Strategies · Moving-average crossover

The golden cross on a SIP: does a moving-average crossover pay?

A moving-average crossover compares two averages of the same index: a fast one and a slow one. When the fast average crosses above the slow one, the trend has turned up and you invest. When it crosses below, you move to a liquid fund. The famous pair is the 50-day and the 200-day: a golden cross when the 50 rises through the 200, a death cross when it falls through. We use the 2-month and 10-month averages of daily closes, checked at each month-end.

−2.70%a year against a plain SIP on the Nifty 50, 1995–2026
+0.47%a year against plain investing on the S&P 500, 1985–2026
20%of 10-year Nifty 50 periods in which it came out ahead

How it differs from a single moving average

The moving-average rule reacts to the price itself crossing the average. A crossover waits for the fast average to cross, which smooths out one-month spikes. It trades less (21 exits on the Nifty 50 in 31 years, against 28 for the 10-month rule) and gets out later.

The result on three markets

2-month over 10-month, whole historyPlain SIPWith the ruleExits10-year periods it won
Nifty 50, 1995–202611.83%9.13%2120%
Midcap Momentum 50, 2005–202620.97%15.25%130%
S&P 500, 1985–20269.07%9.54%1846%

The S&P 500 is the one market where the golden cross came out ahead over the whole history, by under half a point a year, and even there it won fewer than half of the 10-year periods. On the Indian indices it lost: by 2.70 points a year on the Nifty 50 and 5.72 on Midcap Momentum 50, where it did not beat a plain SIP in a single 10-year period. It did cut the worst fall: from −55% to −36% on the Nifty 50, from −69% to −39% on Midcap Momentum 50 and from −53% to −25% on the S&P 500.

Line chart from 1995 to 2026 of a 10,000 rupee monthly SIP in the Nifty 50. The plain SIP ends at 3.26 crore rupees; with the crossover rule, 1.89 crore.
The crossover sat out part of every big Nifty fall, but also the rebound that followed each one.

Other pairs

Changing the averages did not rescue it on Indian data. A 3-month over 12-month crossover made 10.63% on the Nifty 50 and a 5-month over 10-month 10.29%, both below the plain SIP’s 11.83%. On the S&P 500 both trailed plain investing too, at 8.57% and 8.35%.

After tax

With Indian capital-gains tax on both sides, the 2-over-10 crossover made 7.74% against 11.26% on the Nifty 50 and 13.44% against 20.06% on Midcap Momentum 50. Under US tax on the S&P 500 its small lead turned into a loss: 8.11% against 8.56%.

What to take from this

A crossover is a slower, calmer version of the moving-average rule. It fits markets with long trends, which the S&P 500 had more of than the Nifty in these periods. On our Indian data it cost 2.7 points a year on the Nifty 50 and 5.7 on Midcap Momentum 50 for a smaller worst fall. The golden cross in the news is a signal about the trend so far, not a forecast.

The moving-average rules are part of Strategy Lab Premium. The drawdown rule and the strategy search are free.

Investing in the US? The DCA backtest is the same Lab in dollar-cost-averaging terms, on the S&P 500 in dollars.

Questions people ask

What is a golden cross?

A golden cross is when a short-term moving average, usually the 50-day, rises above a long-term one, usually the 200-day. The opposite, the short average falling below the long one, is called a death cross.

Does the golden cross strategy work on the Nifty 50?

Not in our backtest. Investing only while the 2-month average was above the 10-month made 9.13% a year from 1995 to 2026, against 11.83% for a plain SIP, and came out ahead in 20% of 10-year periods.

Does the golden cross work on the S&P 500?

Barely, before tax: 9.54% a year from 1985 to 2026 against 9.07% for plain investing, with a much smaller worst fall. After US capital-gains tax it trailed, 8.11% against 8.56%.

Is a crossover better than a single moving average?

It trades less and reacts later. On the S&P 500 the crossover beat the 10-month rule (9.54% against 8.65%); on the Nifty 50 it did worse (9.13% against 11.45%). Both trailed a plain SIP in India.


Data: official month-end closes from NSE Indices for the Nifty 50 (Nov 1995 to Aug 2026) and Nifty Midcap150 Momentum 50 (Apr 2005 to Aug 2026); S&P 500 month-end closes from Yahoo Finance (Jan 1985 to Aug 2026). All are price indices, so dividends are left out of both sides. A fixed monthly amount is invested at each month-end. Money moved out earns 6% a year in a liquid fund in India and 4% in cash for the S&P 500. Returns are money-weighted (XIRR) and before tax unless a line says otherwise; the after-tax runs use the Strategy Lab’s Indian rules (12.5% long-term and 20% short-term gains, ₹1.25 lakh exemption, liquid-fund interest at a 30% slab) and, for the S&P 500, US rules (15% long-term, 24% short-term). Rolling periods start in every month. Figures come from the same engine the Strategy Lab runs in your browser. The rule moves 100% of the holding when the 2-month simple moving average closes below the 10-month one and buys back when it closes above, with no dead band; the SIP keeps investing throughout. An N-month average is the average of every daily close in the last N months, from NSE Indices’ official daily closes; for the S&P 500, the average of the last N month-end closes. Past returns are not a promise of future returns. Nothing here is a recommendation to buy or sell anything. See the disclaimer.

← All strategies