every model spec’d & versioned · Concordance-tested changelog →

The engines, live

Everything below is computed by the same engine the API and MCP server run — the deterministic household projection, the longevity-aware Monte Carlo, and three coordination strategies — rendered through the embeddable component kit. Fixed inputs, fixed seed: every visitor sees the same numbers, and every number carries a verifiable record.

The Rivera household is fictional — invented for this demonstration. The assumptions are labeled illustrative defaults, not a market view. Nothing on this page is investment, tax, or legal advice.

A fictional household

Alex (65 in 2027) and Sam (63), married filing jointly in California. $90K cash, $650K taxable brokerage ($480K basis), $950K traditional IRA, $150K Roth. Both plan to claim Social Security at 67. Spending: $95K/yr, inflation-adjusted, over a 30-year horizon.

The deterministic projection

One path under the stated assumptions: taxes (federal + California), required minimum distributions, Social Security with spousal coordination, and withdrawal mechanics all firing year by year. The y-axis always includes $0 — a chart that hides the zero line flatters the trajectory.

Ending net worth (nominal)
$2.5M
after 30 years
Money runs out
never (this path)
Lifetime tax (federal + state)
$665K
Projected household net worth by year$0$1M$2M203020352040204520502055

1,000 market sequences, with lifespans

The same projection under 1,000 seeded return paths — sequence-of-returns risk falls out of the real cash-flow mechanics, not a formula. Lifespans are sampled too, from the NCHS 2024 period life table: the second success rate only counts a shortfall as failure if someone is alive to experience it.

Success rate (fixed 30-yr horizon)
65.4%
paths never short of covering spending
Success rate (survival-conditioned)
79.3%
shortfalls after both lifetimes excused
Alex reaches 2057 (age 95)
7.3%
of simulated lifespans
Sam reaches 2057 (age 93)
21.1%
of simulated lifespans
Simulated net worth percentile bands (p10–p90 and p25–p75 envelopes with median)$0$2M$4M$6M203020352040204520502055
  • Death years sampled from the NCHS 2024 period life table (nchs.us-life-table.2024): 2024 cross-sectional mortality, which understates future longevity improvement.
  • Deaths condition the success metric only — cash flows are not death-adjusted in v0.2: income and expense streams continue as entered, and survivor-benefit switching is not modeled.

The coordination engine: ranked, checkable decisions

Each card is a decision object: a named objective, a named baseline, every alternative scored and shown (never just the winner), the actions each implies, and what the ranking does not consider — stated on the card, because an unstated exclusion is a silent claim.

Roth conversion ladder

Objective: maximize final-year household net worth net of estimated IRMAA Part B surcharges across candidate Roth conversion schedules, each run through the full projection (score = ending net worth − cumulative estimated IRMAA, both stated per alternative)

Baseline: no Roth conversions over the horizon

  1. No conversionstop-ranked
    baseline
  2. Convert $60,000/yr for 5 years
    −$39K vs baseline · −$18K tax
    • Convert $60,000 from traditional to Roth in 2027.
    • Convert $60,000 from traditional to Roth in 2028.
    • Convert $60,000 from traditional to Roth in 2029.
    • Convert $60,000 from traditional to Roth in 2030.
    • Convert $60,000 from traditional to Roth in 2031.
  3. Convert $120,000/yr for 5 years
    −$54K vs baseline · −$220K tax
    • Convert $120,000 from traditional to Roth in 2027.
    • Convert $120,000 from traditional to Roth in 2028.
    • Convert $120,000 from traditional to Roth in 2029.
    • Convert $120,000 from traditional to Roth in 2030.
    • Convert $120,000 from traditional to Roth in 2031.
  4. Convert $250,000/yr for 4 years
    −$365K vs baseline · −$301K tax
    • Convert $250,000 from traditional to Roth in 2027.
    • Convert $250,000 from traditional to Roth in 2028.
    • Convert $250,000 from traditional to Roth in 2029.
    • Convert $250,000 from traditional to Roth in 2030.
What this ranking does and does not consider (14)
  • ending net worth: $2,501,486; lifetime tax: $664,856
  • estimated IRMAA surcharges: $0 across 56 enrollee-years (max tier 0); score = ending net worth − surcharges; IRMAA delta vs baseline: $0
  • IRMAA estimate (v0.2): CY2026 Part B thresholds and premiums (cms.partb-premium-irmaa.2026) held constant across the horizon — future brackets are unpublished and MAGI thresholds index annually; MAGI proxied from projection rows (gross income + RMD + conversion) with the SSA two-year lookback; surcharges are NOT fed back into the projection's cash flows; 2 Medicare-age years skipped (two-year lookback precedes the horizon)
  • candidate schedules are caller-specified; bracket-fill optimization is not yet modeled
  • projection: RMDs computed per owner at each owner's age (Uniform Lifetime Table)
  • projection: deficits draw cash → taxable → traditional → roth; traditional draws are taxed the following year
  • projection: realized gains on taxable draws are treated as long-term (holding period not tracked) and taxed the following year
  • projection: social_security streams with a computed-benefit block use the ssa-pia model (eligibility-year dollars, no COLA)
  • projection: state tax treats realized gains as ordinary income (no state preferential rates modeled)
  • projection: state taxable income approximated by federal taxable income
  • ending net worth: $2,462,944; lifetime tax: $646,396
  • ending net worth: $2,447,621; lifetime tax: $444,719
  • ending net worth: $2,136,707; lifetime tax: $363,653
  • estimated IRMAA surcharges: $13,637 across 56 enrollee-years (max tier 2); score = ending net worth − surcharges; IRMAA delta vs baseline: $13,637

Assumptions: profile illustrative-6-3-3@1.0.0 (default; Worthune illustration for demos and documentation — not a market view, not a recommendation. Institutions supply their own profiles.)

Verifiable record: sha256 af5481b5ff40a85f6048497056c6b0f5036ac88fa8ef2cdba0603cc16364851b — recompute from the stored response to prove this decision is unaltered

Pension election

Objective: maximize the actuarial expected present value of the pension election at the caller's discount rate (scores in USD EPV; mortality from the NCHS 2024 period table)

Baseline: the plan's single-life quote — every option priced on the same lives and rate

  1. Lump sum: $420,000 at 65 (EPV $420,000)top-ranked
    +$80K vs baseline
    • Elect the $420,000 lump sum at age 65 (in 2027).
  2. Single-life annuity: $2,400/mo (EPV $340,370)
    no change vs baseline
    • Commence the single-life annuity of $2,400/mo at age 65 (in 2027).
  3. Joint & 100% survivor: $1,850/mo (EPV $328,553)
    −$12K vs baseline
    • Commence the joint & 100% survivor annuity of $1,850/mo at age 65 (in 2027).
  4. Joint & 50% survivor: $2,050/mo (EPV $327,402)
    −$13K vs baseline
    • Commence the joint & 50% survivor annuity of $2,050/mo at age 65 (in 2027).
What this ranking does and does not consider (13)
  • expected present value: $420,000
  • delta vs single-life: $79,630 (EPV, not a projected net-worth path)
  • expected present value: $340,370
  • delta vs single-life: $0 (EPV, not a projected net-worth path)
  • EPVs at a 5% caller-supplied discount rate, valued at 2027 (participant age 65); payments annual, survival measured at year start
  • mortality from the NCHS 2024 PERIOD life table (nchs.us-life-table.2024): it understates future longevity improvement, which systematically understates annuity EPVs relative to the lump sum
  • pre-tax comparison: annuity income taxation vs lump-sum rollover treatment is not modeled
  • joint-survivor EPVs assume the two lives are independent
  • lump sums are conditioned on the participant surviving to commencement; pre-commencement death benefits are not modeled
  • expected present value: $328,553
  • delta vs single-life: $-11,818 (EPV, not a projected net-worth path)
  • expected present value: $327,402
  • delta vs single-life: $-12,968 (EPV, not a projected net-worth path)

Assumptions: profile illustrative-6-3-3@1.0.0 (default; Worthune illustration for demos and documentation — not a market view, not a recommendation. Institutions supply their own profiles.)

Verifiable record: sha256 246854a2ad51e0d1c7493442479ec1e2370e0c7a7509430d41b361964e35bd2a — recompute from the stored response to prove this decision is unaltered

Social Security claiming

Objective: maximize final-year household net worth across joint claim-age combinations, each run through the full projection (scores in ending-net-worth USD)

Baseline: current claim plan: alex at 67, sam at 67

  1. Claim: alex at 67, sam at 67 (current plan)top-ranked
    baseline
  2. Claim: alex at 67, sam at 70
    −$38K vs baseline · +$14K tax
    • Member sam claims Social Security at 70 (in 2034).
  3. Claim: alex at 67, sam at 62
    −$42K vs baseline · −$28K tax
    • Member sam claims Social Security at 62 (in 2026).
  4. Claim: alex at 70, sam at 67
    −$90K vs baseline · +$33K tax
    • Member alex claims Social Security at 70 (in 2032).
  5. Claim: alex at 70, sam at 62
    −$116K vs baseline · +$1,973 tax
    • Member alex claims Social Security at 70 (in 2032).
    • Member sam claims Social Security at 62 (in 2026).
  6. Claim: alex at 70, sam at 70
    −$137K vs baseline · +$49K tax
    • Member alex claims Social Security at 70 (in 2032).
    • Member sam claims Social Security at 70 (in 2034).
  7. Claim: alex at 62, sam at 67
    −$404K vs baseline · −$49K tax
    • Member alex claims Social Security at 62 (in 2024).
  8. Claim: alex at 62, sam at 70
    −$454K vs baseline · −$30K tax
    • Member alex claims Social Security at 62 (in 2024).
    • Member sam claims Social Security at 70 (in 2034).
  9. Claim: alex at 62, sam at 62
    −$459K vs baseline · −$73K tax
    • Member alex claims Social Security at 62 (in 2024).
    • Member sam claims Social Security at 62 (in 2026).
What this ranking does and does not consider (17)
  • ending net worth: $2,501,486
  • the deterministic projection has NO mortality: members do not die, so longevity-insurance value beyond the horizon and survivor benefits are not captured — this ranking holds under this horizon and these assumptions only
  • pension elections are handled by the pension-election strategy, which prices options on the NCHS mortality table
  • projection: RMDs computed per owner at each owner's age (Uniform Lifetime Table)
  • projection: deficits draw cash → taxable → traditional → roth; traditional draws are taxed the following year
  • projection: realized gains on taxable draws are treated as long-term (holding period not tracked) and taxed the following year
  • projection: social_security streams with a computed-benefit block use the ssa-pia model (eligibility-year dollars, no COLA)
  • projection: state tax treats realized gains as ordinary income (no state preferential rates modeled)
  • projection: state taxable income approximated by federal taxable income
  • ending net worth: $2,463,167
  • ending net worth: $2,459,675
  • ending net worth: $2,411,761
  • ending net worth: $2,385,519
  • ending net worth: $2,364,422
  • ending net worth: $2,097,769
  • ending net worth: $2,047,008
  • ending net worth: $2,042,914

Assumptions: profile illustrative-6-3-3@1.0.0 (default; Worthune illustration for demos and documentation — not a market view, not a recommendation. Institutions supply their own profiles.)

Verifiable record: sha256 eadd9340704563b4d7cb77e6524b01c446ca9f53f7233cec2b405a22e8d0ce92 — recompute from the stored response to prove this decision is unaltered

Run this yourself

Everything above is one POST /api/v1/households/{id}/project and three POST …/decisions calls — or the same through the MCP server, or the worthune SDK on npm and PyPI. Every response carries the evidence record shown in the disclosure block; recompute the hash any time to prove the numbers are unaltered.