Backtesting a Simple Momentum Screen on ETFs

By Chanuka Geekiyanage Oct 19, 2026

A momentum screen ranks ETFs by recent price strength and holds the leaders, and a backtest tells you whether that rule would have beaten buy and hold after costs and drawdowns. Most retail backtests of this idea look better than they should, because of lookahead bias, survivorship bias, ignored trading costs, and a rule tuned until it fits the past. This article shows how to build a screen simple enough to audit, run it without fooling yourself, and decide whether the result deserves real money.

Panaprium est indépendant et pris en charge par les lecteurs. Si vous achetez quelque chose via notre lien, nous pouvons gagner une commission. Si vous le pouvez, veuillez nous soutenir sur une base mensuelle. La mise en place prend moins d'une minute et vous aurez un impact important chaque mois. Merci!

What a Simple Momentum Screen Actually Tests

Momentum is the tendency of recent winners to keep outperforming for a while. A screen turns that into rules: rank a fixed list of ETFs by trailing return, buy the top few, and repeat on a schedule.

You are testing two separate claims. The first is relative momentum, meaning the leaders beat the group. The second is absolute momentum, meaning that when an asset is trending down you are better off in cash.

Keep these claims separate in your results. A strategy can win on the first claim and lose on the second, and a blended number hides that.

The Rules I Would Start With

Start with one rule set and write it down before you look at any results. Changing parameters after seeing the output is how an honest test becomes a curve-fitted one.

Parameter	Starting choice	Why it matters
Universe	10 to 12 liquid ETFs, such as the 11 Select Sector SPDR funds	Big enough to rank, small enough to audit by hand
Lookback	Average of 6-month and 12-month returns	Mirrors the horizons used by commercial momentum indexes
Skip period	Ignore the most recent month	Reduces short-term reversal noise
Holdings	Top 3, equal weight	Limits concentration without diluting the signal
Rebalance	Monthly, trade at the next close	Avoids using a price you could not have traded
Trend filter	Move to a T-bill ETF if the ETF is below its 10-month moving average	Tests absolute momentum separately
Costs	Commission plus an assumed bid-ask spread per trade	Turnover is the main cost of this strategy

The 6-month and 12-month structure is not arbitrary. BlackRock’s MTUM product brief describes the fund’s index as using 6-month and 12-month risk-adjusted price momentum scores, and another listing summarizes the method as selecting and weighting stocks on price gains over 6 and 12 months plus low volatility over 3 years. Your screen is a cruder version of the same idea.

Image source: MTUM product page

Choosing the ETF Universe

The universe decides more than the signal does. A screen run on 11 sector ETFs behaves very differently from one run on thematic funds that launched three years ago.

Thematic and narrow ETFs often have short histories and were launched after a strong run. That creates a built-in upward bias in any backtest. If you want to test a screen on that part of the market, first read how thematic AI ETFs differ from broad tech ETFs, because overlapping holdings can make three “different” winners one concentrated bet.

The same warning applies to newer products. Spot bitcoin ETFs only began trading in January 2024, so a crypto-linked ETF in your universe gives you less than three years of one regime, which is too little to judge a rotation rule.

Three Ways to Implement the Idea

You do not have to run your own screen. Here is how the main implementations compare.

Approach	Cost	Strengths	Weaknesses	Best for
DIY sector rotation (monthly top 3)	Your own trading costs	Fully transparent, you control the rules	High turnover, taxes, requires discipline	Investors who want to understand the mechanics
Packaged momentum ETF such as MTUM	0.15% expense ratio per iShares	Rules-based, no trading on your side, 130 or so stocks	Index rebalances can swing sector exposure sharply	Investors who want exposure without running a model
Faber-style 10-month trend filter on broad ETFs	Low, a few trades per year	Low turnover, simple to track	Whipsaws in choppy markets, no relative ranking	Investors focused on drawdown control

MTUM’s holdings show the main risk of momentum funds. One data provider listed Financial Services at 23.03% and Technology at 16.86% as of June 25, 2025. A later snapshot showed Technology at 39.22% as of March 26, 2026. That is a large sector swing in under a year, and your own screen will show the same behavior.

How to Run the Backtest Without Fooling Yourself

Most of the work is removing bias, not writing the ranking code. Check each of these before you trust any number.

Lookahead bias: Compute the signal using data available at month end, then trade at the next close. Using the same close for signal and trade quietly inflates results.
Survivorship bias: If your data source only includes ETFs that exist today, closed and merged funds are missing. Use a universe fixed at the start date.
Total return data: Use adjusted prices that include dividends, or you will understate every fund that pays income.
Trading costs: Deduct costs every time a position changes. Monthly rotation can turn over most of the portfolio many times a year.
Parameter sensitivity: Test 3, 6, 9, and 12-month lookbacks and top 2, 3, and 4-holdings. If only one combination works, you found noise.
Out-of-sample period: Fix your rules using the first half of the data, then judge them only on the second half.

The trend filter has the longest track record of the ideas here. Mebane Faber’s 2007 paper tested a 10-month moving average as a signal to enter and exit asset holdings, and noted that the rule is mechanical and unemotional. The weakness is that the same symmetric signal produces false exits in sideways markets, so check how many whipsaw trades your test generated.

Reading the Results

Do not stop at annual return. Compare the screen with buy and hold on the metrics that decide whether you can actually stick with it.

Maximum drawdown: The worst peak-to-trough loss. A momentum screen that matches the market’s return with a shallower drawdown is a real improvement.
Turnover and costs: Net return after trading costs is the only return that counts.
Return versus a packaged alternative: If MTUM at 0.15% would have done as well, your DIY effort needs a better reason to exist.
Concentration: Check how often one sector or theme made up most of the portfolio. Overlapping holdings can hide this, and the guide to AI ETF fees and holdings shows how top holdings repeat across funds that look different on the surface.
Tax drag: Frequent rotation in a taxable account turns gains into short-term gains, which can erase a small edge.

Momentum also has a known failure mode. It tends to lag sharply when markets reverse, because last year’s leaders become this year’s laggards. A backtest that skips a major reversal period tells you little.

When This Screen Makes Sense

It makes sense when you hold the position in a tax-advantaged account, use a liquid universe, and have a defined rule for trading and for ignoring the signal. It also fits investors who mainly want to avoid deep drawdowns, since the trend filter targets that.

It does not make sense when the universe is a handful of young thematic funds, when you cannot tolerate months of underperformance against the index, or when trading costs and taxes eat most of the edge. In those cases, a single packaged fund or plain buy and hold is the better choice.

My Take

For most readers, the best use of a DIY momentum backtest is learning, not replacing the core portfolio. Build the sector rotation test, add the 10-month filter, and compare both against buy and hold and against MTUM. If your version does not clearly win on drawdown or net return, use the packaged fund or skip the strategy.

I would trust the trend filter on broad ETFs more than the relative ranking on narrow ones. It has fewer parameters, lower turnover, and a longer documented history, so it has less room to hide overfitting.

Before using any version with real money, I would check four things: that signals trade at the next close, that the universe includes funds that later closed, that results hold across several lookbacks, and that the strategy still looks acceptable after taxes. A good fit is a patient investor with a core index holding who uses the screen on a small satellite portion. A poor fit is someone who will abandon the rules after the first bad quarter.

Conclusion

A simple momentum screen on ETFs is worth backtesting, but only with fixed rules, a stable universe, next-day execution, and realistic costs. Judge it on drawdown, turnover, and net return against both buy and hold and a packaged momentum ETF, not on a headline return.

The main risks are overfitting, hidden concentration in a few sectors, and sharp lags when markets reverse. If your test survives those checks, the practical next step is a small allocation inside a tax-advantaged account, reviewed against your written rules each month.

FAQs

1. How many years of data do I need to backtest an ETF momentum screen?

Aim for at least 10 to 15 years that include more than one market decline and one sharp reversal. Shorter samples mostly reflect one regime and tell you very little about how the rule behaves when conditions change.

2. Should I skip the most recent month when ranking ETFs?

Many momentum studies skip the latest month to reduce short-term reversal noise. Test both versions on your own universe, because the effect varies and you should not assume it helps.

3. Is it better to rebalance monthly or quarterly?

Monthly reacts faster but raises turnover and costs, while quarterly cuts trading but responds more slowly to reversals. Run both with realistic costs and keep the one with the better net return and smaller drawdown.

4. Can I just buy MTUM instead of running my own screen?

Yes, and for most investors that is the simpler choice at a 0.15% expense ratio. A DIY screen is only worth the effort if you want transparency, a custom universe, or a specific trend filter.

5. Why do my backtest results look better than live results?

The usual causes are lookahead bias, survivorship bias, ignored costs, and a rule tuned to the past. Fix signal timing, fix the universe at the start date, deduct costs on every trade, and judge the rule only on data you did not use to build it.

References

iShares MSCI USA Momentum Factor ETF product page (BlackRock): https://blackrock.com/latamiberia/products/251614

iShares MSCI USA Momentum Factor ETF product brief (BlackRock): https://blackrock.com/us/individual/literature/product-brief/smart-beta-style-box-mtum-en-us.pdf

MTUM fund details and sector weights as of June 25, 2025 (Robinhood): https://robinhood.com/stocks/MTUM/

MTUM fund details and sector weights as of March 26, 2026 (Robinhood): https://robinhood.com/us/en/stocks/MTUM/

MTUM Quantpedia strategy summary, Asset Class Trend-Following: https://quantpedia.com/strategies/asset-class-trend-following

Revisiting “Quant Approach to Tactical Asset Allocation” (The Big Picture): https://ritholtz.com/2012/01/revisiting-quant-approach-to-tactical-asset-allocation

Cet article vous a-t-il été utile ? S'il vous plaît dites-nous ce que vous avez aimé ou n'avez pas aimé dans les commentaires ci-dessous.

About the Author: Chanuka Geekiyanage

Chanuka is a content writer, editor and contributor at Panaprium, specializing in long-form editorial articles. He conducts in-depth research and produces well-structured, carefully edited content with a strong focus on clarity, SEO best practices, and consistent formatting. His work spans lifestyle, fashion, wellness, and technical topics, and all articles are reviewed and refined to meet Panaprium’s editorial standards.

Contre Quoi Nous Luttons

Les groupes multinationaux surproduisent des produits bon marché dans les pays les plus pauvres.
Des usines de production où les conditions s’apparentent à celles d’ateliers clandestins et qui sous-payent les travailleurs.
Des conglomérats médiatiques faisant la promotion de produits non éthiques et non durables.
De mauvais acteurs encourageant la surconsommation par un comportement inconscient.
- - - -
Heureusement, nous avons nos supporters, dont vous.
Panaprium est financé par des lecteurs comme vous qui souhaitent nous rejoindre dans notre mission visant à rendre le monde entièrement respectueux de l'environnement.

Si vous le pouvez, veuillez nous soutenir sur une base mensuelle. Cela prend moins d'une minute et vous aurez un impact important chaque mois. Merci.