How every number is worked out
Nothing on this site asks you to take our word for it. This page is the working. If a figure on the simulator or the portfolio check cannot be traced to a rule here, write to us and we will fix the page.
The three lines
Put in
Every rupee that left your bank, including instalments that were skipped and sat in savings. All three lines put in exactly the same money on exactly the same dates, so the only thing that differs between them is what was done with it.
Held, untouched
Your instalment buys the fund at that month's closing NAV, every month, nothing sold. Worth is units held times the latest NAV. This is what the fund delivered to someone who did nothing.
What you did
The same instalments with a behaviour on top. Sold in a fall sells everything at the deepest drawdown in your window, keeps the money in savings at 3.5% a year for twelve months while the SIP also pauses, then puts it all back. Paused stops the SIP for the months you choose and leaves that money in savings. Booked profits sells a quarter of the holding whenever it stands 30% above cost, at most once a year, and keeps the proceeds in savings. Missed some skips one instalment in six. Money that leaves the fund is always credited at savings interest, never lost, so the gap is the conservative figure.
The routine
Each month, the fifteen funds in the universe below are ranked on 12-1 momentum: the NAV one month ago divided by the NAV twelve months ago, ignoring the most recent month because one-month returns tend to reverse. That month's fresh money is split equally across the top three. Nothing is ever sold and no fund is ever switched out of; a fund that drops out of the top three simply stops receiving new money. The routine needs twelve months of history, so it starts no earlier than February 2014 whatever start you pick.
The market check
Added in September 2026. Before a month's money goes anywhere, the fifteen funds are checked as a group: an equal-weight index of all fifteen, from twelve months ago to last month, the same window the ranking uses. If the group rose, the money follows the routine. If it was flat or fell, that month's money goes into your own fund as a plain SIP, and the routine resumes the first month the year turns positive again. Nothing is sold either way, so the check never triggers tax or an exit load. The rule comes from published research (Cooper, Gutierrez and Hameed, "Market States and Momentum", 2004; Daniel and Moskowitz, "Momentum Crashes", 2016) showing momentum pays after rising markets and fails after falling ones. We fixed it before testing and did not tune it. Since 2014 it has sent money to a plain SIP in 20 of 152 months, across five episodes: early 2016, late 2018 to early 2019, the 2020 crash and rebound, mid 2022, and three months of 2025-26. The simulator also shows momentum without the check, so you can see what the check did in your own window.
Return, a year
XIRR: the single annual rate that discounts every instalment and the final worth back to zero, on month-end dates. It is the right measure for money that went in over time, and it is what your app reports too.
The gap
Held minus what you did, in rupees. A positive gap is what the behaviour cost. Over some windows a behaviour happens to make money, and the simulator says so; one decade of one fund is not a rule.
The universe
Fifteen diversified equity funds with long records and stable mandates, the same fifteen the published backtest ran over. A short list is easier to watch than a long one.
Naming a scheme here is not a recommendation to buy it. Which funds fresh money goes to each month is worked out by the rule above and, once we are registered, shared only with advisory clients after risk profiling and a signed agreement.
Flexi cap and value
Parag Parikh Flexi Cap, HDFC Flexi Cap, ICICI Prudential Value.
Mid cap
Axis Midcap, Kotak Mid Cap, HDFC Mid Cap, DSP Midcap, Nippon India Growth, Franklin Mid Cap.
Large and large-mid
Mirae Asset Large Cap, HDFC Large Cap, ICICI Prudential Large Cap, SBI Large Cap, SBI Large & Midcap, Franklin Large & Mid.
Excluded on purpose
One sector fund sits in the data so the sector-rotation test can be reproduced. It is never eligible for the routine, because sector rotation was tested and rejected.
The data
Source
Month-end NAVs of direct growth plans from AMFI's public feed, for 284 diversified and sector equity funds. Where a fund's direct plan started after January 2013 its history begins there.
Freshness
The files are regenerated after each month-end and the date they run to is printed at the foot of the simulator. Fund returns in the search list are as of the same date.
Where it runs
In your browser. The NAV files download to you and every calculation runs on your device; there is no server-side computation on this site. Without an account, what you type never leaves your device. With a free account, the funds and watchlist you choose to save are stored in your account, readable only by you.
Limits
Month-end prices, not the exact day your SIP debits. Behaviour presets are stand-ins for what people actually do, not a record of what you did. Dividends are not modelled because the series are growth plans. Fund mergers and category changes are handled by using the surviving scheme's history.
What we tested and killed
Forty-five configurations, September 2015 to August 2025, over the same fifteen funds, before settling on the boring one, and, in September 2026, nine rules on when to skip it and four wider pools of funds to pick from. We publish the losers, not just the winner.
Buying the dip
Holding cash in reserve and deploying it on falls beat a plain SIP in none of the fifteen funds. Markets rise in most years; the waiting itself is what costs you.
Sector rotation
Chasing the hottest sector fund backtested at 17.45% a year against 18.38% for doing nothing. Sector funds also show the widest gap between fund returns and investor returns.
Market-timing exits
Every exit rule tried, valuation triggers, moving averages, profit booking, cost about 4.2 percentage points a year against staying put. So there is no sell signal, and that is in writing.
When to skip the routine
Tested in September 2026 over every 3, 5 and 7-year SIP window since 2014, against a plain SIP in each of the fifteen funds and in all fifteen together. Four ways of deciding when to skip momentum made things worse: skipping after the routine's own bad year (its record tends to bounce back), skipping when the funds are bunched close together, sticking with any fund that did well over three years (three-year winners did not stay winners), and letting the data pick the best rule each year (it picked plain momentum almost every time). The market check above survived. Over five-year SIPs it beat a plain SIP in 61.5% of windows against 59.1% for momentum alone, lifted the average edge from 0.50 to 0.65 of a point a year, and cut the average of the worst tenth of outcomes from 5.7 to 4.2 points a year behind. Against all 211 funds with five years of history, the worst tenth went from 8.1 to 5.4 points behind. That is suggestive, not proof: it rests on five episodes, and a placebo test put the chance of a result this good arising by luck at about one in fourteen. It also does not fix the main reason a plain SIP sometimes wins, which is that your fund turned out to be one of the best of its time. Nobody knows that in advance.
Picking from every fund, not fifteen
Tested in September 2026: the same routine picking its top three from all 279 diversified funds in our list, from all 284 including sector funds, from the 279 with at least three years of history, and with at most one fund per category. Over five-year SIPs the wider pools finished ahead of the fifteen-fund routine in about six windows in ten, by a median of 0.4 to 1.3 points a year, but they lost in the windows that began in 2014-15 and in 2021, the ones that ran through a small-cap fall. They still fell behind a plain SIP in about three windows in ten. We kept fifteen because around 40% of the wider pools' money went into small-cap funds, against none today; their worst fall was 29 to 35%, against 28%; they bought 69 to 88 different funds over twelve years, against 15; and their edge is about the size of the survivorship bias in our data, which holds only funds alive today, after the 2017-18 SEBI recategorisation merged many that were not. Anyone with a free account can run the 279-fund pool in the simulator and judge for themselves.
What survived
Top-three diversified momentum for fresh money with no selling: 19.13% a year against 18.38% for a plain disciplined SIP. That 0.75 of a point is the bonus, not the promise. The product is closing a behaviour gap that runs to several points a year for most people.
All figures are hypothetical backtests on public NAV history, not the returns of any client. They include no fees, no taxes and no trading days lost. We will publish an out-of-sample review of the routine every year, whichever way it goes.
How the service will work
Once SEBI grants registration. Until then, none of this is on offer.
- Measure your gap, free. The simulator and the portfolio check, with no account and nobody calling you.
- Join on screen. Risk profile, suitability assessment and an advisory agreement signed with Aadhaar eSign. If the assessment finds we do not suit you, it says so and refuses your money.
- Accounts in your name. Execution on BSE StAR MF or MFU: direct plans, your own folio, your own bank mandate. Money moves only from your bank to the fund house. We are structurally unable to touch it.
- Run the routine. SIPs debit on the day salary lands. On the 1st of each month the review directs fresh money to that month's top three. You approve every order in your own app.
- A human within a day. Any question gets a human reply within 24 hours. On the days the market falls hard, we call yearly-plan clients before they call us. The annual gap report shows, in rupees, what the discipline was worth.
If any platform asks you to transfer a SIP amount into its account rather than the fund's, that is a red flag under SEBI's 2021 anti-pooling rules, including if we ever ask.