There's no silver bullet here. Refill doesn't reliably beat a plain waterfall on raw survival odds - what it offers instead is a smoother, more protected spending experience along the way. Every strategy on PortfolioCube makes its own trade-off; the right one depends on your own circumstances, not a single "best" answer.
Refill Strategy Lab: Dynamic Bucket Rebalancing for Retirement Withdrawals
A free tool to test a three-bucket retirement strategy that periodically refills your Safety and Income buckets from Growth - a PortfolioCube Framework - and compare it side by side against a plain, drain-only waterfall, using the same Monte Carlo stress-testing this platform's Advance Retirement Calculator is built on.
Why a single retirement corpus number isn't the same as a withdrawal plan
Most retirement planning stops at one question: do I have enough? This tool starts from that same Required Retirement Corpus check - using the same real-rate, Fisher-equation methodology as the Advance Retirement Calculator - but treats it as the beginning of the real question, not the end of it. Even a fully-funded corpus still has to survive being drawn down every month for two, three, sometimes four decades, and *how* it's drawn down - what gets spent first, what's left untouched the longest, and whether anything ever gets replenished - changes how long it actually lasts, independent of whether the starting number was ever big enough.
The three buckets: Safety, Income, and Growth
This tool organizes a retirement corpus into three buckets, each carrying a different investment character. Safety (B1) is low-volatility, debt-like money - built from assets like EPF and Fixed Deposits - meant to fund the earliest years of retirement without ever needing to be sold during a market downturn. Income (B2) sits in between - Conservative Hybrid-type exposure (modeled here through NPS), carrying more return potential than Safety with meaningfully less volatility than pure equity. Growth (B3) is the corpus's equity-and-other-growth-asset engine, given the longest possible runway to compound before it's ever touched. The names are deliberately about *role*, not asset label - a bucket strategy is a sequencing decision, not just a relabeled asset allocation.
Plain Waterfall vs. Refill: two different rules for what happens once a bucket runs low
A plain waterfall follows one simple rule: Safety is drained fully before Income is touched at all, and Income fully before Growth is touched at all - and once a bucket is empty, it stays empty for the rest of retirement. Refill Strategy keeps the same three buckets and the same underlying sizing, but adds a second rule on top: at a periodic checkpoint, Income tops Safety back up to its original target, and Growth tops up Income - but only in a checkpoint where Growth itself posted a return above a configurable threshold, so it never sells at a loss just to hit a schedule. If a checkpoint can't be fully funded even after that, a Crisis Rule decides what happens next: Capitulation draws whatever's needed from Growth regardless of its recent performance, treating Safety's spending as non-negotiable; Corridor-Cut instead permanently lowers the spending target itself - inflated forward, haircut applied, and never allowed below an absolute floor you set. An optional Stress Test can also force Growth through a fixed multi-year crash, to see how each rule behaves under a genuine downturn rather than just an average one.
Deterministic projections vs. Monte Carlo simulation: two different questions, not two versions of the same answer
Both the Refill and Waterfall cards on this page show every scenario two ways. The deterministic chart and audit table assume each bucket earns exactly its own average return, every single year - useful for seeing the basic shape of a drawdown, but it structurally cannot show what a bad sequence of returns does to a plan, since it has no variation to be bad. The Monte Carlo panel runs 1,000 randomized trials instead - Safety, Income, and Growth each drawing their own return around their own volatility every year - and reports the share that survived the full retirement horizon. It's common, and expected, for these two views to disagree meaningfully for the same inputs; the gap between them is itself the measure of how much a plan's real-world survival depends on sequence of returns, not just average return.
Refill vs. Plain Waterfall: what the comparison on this page actually shows
Card 7's own comparison panel runs both strategies on the exact same starting buckets and the exact same Return and Volatility assumptions, so any difference in outcome is attributable to the mechanism, not to one side getting luckier draws than the other. Tested across a wide range of settings, the honest finding is this: Refill does not reliably produce a higher Monte Carlo success rate than a plain waterfall - in most tested scenarios, the waterfall's own success rate is at least as good, sometimes meaningfully better, because leaving Growth completely untouched for as long as possible is itself a real, measurable advantage that periodic refilling trades away in exchange for something else. That something else is real too: smoother year-to-year spending, protection against a discontinuous cut to Safety's own withdrawal, and a Growth bucket that's never forced to liquidate at a bad moment specifically because a schedule demanded it. Neither strategy is a strictly better version of the other - they optimize for different things, and the right one depends on what you're actually trying to protect against, not a single headline number.
Does the refill checkpoint's own timing matter? Comparing three rebalancing cycles
Refill Strategy's own checkpoint doesn't have to fire every year - this page's underlying model was tested across three different rebalancing cycles, all using the exact same mechanism (the same circuit breaker, the same Crisis Rule), varying only how often the checkpoint itself fires. The comparison below uses this page's own default inputs - Safety (B1) 3 years, Income (B2) 4 years, Capitulation, Stress Test off - run as 5,000 Monte Carlo trials per option, each option tested against identical randomized market sequences for a fair, apples-to-apples comparison.
| Rebalancing Cycle |
How Often Safety Gets Checked |
Monte Carlo Success Rate |
| Annually |
Every single year |
66.5% |
| Safety's own Years of Expense |
Every 3 years (this page's own default) |
71.3% |
| Safety + Income's combined Years of Expense |
Every 7 years |
72.7% |
The pattern is consistent: the less frequently the checkpoint fires, the closer Refill's own success rate moves toward a plain waterfall's - because a longer gap between checkpoints means Growth spends more time compounding completely undisturbed, the same advantage a plain waterfall has by design. Checking in annually is the most active form of Refill, and also the one furthest from matching a plain waterfall's own raw survival odds; checking in only once Safety and Income's combined runway is exhausted is the closest Refill can get to a waterfall's own behavior, while still keeping the periodic discipline that gives Refill its own value. This is a real, measurable trade-off, not a reason to pick one cycle over another by default - a shorter cycle checks in more often and catches a struggling Safety bucket sooner, at some cost to Growth's own compounding time.
Common bucket-strategy planning mistakes
The gaps this tool is specifically built to surface: assuming a rebalancing or refill mechanism automatically improves long-run survival odds, when the honest comparison on this very page often shows the opposite; sizing buckets by feel rather than checking the actual Monte Carlo Success Rate and worst-case outcome each configuration produces; treating "I have a bucket strategy" as a substitute for first confirming the underlying corpus is even large enough via the Required Retirement Corpus check; and picking a Crisis Rule or Stress Test setting without ever looking at the year-by-year audit table to see what it actually does to your own numbers in a genuinely bad stretch, not just the headline success rate.
How to use this tool
Start with Retirement Basics - your retirement age, life expectancy, monthly expense, and post-retirement inflation assumption. Enter your actual holdings across Equity, EPF, NPS, FD, and Misc, each with its own return and volatility. Check the Required Retirement Corpus card first, before anything else - neither bucket strategy removes a genuine shortfall, they only change how gracefully it plays out. From there, review the plain Waterfall's own bucket sizing and its Range of Simulated Outcomes, then move to Refill Strategy - adjust its B1/B2/B3 Years, Return, and Volatility (auto-synced with the Waterfall card by default, so both start from the identical setup), choose a Crisis Rule, and optionally turn on Stress Test or Monte Carlo mode. Use the live Refill vs. Plain Waterfall comparison panel to see both success rates side by side, and the year-by-year audit tables under each card to see exactly what's driving the number, not just the number itself.
Terms and Definitions used in this Refill Strategy Lab
Examples use this page’s default inputs: retirement at 50, life expectancy 90, ₹1,00,000 monthly expense at retirement, 6% post-retirement inflation, and ₹1.5 crore in equity (12%), ₹55 lakh in EPF (8.25%), ₹30 lakh in NPS (10%) and ₹15 lakh in fixed deposits (6%). Monte Carlo figures vary slightly with every set of simulated paths.
Retirement Age and Life Expectancy – together they set how long your corpus must last: life expectancy minus retirement age. Example: retiring at 50 with a life expectancy of 90 means a 40-year retirement.
Monthly Expense at Retirement – what you expect to spend each month in the first year of retirement, in rupees of that time; it then rises with post-retirement inflation every year. Example: ₹1,00,000 a month, or ₹12,00,000 a year.
Inflation – the yearly rise in prices. Here, post-retirement inflation raises your withdrawal every year so it keeps the same buying power. Example: 6%.
Multi-Asset Investments – what you hold in Equity (stocks and mutual funds), EPF, NPS, Fixed Deposits and Miscellaneous assets, each with its own return and volatility; together they make up your Available Corpus at retirement. Example: ₹1.5 crore + ₹55 lakh + ₹30 lakh + ₹15 lakh = ₹2.5 crore.
Expected Annual Return – the average yearly return you assume. Each asset class carries its own return, and the corpus as a whole earns their weighted average. Example: Equity 12%, EPF 8.25%, NPS 10%, Fixed Deposit 6%.
Return Volatility – how much returns swing from year to year around the average, used by the Monte Carlo simulation. Set it to 0 to see a steady-return world. Example: at a 10.57% return and 8.05% volatility, about two in three simulated years fall between +2.5% and +18.6%.
Required Retirement Corpus – the lump sum you need on the day you retire to fund your inflation-rising expenses until life expectancy, while the remaining corpus keeps earning its return. It uses the Fisher real rate of return and an annuity-due present value (withdrawals at the start of each month) – the same calculation as the Retirement Corpus Required (RCR) in the Advance Retirement Calculator. Example: ₹2,32,00,706 for the default plan.
Available Corpus (Total Available Corpus at Retirement) – the combined value of your multi-asset holdings at retirement, compared against the Required Retirement Corpus to show a surplus or deficit. Example: ₹2.5 crore available vs. ₹2.32 crore required – a surplus of ₹17,99,294.
Weighted Avg. Return and Weighted Avg. Volatility – each asset’s return and volatility weighted by its share of the corpus – the single return and volatility the corpus as a whole is modelled with. Example: 10.57% return and 8.05% volatility.
Asset Allocation – the doughnut chart showing how your corpus splits across asset classes, with an Asset view (Equity, Debt = EPF + FD, Hybrid = NPS, Misc) and a Sub-Asset view (each holding separately). Example: 60% Equity, 28% Debt, 12% Hybrid.
Sequence of Return Risk (SORR) – the risk that the order of returns, not just their average, decides whether a corpus lasts: a market fall early in retirement, while you are withdrawing, does far more lasting damage than the same fall late in retirement.
Monte Carlo Simulation (1,000 trials) – the plan is replayed 1,000 times, each with its own random sequence of yearly returns drawn from your expected return and volatility, to show the full range of possible outcomes.
Monte Carlo Success Rate – the share of the 1,000 trials in which the corpus lasts through your full retirement, up to life expectancy.
P10 / P50 / P90 (Pessimistic, Median, Optimistic) – throughout this lab, P10 always means the pessimistic outcome (only 10% of paths do worse), P50 the median and P90 the optimistic outcome (only 10% of paths do better). For a corpus, pessimistic means a lower value.
Terminal Wealth – the value left at the end of the simulated retirement horizon (life expectancy) – or zero, if the corpus ran out before reaching it.
Volatility Drag – the gap between a steady return and the same average return delivered unevenly; the reason a plan that works on steady returns can still fail in many simulated markets.
Bucket Strategy – splitting one retirement corpus into three pots with different jobs, time horizons and risk levels, and spending from them in a fixed order rather than from one undivided pool.
Safety (B1), Income (B2) and Growth (B3) – the three buckets. Safety holds the nearest years of spending in low-risk assets; Income holds the next few years in steadier, income-oriented assets; Growth holds the rest, invested for the long run and not touched until the other two are used.
Waterfall (No Rebalancing) – the withdrawal order used here: Safety (B1) is drained fully first, then Income (B2), then Growth (B3), with no money ever moved back between buckets.
Bucket Sizing by Years of Expense – each bucket is sized as a number of years of inflation-adjusted annual expense, with Growth (B3) taking whatever corpus remains. Example: B1 = 3 years (₹38.20 lakh), B2 = the next 4 years (₹62.52 lakh), B3 = the remaining ₹1.49 crore.
Years Each Bucket Supported Retirement – how long each bucket paid for retirement before running out on a steady, average-return path, and the age at which it was depleted. Example: with 3 + 4 years of expense, Safety lasts to about age 53.3 and Income to age 60.
Bucket-Aware Monte Carlo – the 1,000-trial simulation run bucket by bucket: Safety, Income and Growth each draw their own independent yearly return (no cross-bucket correlation is modelled), with the same withdrawal order as the waterfall.
Range of Simulated Outcomes – the chart tracing the optimistic (P90), median (P50) and pessimistic (P10) total corpus year by year through retirement, with a table of success rates for surviving 5 to 40 years.
Inference (P10 run-out age) – in a worse-than-typical market (the 10th percentile), the age at which the corpus could run out. Example: around age 77–78, about 27–28 years into retirement.
Refill Strategy (Calendar Reset) – instead of draining each bucket in turn, Safety (B1) runs down on its own for a full cycle of its own years of expense, then a refill checkpoint tops it back up from Income (B2), and Income from Growth (B3) – so the waterfall is rebalanced on a schedule rather than never. Example: with B1 = 3 years, a checkpoint fires every 3rd year.
Refill Checkpoint – the year Safety is topped back up to its full target. It also fires early if Safety runs dry before its scheduled year, and Growth’s own top-up of Income is deferred and rechecked each year if Growth’s return that year does not clear the B3 Refill Threshold.
Capitulation – one of the two Crisis Rules. When a refill checkpoint can't be fully funded by Income (B2) alone, Capitulation draws whatever's still needed directly from Growth (B3), regardless of Growth's own recent performance, so Safety's full spending target is always restored. Example: Growth is down 15% for the year, but Capitulation still draws ₹5L from it to fully refill Safety anyway - Safety's spending is treated as non-negotiable.
Corridor-Cut – the other Crisis Rule. Instead of forcing a full draw from Growth, Corridor-Cut permanently lowers Safety's own spending target - inflated forward one year, a haircut applied, then clamped so it can never fall below your Absolute Spending Floor. Example: a checkpoint can't be fully funded, so a ₹1,20,000/month target is permanently cut by 20% to ₹96,000/month (or the Spending Floor, whichever is higher) - and stays at that new, lower level going forward, not just for one year.
Crisis Rule – the toggle deciding which of the two rules above applies whenever a checkpoint can't be fully funded by Income alone. Example: switching this from Capitulation to Corridor-Cut changes what happens the next time B2 can't cover B1's full need, without changing anything else about the simulation.
B3 Refill Threshold – the minimum return Growth must post in a checkpoint year before it's allowed to top up Income at all - this is the circuit breaker that protects Growth from selling at a loss just to hit a schedule. Default: 0%, meaning any positive return clears it. Example: with the default 0% threshold, a checkpoint year where Growth returns -3% skips the Growth-to-Income refill entirely that year, even though Income needs it - the refill is deferred and rechecked the following year.
Corridor-Cut Haircut – the percentage by which Safety's spending target is permanently reduced the moment Corridor-Cut fires. Default: 20%. Example: an inflation-adjusted target of ₹1,20,000/month becomes ₹96,000/month after a 20% haircut is applied.
B1 Absolute Spending Floor – the minimum monthly amount Corridor-Cut can never cut below, no matter how large the haircut would otherwise be. Default: ₹70,000/month. Example: if repeated haircuts would otherwise push the effective spending target below ₹70,000/month, it's held at ₹70,000/month (inflation-adjusted) instead.
Stress Test – an optional toggle that forces Growth through a fixed, scripted market crash for a set number of years, instead of letting it draw its usual average or randomized return during that window - used specifically to see how each Crisis Rule behaves under a genuine downturn rather than an average one.
Crash Window – the specific years during which the Stress Test's fixed crash return applies to Growth. Default: Years 2 through 4 of retirement. Example: with the default window, Growth's own return is overridden by the Crash Return for Years 2, 3, and 4, then reverts to normal from Year 5 onward.
Crash Return – the fixed, negative return forced onto Growth for every year inside the Crash Window when Stress Test is switched on. Default: -15% per year. Example: across the default 3-year Crash Window, Growth loses roughly 39% cumulatively (1 - 0.85³), regardless of what its own average return or volatility would otherwise have produced.
Zone – the audit table's own label for what happened to the three buckets in a given year. Five possible values: Growing (no checkpoint this year - buckets simply grow and Safety pays its own way, exactly like a plain waterfall would); Reset (a checkpoint fired and both the Income-to-Safety and Growth-to-Income refills completed successfully); Reset (B2 Frozen) (Safety's own refill from Income succeeded, but the Growth-to-Income refill was skipped because Growth's return didn't clear the B3 Refill Threshold that year); Crisis (Income alone couldn't cover Safety's full need, so the Crisis Rule fired); Ruined (the portfolio ran out entirely and every later year is shown as zero). Example: a table showing "Growing" for Years 1-2, then "Reset" at Year 3, is a checkpoint cycle working exactly as intended, with no crisis at all.
B2→B1 Transfer and B3→B2 Transfer – the actual rupee amount moved from Income into Safety, and from Growth into Income, at a given checkpoint. Example: a "Reset" year showing a ₹4.2L B2→B1 Transfer and a ₹1.1L B3→B2 Transfer means Income gave ₹4.2L to Safety, and Growth then gave ₹1.1L back to Income to help refill it.
Active Bucket – used only in the plain Waterfall's own audit table, showing which bucket (or combination) actually funded that year's withdrawal under its drain-only rule. Example: "B1" means Safety alone covered the year; "B1+B2" means Safety ran out partway through the year and Income covered the remainder - the one year this handoff happens.
Chosen Condition – the tile showing which Crisis Rule and Stress Test setting are currently selected, so every result beside it is easy to interpret at a glance without having to scroll back up to check the toggles.
Years in Crisis – a summary tile counting how many years across the full simulation triggered the Crisis zone - a quick way to see how often the Crisis Rule actually had to step in, without reading every row of the audit table.
Refill vs. Plain Waterfall – the side-by-side panel comparing the 1,000-trial Monte Carlo Success Rate with Refill against the plain waterfall on the identical starting buckets and assumptions, so you can see whether scheduled refills help or hurt your particular setup. Example: around 70% with Refill vs. around 71% for the plain waterfall at the defaults.
Monte Carlo Simulation (On / Off) – the toggle on the Refill card: Off shows one deterministic average-return path; On shows one randomized trial in the chart and audit table, plus the 1,000-trial histogram.
Re-run Simulation – draws a completely fresh set of 1,000 market paths; results shift slightly each time, which is itself a reminder that the future is a range. On the Refill card it also redraws the single randomized trial shown in the chart and audit table when Monte Carlo mode is On (with it Off, the path is deterministic and re-running gives the same result).
Reset to Default – puts the card’s inputs back to the page’s default values.
