Scenarios — decision workspace

Composed from this surface's own capability modules. 10 of 24 inventory capabilities are green under this surface's own tests; the rest are held by published needs and are listed as held rather than mocked.

Start here — which of these is closest to you?

These three are the people this instrument opens on, and what differs between them is shown below. What each would actually end up with is not shown, because computing it needs the simulator capability, which is not published yet — and this surface will not invent it.

What differs between them, and why each difference matters:

Outcomes for these three are not shown: held by simulator.

Plan

Life events
Custom component
Glidepath
Leverage cap

Commit, share, cross

Draft

Resilience

Stress presets

Selection only. Every selected preset is compared against the baseline — the no-shock projection, which is not itself a preset — and that comparison is a stress evaluation this surface does not compute.

How each preset is anchored

Optimize

The order is the argument. A result shown without the naive run it improved on is a number a reader has to take on trust; the same result shown second, against a baseline nobody argues with, is evidence. No control here yet — both runs are held by a published need, and this surface states the sequence rather than inventing a figure to put in it.

  1. Run it naively The plan under a recognisable default — a constant allocation by preference, a target-date glidepath by choice — with its band, on the draws every later arm is compared against. Held by policy-batches.
  2. Pin the goal already held Wealth, time and confidence are not independent, so the goal a reader already has is fixed at the confidence they already hold. Nothing is being chosen for them at this step; the point of it is that nothing is. Held by wtc-solver.
  3. Free wealth or time, holding confidence With confidence held, one of the other two is allowed to move: more wealth for the same date, or the same wealth sooner. Where both improve, the time reading leads — years are a unit nobody has to be taught. Held by policy-batches.

Analysis

Honest async

Resolve refuses by name. A result would be a simulator output, and this surface does not compute one or invent one.

State

Workspace not started — this page is static until its module runs.



Capabilities

Status is read from this surface's own test corpus at build time, not written by hand. "Held by" names the published need that is outstanding.

#CapabilityStatusOn this pageHeld by
1draft/commit/revertgreendriven here—
2historical lens + inflation bridgepartialnot wired here yetdeck-statistics
3life eventsgreendriven here—
4loan eventgreendriven here—
5custom event componentsgreendriven here—
6glidepath control pointspartialdriven herekit-relay
7plan-vs-policy modepartialnot wired here yetpolicy-batches
8WTC pick-two-solve-thirdpartialnot wired here yetwtc-solver
9leverage capgreendriven here—
10optimize (CMA-ES / tabular-RL)partialnot wired here yetpolicy-batches
11resilience across stress presetspartialdriven herewtc-solver
12balance-sheet attributionpartialnot wired here yetsimulator
13honest async per sectiongreendriven here—
14share permalinkgreendriven here—
15concentration navigationgreendriven here—
16Roth Ledger surfacepartialserved on another surfacefront-door-relay
17Roth One Question surfacepartialserved on another surfacefront-door-relay
18Roth Hinge surfacepartialserved on another surfacefront-door-relay
19startup equity liquidity A/B surfacepartialserved on another surfacefront-door-relay
20Big-5 onboarding entrypartialserved on another surfacefront-door-relay, big5-concepts
21a visual driven by real valuespartialserved on another surfacekit-relay
22one Roth design linked, not threepartialserved on another surfaceroth-single-entry
23every variant capturedgreenserved on another surface—
24the generations gallerygreenserved on another surface—

What this artifact is not

It carries no simulation, no solver and no optimizer: those are consumed capabilities, and the ones this surface needs are not published yet. Nothing here renders a projected number. The component layer is a separate concern's, so the presentation is deliberately plain rather than a stand-in for it.