Why identity rather than possession
The loop writes to the risk record — each strategy's unrealized figure, and the account's. A lock that merely exists would let the loop and the record's own readers hold different locks and write at once.
So the check is identity: the object handed in must be the object the record carries. A record carrying none at all fails it too.
The case it does not check
A loop given no lock. The check begins with the lock being there, so a loop built without one is built without the question being asked — and then nothing stands between its writes and anyone else's.
The interval
A whole or ordinary number, with a true-or-false refused outright by a test of its own, and compared with nought. Finiteness is not asked, so a figure that is no number by value passes the comparison and is kept — the same gap the health loop and the checkpoint timer leave.
What is kept is the figure as a plain number.
Related
Source evidence
---
row: N52-01
baseline: 49aa659
created: 2026-11-08 17:43 UTC
modified: 2026-11-09 03:10 UTC
evidence:
- claim: "Where a positions lock is given, it must be the very lock the risk record holds"
source: "`nexus/core/mtm_loop.py` 87-90 into 91-92 and 107"
- claim: "A risk record carrying no lock at all fails that same check"
source: "`nexus/core/mtm_loop.py` 88, which asks whether the record has one"
- claim: "A loop given no lock is built without the check being made"
source: "`nexus/core/mtm_loop.py` 87, whose test begins with the lock being there"
- claim: "Its interval is refused unless it is a whole or ordinary number, with a true-or-false refused outright, and refused at or below nought by comparison"
source: "`nexus/core/mtm_loop.py` 79-85"
- claim: "What is kept is that interval as a plain number"
source: "`nexus/core/mtm_loop.py` 111"
---