Where the Minervini Trend Template Works: A Backtest Across 14 Stock Universes

By MarketInOut · October 21, 2026

Mark Minervini's Trend Template keeps only stocks in a strong Stage 2 uptrend: price above rising moving averages, near its 52-week high and outperforming most other stocks. It's a popular filter, but does it actually beat the market, and where?

We ran the same Trend Template strategy through MarketInOut's backtester on 14 stock universes in seven countries, with up to 26 years of data each, and compared every result with a local index fund. The biggest factor wasn't the country. It was which stocks the strategy was allowed to buy.

The short version

Limited to the members of a large-cap index, the strategy beat the local index fund in 4 of 9 tests, by up to 3.9 points a year, and roughly matched it in 2 more.
Run on every liquid stock of an exchange, it beat the index fund only once in 5 tests. On all liquid US stocks it lost money over 26 years.
It did worst in small caps (Russell 2000, FTSE 250) and in a very small index (DAX, 40 stocks).
Where it worked, it often fell much less than the market: 35% vs 82% for QQQ, 19% vs 44% for the FTSE 100 fund.

Where the Trend Template beat the index fund

Annual return of the strategy minus the annual return of the local index fund over the same period, in percentage points:

Full results

Universe	Since	Strategy per year	Strategy max drawdown	Index fund	Fund per year	Fund max drawdown
Nasdaq-100 members	2000	12.3%	35%	Invesco QQQ	8.4%	82%
S&P/TSX Composite members	2000	11.1%	37%	iShares S&P/TSX 60 (XIU)	8.2%	50%
FTSE 100 members	2007	6.9%	19%	iShares Core FTSE 100 (ISF)	4.8%	44%
Hang Seng members	2008	1.6%	35%	Tracker Fund of Hong Kong (2800)	0.9%	56%
Russell 1000 members	2012	14.3%	42%	iShares Russell 1000 (IWB)	14.3%	33%
S&P 500 members	2000	7.9%	48%	SPDR S&P 500 (SPY)	8.2%	54%
FTSE 250 members	2007	3.5%	26%	iShares Core FTSE 100 (ISF), as a UK proxy	4.8%	44%
Russell 2000 members	2006	3.8%	66%	iShares Russell 2000 (IWM)	8.2%	57%
DAX members	2009	2.6%	21%	iShares Core DAX (EXS1)	8.8%	37%
All liquid TSX stocks	2000	9.6%	38%	iShares S&P/TSX 60 (XIU)	8.2%	50%
All liquid LSE stocks	2007	4.7%	41%	iShares Core FTSE 100 (ISF)	4.8%	44%
All liquid Tokyo stocks	2000	0.0%	72%	iShares MSCI Japan (EWJ, US$)	3.1%	58%
All liquid US stocks	2000	0.0%	76%	SPDR S&P 500 (SPY)	8.2%	54%
All liquid BSE stocks	2013	1.3%	30%	SBI S&P BSE Sensex ETF	11.2%	33%

Each test ends on October 19, 2026 and starts in the year shown: the first year with index membership history for that universe. Index funds are bought and held, with dividends reinvested. Results are in each market's own currency, except the Japan fund, which trades in US dollars.

Why the universe matters so much

Index membership filters out short-lived spikes

On a whole exchange, thousands of small companies can pass the Trend Template for a few weeks after a sharp rally, then collapse. The strategy buys the strongest of them first, so it fills up with exactly these stocks. Index members are larger, more established companies whose uptrends tend to last, so the same rules catch longer trends.

Small caps are the hardest place for it

The Russell 2000 version lost 4.3 points a year against its index fund, with a 66% drawdown. The FTSE 250 version trailed too. Small-cap uptrends break more often and more violently, and the strategy's exit, a close below the 200-day average, comes too late.

The index needs enough members

The DAX has only 40 members, so at any time only a few of them are likely to pass the Trend Template. The strategy then holds fewer than its 20 positions and keeps the rest in cash, which makes it hard to keep up with the index. A pool of a few hundred stocks suits it better.

The largest US stocks are hard to beat

With S&P 500 and Russell 1000 members, the strategy roughly matched the index fund. The biggest US companies have dominated the index for years, and buy-and-hold already captures their gains. The Nasdaq-100 looks like an exception: the strategy beat QQQ by almost 4 points a year. But the whole edge came from 2000–2002, when few technology stocks passed the template and the strategy sat out most of the crash. From 2003 on, it trailed QQQ slightly: 15.0% a year vs 15.9%.

What about selling everything at year end?

We also tested a simpler exit: hold every position until the end of the year, sell everything, and start again with the new year's leaders. Annual return and maximum drawdown:

Universe	Exit below 200-day	Sell all at year end	Index fund
Nasdaq-100 members	12.3% / 35%	10.5% / 58%	8.4% / 82%
S&P/TSX Composite members	11.1% / 37%	10.6% / 58%	8.2% / 50%
S&P 500 members	7.9% / 48%	10.9% / 63%	8.2% / 54%
Russell 1000 members	14.3% / 42%	19.3% / 42%	14.3% / 33%
Russell 2000 members	3.8% / 66%	7.2% / 67%	8.2% / 57%
FTSE 100 members	6.9% / 19%	7.0% / 40%	4.8% / 44%

Selling at year end raised returns for S&P 500, Russell 1000 and Russell 2000 members, most of all with the Russell 1000 (19.3% a year). But it lowered returns for the Nasdaq-100 and TSX, and made the worst drawdown much deeper in most markets: 58% instead of 35–37% for the Nasdaq-100 and TSX. Using both exits together (sell below the 200-day average or at year end, whichever comes first) was worse than the 200-day exit alone in most markets. We kept the 200-day exit as the main rule because it held up better across markets.

A warning about these results

We tested 14 universes and several exits, and some results will look good by chance. Big edges often come from one period: the Nasdaq-100 result depends on 2000–2002. Don't pick the best row and expect the same in the future. The pattern across many tests is more reliable than any single number: the strategy did better on large, established stocks and worse on small, speculative ones.

How we tested

Entry: the eight Trend Template rules, unchanged from our Trend Template screen. When more stocks qualified than there were open positions, the strategy bought those with the highest one-year relative strength first.
Exit: a close below the 200-day moving average. No stop-loss. In our TSX test, this patient exit did much better than a quick one.
Positions: up to 20 at a time, 5% of capital each, bought and sold at the closing price.
Costs: 0.1% commission per trade plus a small spread.
Data: prices adjusted for dividends and splits, delisted stocks included, and index membership as of each date, so the strategy only bought stocks that were in the index at the time.
What's not tested: Minervini's full method, which uses the template as a watchlist and buys only on setups such as a volatility contraction pattern (VCP), with tight risk control.

Run it yourself

Opening formula, using Nasdaq-100 members as the example:

index(nasdaq100) and price > sma(50) and sma(50) > sma(150) and sma(150) > sma(200) and sma(200) > sma(200)@21 and price >= 1.25 * w52low and price >= 0.75 * w52high and rs(252) >= 70 and sortby(rs(252), desc)

Closing formula:

price cb sma(200)

Paste both into the Strategy Backtester and change the universe, for example to index(tsx), index(ftse100) or index(sp500). To test the year-end exit, leave the closing formula empty and set “Close position at the end of” to Year. To see which stocks pass the template today, open the Trend Template screen for the US, Canada, UK or India.

Open the Strategy Backtester → See today's Trend Template stocks →

Backtests run on MarketInOut data on October 20–21, 2026. Past results don't guarantee future returns.