ForecastBudgetsSubscriptions

Rest-of-year forecast

Monte Carlo over the 30 complete weeks remaining in 2026, drawn from the last 52 complete weeks of spend (as of 2026-05-28).

Spent YTD
$36,587
Monthly average
$7,317
Projected EOY (P50)
$95,524
P10 — better than this
$51,787
Median (P50)
$58,937
P90 — worse than this
$66,826
Mean
$59,174

Distribution of remaining 2026 spend

P10P50P90$39,747$84,575

Projected rest-of-year by category

The Spent YTD column shows what you've already spent this calendar year. Pin an annual commitment on any row (e.g. "$5k for the year") and the rest-of-year drops to annual − YTD — what's left of your commitment. Pinned categories are excluded from the bootstrap; everything else uses YTD pace × weeks remaining (or falls back to T12 prorated when the year is too young to have a stable pace). Lumpy categories always use T12 prorated.

CategorySpent YTDProjected RoYSourceSet annual $
rent$14,250.00$21,375.00YTD pace
travel PINNED$6,000.00annual − YTD
restaurants$3,865.61$5,798.42YTD pace
groceries$3,747.27$5,620.90YTD pace
shopping$3,347.98$5,021.97YTD pace
household_goods$2,625.37$3,938.06YTD pace
education$1,724.80$2,587.20YTD pace
taxes LUMPY$2,451.92T12 prorated
utilities$1,483.50$2,225.25YTD pace
transportation$1,481.94$2,222.91YTD pace
charity PINNED$1,800.00annual − YTD
insurance LUMPY$1,380.00$1,592.31T12 prorated
coffee$694.59$1,041.88YTD pace
uncategorized$659.45$989.18YTD pace
personal_care$432.00$648.00YTD pace
subscriptions$343.80$515.70YTD pace
fitness$195.00$292.50YTD pace
entertainment$181.76$272.64YTD pace
health$173.81$260.72YTD pace
Total$36,586.88$64,654.56

How this is computed

10,000 simulated rest-of-2026 totals. Each draw samples 30 weeks with replacement from the trailing 52 weeks of expense totals (combined accounts) and sums them. Lumpy categories (insurance, taxes) contribute $4,044 over the remaining horizon as a deterministic add (annual T12 × 30/52) — they don't bootstrap. User-pinned categories (charity, travel) contribute $7,800 as a deterministic add (annual commitment − YTD already spent, clamped at zero). Their weekly history is excluded from the bootstrap — you know the number, the model defers to you. P10 / P50 / P90 are percentile boundaries of the simulated totals.

Weekly resampling sits between monthly (only 12 samples — too noisy) and daily (~365 samples but high autocorrelation makes effective sample size much smaller). Trade-off: monthly bills land in ~30/4.33 of the 30 sampled weeks on average — right expected count, with phase noise that shrinks proportionally as the horizon grows.