted cash minus entered pending debits. It intentionally does not add pending credits. The amount is an entered commitment snapshot, not a time-distributed forecast or recommended spending limit. A large pending debit can make that snapshot negative while every posted observation stays positive. That differs from a posted shortfall. No probability of settlement or expected arrival date is estimated. A later status change needs a later observation record. A possible credit should not become a reason to spend merely because a scenario has formatted it neatly alongside actual posted amounts.
Groups reconcile without selecting attempts
Each attempt and category has the same four buckets as the whole journal. Groups reconcile to the total without treating a reset as a strategy result. Opaque A1/A2 organize the teaching population. A complete actual attempt list remains separate because an omitted attempt cannot be detected from supplied rows. Zero posted charges do not justify omitting a group containing pending or void observations. These totals describe the entered population, not typical participant costs or success. A favorable credit on one attempt should not erase the charges on another. Preserve the declared period before comparing any resulting amounts.
Exact decimals, bounded evidence
Amounts are plain positive decimal strings up to one trillion units with at most six fractional places. Starting cash may be zero. The decimal case keeps 0.1 and 0.2 exact rather than relying on a binary approximation. Duplicate event identities, invalid dates, out-of-order observations and conflicting category directions are rejected. Those checks establish input and arithmetic consistency, not authenticity. Account approval, payout eligibility, continuous monitoring and strategy validation remain null. A technically valid journal may still be incomplete or wrongly classified. The exported inputs preserve that limited computation for review without turning it into an expected-return model or affordability determination.