Does a moving average rule beat simply holding the Nifty?

What happened when a moving average rule was tested on 255 months of NIFTY 50 closes?

No. Across 255 months of NIFTY 50 month-end closes from April 2005 to August 2026, 0 of 5 moving average windows beat simply holding, which returned 12.7% a year. But 5 of 5 cut the worst fall, which was -55.1% for holding. The rule trades return for a gentler ride.

Key Takeaways

• Just holding the NIFTY 50 gave 12.7% a year over the 255 months tested, with a worst fall of -55.1%.
• Of 5 moving average windows tested — 6, 8, 10, 12 and 14 months — 0 of 5 beat holding on return.
• All 5 windows had a smaller worst fall than holding, so the rule bought comfort, not money.
• The test assumes idle money earns nothing, and index levels carry no brokerage, stamp duty or capital gains tax.
• Every exit and re-entry in real life costs you charges and resets your holding period for tax.
• The honest use of such a rule is behaviour control for someone who panics — not a way to earn more.

What is a moving average rule, in plain English?

A moving average is just the average closing level of the index over the last few months. If you take the NIFTY 50's month-end close for the last twelve months and average them, that is a 12-month moving average. It moves because every new month pushes the oldest month out.

The rule tested here is as plain as it gets. At each month end, look at where the NIFTY 50 closed. If it closed above its own average of the last N month-end closes, you hold the index through the next month. If it closed below, you hold nothing. The average uses only months that have already finished, so there is no peeking at the future. Five versions were tested: 6-month, 8-month, 10-month, 12-month and 14-month.

People like this rule because it has no opinion. It does not care about the budget, about FII flows, about what a TV panel thinks of the rupee. It is arithmetic on closing prices. That is also why it is worth testing properly instead of repeating as folklore.

The moving average test on the NIFTY 50, April 2005 to August 2026

What was measured	Result
Months tested	255
Just holding the index	12.7% a year
Worst fall while just holding	-55.1%
Averages tested	6, 8, 10, 12, 14 months
Windows that beat holding on return	0 of 5
Windows with a smaller worst fall	5 of 5

How was this rule tested on the NIFTY 50?

The test runs from 1 April 2005 to 3 August 2026 — 255 months. That start date is not cherry-picked to flatter anyone. It is the first month at which a 14-month average exists, so every one of the five rules and the plain hold-it line cover exactly the same months. Comparing a rule that starts in 2005 against a hold-it line that starts in 2003 is the oldest way to fake a result, and it is avoided here.

This is a published rule tested on published index closes. It is not a strategy of ours and not the record of any service of ours; the same closes give anyone the same answer. Index levels carry no cost or tax. If you download NIFTY 50 month-end closes and run the same arithmetic, you will land where we landed.

Two assumptions matter. First, when the rule is out of the market, the money earns nothing — no interest is assumed on idle cash. Second, the NIFTY 50 price index carries no dividends, no brokerage and no tax on either side. Both assumptions cut in different directions, and we deal with each below rather than hiding them in a footnote.

Why did the rule lose on return but win on the worst fall?

Both halves of the result come from the same behaviour. The rule is slow. It only acts on month-end closes, so it never sells at the top and never buys at the bottom. In a long, deep fall — 2008, March 2020 — being slow is a gift. You are out for most of the damage, and the worst fall you ever live through is smaller. That is why every window tested came out ahead of holding on this measure.

In a choppy market, being slow is expensive. The index dips below its average in a routine 8% correction, the rule sells, the market turns two weeks later, and the rule buys back higher. That is a whipsaw. Indian equity has delivered plenty of them — long sideways stretches where the index crosses its own average again and again without going anywhere. A handful of whipsaws is enough to eat the money the rule saved in one crash.

There is a second leak. Big up-days cluster right after big down-days. The rule is usually sitting in cash on exactly those days, because it only re-enters at a month end once the index is back above the average. Missing a small number of very large up-months over 255 months is enough to explain the gap, and that is before you pay a single rupee of cost.

What do brokerage, STT and capital gains tax do to this rule?

Everything above happens on raw index levels. Real money does not trade at index levels. Every time the rule tells you to exit, you sell units of an index fund or an ETF and pay brokerage, exchange charges, stamp duty, GST and securities transaction tax on the way out. Every re-entry pays again, plus the bid-ask spread if you are using an ETF with thin volume on the NSE.

Tax is the bigger bite. In India, your holding period resets every time you sell. A rule that moves you in and out repeatedly keeps dragging your gains into the short-term bucket, which is taxed at a higher rate than long-term equity gains, and it burns through your annual long-term exemption unevenly. Holding does the opposite: it defers the whole tax bill to a single distant sale.

Here is an illustrative sense of scale — these are made-up numbers, not a measurement. Suppose an illustrative ₹10 lakh portfolio makes twenty round trips over two decades and each round trip costs 0.5% in charges and forces an extra tax event. That drag is a real, recurring cost taken out of a rule that already did not beat holding before costs. The direction is not in doubt even if the exact figure is.

Does parking money in a liquid fund change the answer?

It helps, and the test deliberately does not give the rule that help. When the rule is out of the market, this test assumes the money earns nothing. In practice you would park it in a liquid fund, an overnight fund or a sweep FD, and it would earn something. So the fair reading is that the real-world rule sits somewhere above the version tested here.

But you have to take away as much as you add. Add the interest on idle money, then subtract brokerage and tax on every switch, and subtract the dividends the index itself pays that a price index does not count. The interest you earn while sitting out is partly cancelled by the cost of getting out and back in, and by the dividends you miss while in cash.

The useful conclusion is not a precise adjusted number — we have not computed one and will not guess at one. It is that the rule's shortfall on return is not a rounding error that better cash management quietly fixes. Zero of five windows beat holding, and the gap is being blamed on the cash assumption more often than the cash assumption can bear.

Who should actually consider a rule like this?

The person this rule helps is not the person chasing a higher number. It is the person who has sold at the bottom before. If a -55.1% fall would make you liquidate your entire equity portfolio in a panic and stay out for three years, then a rule that caps your worst experience at something smaller is not a performance tool. It is a device that keeps you invested at all.

Judge it that way and the trade-off becomes honest. You are knowingly accepting a lower long-run return in exchange for a smaller maximum fall and a written instruction that tells you what to do on the worst day, so you do not have to decide while frightened. For some investors that is a good deal. For someone with a twenty-year horizon, a stable job, a monthly SIP and the temperament to ignore the screen, it is a bad deal — they are paying for insurance against a risk they do not actually carry.

Be careful about one thing. The rule only works as protection if you follow it mechanically, including the re-entries. Investors who exit on the signal and then refuse to re-enter because the news still looks bad end up with the cost of the rule and none of its benefit. That is the most common way this idea fails in real portfolios.

How do I check a claim like this for myself?

Four checks catch most bad backtests you will see on Indian finance social media. First, does the rule start on the same date as the hold-it comparison? If the rule needs fourteen months of history, the comparison must also start after those fourteen months. Second, is the signal computed on closed months only, or does it use the very month it is trading in? The second one is looking at the answer sheet.

Third, what does idle money earn in the test, and is that stated? Fourth, has the author shown every window they tested, or only the one that won? Showing 6, 8, 10, 12 and 14 months together is the point — if only the 10-month version were published, you would have no way of knowing that it was picked after the fact from a group where 0 of 5 beat holding.

You do not need a research terminal for this. NIFTY 50 month-end closes are public on the NSE website. A spreadsheet with a column of closes, a column of averages and an IF statement reproduces the whole thing in an afternoon. A claim that cannot survive that afternoon is not worth your savings.

Does this same logic apply to picking individual stocks?

The habit transfers even though the rule does not. A price rule on a single stock has no floor under it — an index cannot go to zero, a company can. What does transfer is the discipline of applying one stated test to the whole market and then living with however few names survive it.

Look at what a basic quality bar does. Of 1,518 listed Indian companies with full-year fundamentals in our universe as of August 2026, only 22.0% clear return on equity above 15%, return on capital employed above 15% and debt-to-equity below 1 at the same time. Add a simple valuation check of PE under 40 and you are left with 15.9%. For context, the median PE in that universe is 24.0. Most of the market fails a test that sounds undemanding when you say it out loud.

One honest caveat on that screen: a flat debt-to-equity rule is unfair to banks and NBFCs, because borrowing is their business model, not a warning sign. Screens are for narrowing a list, not for delivering a verdict. The verdict still needs you to read the annual report and understand how the company makes money.

Click Here – See BossInvestor's Data-Driven Stock Screens

---

Conclusion

On the NIFTY 50 over 255 months to August 2026, the moving average rule did not beat holding on return in any of the five windows tested, and it reduced the worst fall in all five. So it is not a way to earn more — it is a way to feel less, paid for in return, brokerage and tax. Judge it as behaviour management, test it yourself on public closes, and be suspicious of anyone who shows you only the window that won. Anything that amounts to an actual call on a named stock sits behind KYC in the app, where regulation says it belongs.

BossInvestor is a SEBI Registered Research Analyst (INH000024143). This article is for educational and informational purposes only. It explains a method and is not investment advice, nor a recommendation to buy, sell or hold any security. Market-wide figures are computed from our universe as of August 2026; past metrics do not guarantee future results. Always do your own research or consult a registered adviser before investing.

---

Frequently Asked Questions

Which moving average window worked best on the Nifty?

None of them beat holding the index. Five windows were tested — 6, 8, 10, 12 and 14 months of month-end closes — and 0 of 5 beat simply holding, which returned 12.7% a year over the 255 months from April 2005 to August 2026. All 5 did reduce the worst fall below holding's -55.1%. Naming a single best window would be picking a winner after the fact from a group that collectively lost on return.

Why not use the 200-day moving average instead of a monthly rule?

A daily 200-day rule trades far more often than a monthly rule, which makes the cost problem worse, not better. Every extra exit and re-entry on the NSE adds brokerage, stamp duty, securities transaction tax and spread, and each sale resets your holding period for capital gains tax. The monthly version tested here is the gentlest, cheapest form of the idea, and it still did not beat holding. A noisier daily version starts from a harder position.

Does this result mean timing the market never works in India?

No. It means this specific published rule, on this index, over these 255 months, did not beat holding on return. That is one clean test, not a law of nature. What it does show is that the burden of proof sits with anyone selling a timing method: they should publish the rule, the exact period, every window they tested and what idle money was assumed to earn. Most claims you will see online do not.

Should I stop my SIP when the Nifty falls below its moving average?

That is a different question from the one tested here, and this page does not answer it. The test covers a lump sum moving fully in and out on a monthly signal, not a monthly contribution. Do note the mechanical difference: an SIP running through a fall buys more units at lower prices, which is the opposite behaviour to a rule that sits in cash during falls. Decide with your adviser based on your horizon and cash flow.

Does the -55.1% worst fall include dividends?

No. The test runs on NIFTY 50 price index closes, which exclude dividends entirely. That understates the return of simply holding, because a real index fund investor also receives or reinvests dividend income. It also means the comparison is slightly generous to the moving average rule rather than harsh on it. Index levels carry no cost or tax either way, so brokerage, stamp duty, securities transaction tax and capital gains tax are outside this measurement for both lines.