Status: Proposed, 12 September 2026. Not decided. Awaits the contract engineer and a reviewer (blueprint R05). Nothing here is production configuration. The figures the simulation uses (10 and 20 units per lot, a cap of 1,000 lots) are illustrative and are labelled so wherever they appear; the method below is what is proposed, and its inputs are not yet known.
Context
A lot is the unit of a position: a fixed number of base units of A and of B. Choosing those numbers decides how a holder's rights can be computed with no oracle (§7), how large the smallest position is, how much of a small wrapper's supply one lot would be, and how much the series may owe in total.
Decision (proposed method)
1. Integer base units, fixed at construction. qA and qB are positive integers of each component's own base units (both candidates have 18 decimals as read on chain for A; B's decimals are not yet read). They are immutable (ADR-001). 2. The receipt has 0 decimals. One receipt is one whole lot. Fractional lots do not exist, so no rounding exists in mint, exit or claims: every claim is lots × q, exactly. 3. Lot size is chosen so that one lot is economically small and unit-round. Proposed: choose qA and qB so that one lot corresponds to a comparable economic exposure to the company through each component at the time the series is made, rounded to whole units of each token where the token's unit is a share-equivalent, and to a round number of base units otherwise. Because the two components' units are not equal (A's wrapper share and B's total-return unit each drift from one share over time), a lot is not one share and not one dollar, and the ratio between qA and qB is fixed for the life of the series even as the two components' values diverge. This is the "no rebalancing" rule seen from the lot's side. 4. The cap counts liability, reserved included. capLots bounds n + (reserved / q); a lot allocated for exit still counts until its claims are paid. The cap for a first pilot is proposed as a small fraction of the smallest component's circulating supply — the fork evidence shows the current wrapper for A holds about 10.16 raw tokens and has about 10.13 shares in all, so any cap in whole lots of 10 units would be *the entire wrapper*; that is a gate G6 fact, not a configuration. 5. Decimals of the components are read, not assumed. The verification reads decimals() daily; a change is a DARK drift and a reason to stop mints.
Consequences
- —A holder's rights are computable from
lots,qA,qBalone; the site's previews and the ledger model do exactly that. - —Small holders may be excluded by the lot size; that is a product decision to be tested in interviews (R04), not something the contract can soften without fractional lots.
- —If the wrapper for A stays as small as it is, no cap makes a pilot sensible on that component; the candidate would have to change or the pilot wait.
What exists today
The contract takes qA, qB, capLots as constructor arguments, refuses zero and overflow, counts reserved liability against the cap (T05); the receipt has 0 decimals; the ledger model and simulation use illustrative values; the fork evidence records the wrapper's size.
Open
- —The actual
qA,qBand cap for a pilot — inputs missing: B's address and decimals, both components' unit economics on a chosen date, the interview results. - —Whether the cap should be lowered by the operator during a pilot (today: immutable; lowering would need a new series or a mint pause).