In this issue · 5 sections
A lotus grows in a pond. It doubles in size every single day. On Day 30, it covers the pond completely.
Question for you: on which day was the pond half covered?
Day 29. You knew that one - it's the famous version of the riddle.
So here's the better question. On which day was the pond 3% covered?
Day 25.
Think about that for a minute. For five-sixths of the lotus's life, a man standing at the edge of that pond would have told you nothing was happening. Twenty-five days of a nearly empty pond. Then, in the last five days, the entire growth took place.
Nobody quits on Day 29. They quit on Day 14, when the pond is 0.0015% covered and they've been staring at it for a fortnight.
Our human intuition is linear, whereas wealth creation is exponential, and the gap between those two is where almost everyone loses their money.
§01The maths, in one line
Here's the shortcut: at 12% a year, money roughly doubles every 6 years — that's the Rule of 72 (72 ÷ 12 = 6). So a lump sum left alone for 30 years isn't "30 years of growth." It's five doublings. And here's the part nobody tells you: the fifth doubling is worth more than the first four put together.
That's the riddle in the abstract. Here's what it looks like in real money.
I ran a ₹10,000-a-month SIP against the actual Nifty 50, every month from August 1991 to August 2026. Thirty-five years. It finished at ₹4.72 crore.
Now the part that should stop you. Twenty-seven years in — 2018 — that portfolio was worth ₹2.19 crore. Less than half of what it would eventually be. By then you'd already made three-quarters of every payment you were ever going to make.
That last quarter of the payments overlapped with 53% of the final value arriving. And 31% of the entire final value arrived in just the last five years.
A SIP isn't quite the pond, because you keep feeding it. But the shape is the same: the money goes in evenly, and the value shows up at the end.
And stepping up your contributions doesn't rescue you from the wait. The 10% step-up version put in eight times more money and landed in exactly the same shape — 37% of its final value arrived in the last five years alone. Feeding the machine more doesn't make the quiet years shorter. It just makes the reveal bigger.
§02So why doesn't everyone just wait?
Because waiting is the hard part. Not the investing. Not the maths. The waiting.
Here's the data. Every year, AMFI publishes a breakdown of how long the money sitting in SIPs has actually been sitting there. Not how long it could stay. How long it has.
As of March 2026, only 34% of the money in regular-plan SIPs had been invested for more than five years. For direct plans — the ones investors manage themselves — it was 20%. Meanwhile, 29% of all direct-plan SIP money was less than a year old.
The pond doesn't quit. The gardener does.
Now look at the gap between those two columns, because it points at something most people get backwards.
Direct plans exist so you don't pay a distributor a commission. No middleman, no advisory fee, more of your return stays yours. On paper, the direct investor should end up richer.
But direct-plan money is consistently younger. Some of that is simply because direct plans are newer — they've only existed since 2013, and India's count of contributing SIP accounts grew by a net 1.6 crore last year alone, so a lot of that "less than a year old" money is genuinely new rather than about to be pulled out. I don't want to overclaim here. Worth one more caveat: this is a share of assets, not of investors. Money invested six years ago has grown since, which inflates the "5+ years" bucket on its own.
What the data does say plainly is this: the plan that costs you less is not, on the evidence, the plan people hold longer. The cheapest plan is worthless if you don't stay in it.
That's the part nobody tells you when they sell you on "direct is better." Direct is better — for someone who was staying anyway. For everyone else, it's a small, real saving that funds a much larger, invisible loss.
§03What quitting actually costs
So I ran two more versions of the same SIP.
One increased the monthly amount by 10% every year — the step-up. It finished at ₹14.24 crore.
The other panicked. Every time the Nifty fell 20% from its high, it sold everything and sat in cash, waiting to feel safe again, only buying back in when the index hit a new all-time high. It finished at ₹0.43 crore.
Two honest caveats, because I'd rather you trust the numbers than be impressed by them.
First, the panic investor didn't just sell — they also stopped contributing while they were out. They put in ₹21.2 lakh against the flat SIP's ₹42.1 lakh. So the gap between those lines isn't purely the cost of panicking; a chunk of it is simply investing less. The honest comparison: the flat SIP turned ₹42 lakh into ₹4.72 crore, an 11x multiple. The panicker turned ₹21 lakh into ₹43 lakh — barely 2x, across thirty-five years. After inflation, that's close to standing still.
Second, nobody actually follows a written rule to sell at −20%. They just quietly stop the SIP and never restart it. This model isn't a strategy. It's a model of an emotion.
One more thing on the step-up, since the headline number is misleading. Yes, ₹14.24 crore beats ₹4.72 crore — but it invested ₹3.28 crore to get there, a 4.3x multiple, versus the flat SIP's 11.2x. The flat SIP earned more per rupee because its rupees arrived earlier and got more doublings. The step-up won on absolute rupees because it fed the machine more.
Both of them destroyed the one that quit. That's the only comparison that really matters.
§04Why you're not going to be that person
Look at that chart one more time.
Notice what the two winning lines didn't require. No stock picking. No timing. No opinion on rates, elections, or the rupee. Neither of them did anything clever. They did exactly one thing the third line didn't:
They were still there.
That's the whole edge. Not information — every person reading this can look up a NAV. Not intelligence — the maths is a Rule of 72 you learned four paragraphs ago. The edge is staying.
Which is oddly good news. Because "stay" isn't a talent. It's a design problem. And design problems have solutions.
- Invert the order. Salary → invest → spend what's left. Not salary → spend → invest what's left. Same income, same person, completely different outcome — because the second version makes your future self bid against your present self, every single month, forever. Fix this with an auto-debit dated the day after your salary lands. Remove the decision, and there's nothing left to talk yourself out of.
- Automate the step-up too. A 10% annual increase, set once, tracks your raises without asking you to decide anything, ever again. Every decision point is a chance to quit. This one just removes the point.
- Reduce the number of times you look. The crash isn't what makes people sell. Watching the crash is. Check quarterly. Not daily. The market doesn't reward attention — it charges for it.
You are not going to out-think the market. You're going to out-wait it.
Twenty-five boring days. And then the pond.
If this was useful, forward it to the person in your life who stopped their SIP in March 2020. They're the one who needs it.
§05How I built this
Every chart in this issue runs on real data, and the code is public — check my work, break it, or build on it.
Market data: NSE Nifty 50 monthly closing levels, August 1991 – August 2026, pulled via jugaad_data (index_df). Portfolio simulations (flat SIP, 10% step-up SIP, and the panic-sell scenario) are built month-by-month from those closes in Python (pandas + matplotlib/plotly).
What's excluded, on purpose: this is a price index, not a total-returns index — dividends aren't reinvested — and no expense ratio or exit load is modeled. Those cut in opposite directions: leaving dividends out understates the result, leaving costs out overstates it. Dividends are much the larger of the two over thirty-five years, so on net every number in this issue is a conservative estimate of what a real SIP would have earned, not an inflated one.
The Nifty series before 1996 is a back-computed one, and no investable Nifty index fund existed in India until around 1999. Treat the early years as a maths exercise, not a trade you could have placed.
In the panic scenario, cash sits at 0% while out of the market. A real panicker would have earned something in a savings account, so that line is a floor, not a forecast.
SIP holding-period data: AMFI's Fiscal 2026 Annual Report, "Investors embrace long-term approach with SIPs" — the holding-period breakdown of SIP AUM as of March 2026.
This newsletter is for educational purposes and is not investment advice. Past returns are not indicative of future results.
Code: github.com/shanzxt/n1-Compounding-newsletter — data pull, portfolio simulation, and every chart script for this issue.
If you spot a bug in the maths, that's the whole point of shipping the code — tell me.