Sync docs/project-layout.md backtest example to the bias/R/cost-model reframe (#117 + #116) #151

New Issue

2026-06-28T14:49:56+02:00

Brummel commented

2026-06-28 14:49:56 +02:00

The "Backtest" worked example in docs/project-layout.md (steps 3-4, ~lines 95-110) is stale on two retired reframes:

#117 (exposure → bias): it still describes an "exposure stream" and the sim-optimal broker integrating exposure·return into pip-equity; the live model is an unsized bias stream scored in R.
#116 (2026-06-28 C10 rework): it still sells "a realistic broker reads the derived position-event table" with --broker pepperstone producing a realistic currency curve. The realistic broker is retired — cost is now the cost-model graph (in R: net R = gross R − cost-in-R, the ConstantCost node + net_r_equity tap, shipped cycle 0081), and money / a real broker live only at the later live/deploy edge (#149).

Rewrite the backtest + experiment worked examples to the bias / R / cost-model-graph model (including the {sim-optimal, pepperstone} matrix axis in step 4). While here, sweep the rest of docs/ prose for sibling exposure / realistic-broker drift.

context: surfaced by the cycle-0081 audit (architect drift review); pre-existing, predates the cost-model cycle. Not a code defect — docs-only.

The "Backtest" worked example in `docs/project-layout.md` (steps 3-4, ~lines 95-110) is stale on two retired reframes: - **#117 (exposure → bias):** it still describes an "exposure stream" and the sim-optimal broker integrating `exposure·return` into pip-equity; the live model is an unsized **bias** stream scored in **R**. - **#116 (2026-06-28 C10 rework):** it still sells "a realistic broker reads the derived position-event table" with `--broker pepperstone` producing a realistic currency curve. The realistic broker is **retired** — cost is now the **cost-model graph** (in R: `net R = gross R − cost-in-R`, the `ConstantCost` node + `net_r_equity` tap, shipped cycle 0081), and money / a real broker live only at the later **live/deploy edge** (#149). Rewrite the backtest + experiment worked examples to the bias / R / cost-model-graph model (including the `{sim-optimal, pepperstone}` matrix axis in step 4). While here, sweep the rest of `docs/` prose for sibling `exposure` / realistic-broker drift. context: surfaced by the cycle-0081 audit (architect drift review); pre-existing, predates the cost-model cycle. Not a code defect — docs-only.

Brummel referenced this issue from a commit

2026-06-28 15:26:33 +02:00

audit(0081): cycle close — drift-clean (code); ledger C10 realization note + stale-broker doc sync (refs #148)

Brummel referenced this issue

2026-06-29 02:33:50 +02:00

Cost-flag CLI ergonomics: units hint, negative-rate diagnostic, non-R-harness handling (+ C10 spec_gap) #153

Brummel referenced this issue from a commit

2026-06-29 10:20:56 +02:00

fieldtest(milestone): cost-model graph (#148) — capability delivered green

Brummel referenced this issue from a commit

2026-06-29 12:22:46 +02:00

docs(project-layout): retire exposure/realistic-broker framing → bias/R/cost-model