In a backtest of seven US-listed exchange-traded funds (ETFs), no moving average pair came top on more than two of them. The pair that came top on one fund ranked 6.1 out of ten on the other six, on average.
So which moving average is best? On this evidence, none of them is.
What is the best moving average? There is no single best moving average. In a backtest on seven funds, the pair with the highest CAGR (compound annual growth rate) over annualised volatility came out as six different pairs.
Seven funds gave six different top pairs. No pair came top on more than two.
The pair that came top on one fund ranked 6.1 out of ten on the other six, on average. Only 8 of those 42 placements reached the top three.
On SPY (the S&P 500 ETF) the ten pairs ranged from 0.39 to 0.78 on that measure, against 0.58 for holding. On no fund did the lowest of the ten beat holding on it.
How were the moving average pairs tested?
A moving average is the mean of the last n daily closing prices, as What Is a Moving Average? explains. The test used lengths of 10, 20, 50, 100 and 200 days.
A crossover pair holds the fund while the fast average is above the slow one, otherwise cash. A pair is written fast/slow, so 50/200 pairs a 50-day average with a 200-day average. Five lengths give ten pairs, including 50/200, covered in What Is a Golden Cross?.
Five single-length rules hold the fund while its price closes above its own average. That makes 15 rules per fund.
Each signal is read at the close and earns the next day's move. Holding means buying the fund and keeping it, also called buy and hold. Cash earns nothing. Each change of position costs 5 basis points, or 0.05%.
The data are daily adjusted closes to 16 October 2026 for seven US-listed funds. Each fund's test starts on its first day with a 200-day average, where holding starts too, giving 19.6 to 32.9 years.
The worst drawdown is the largest fall from a peak to a later low, as What Is a Drawdown? covers. CAGR over annualised volatility divides the yearly growth rate by the size of daily swings, scaled to a year.
We do not name a winning pair for any fund. A pair that comes top is the pair that suited that fund's history. That is the finding here. We are not recommending it.
How far apart are the ten pairs on one fund?
The seven funds are SPY (the S&P 500 ETF), QQQ (the Nasdaq-100 ETF), DIA (the Dow Jones Industrial Average ETF), IWM (the Russell 2000 ETF), GLD (the gold ETF), SLV (the silver ETF) and TLT (the long-term US Treasury bond ETF).
Source: YX Insights
Chart 1 shows the range of CAGR over annualised volatility across the ten pairs on each fund, with holding marked.
Fund | Lowest of the ten pairs | Highest of the ten pairs | Holding |
|---|---|---|---|
SPY | 0.39 | 0.78 | 0.58 |
QQQ | 0.21 | 0.60 | 0.33 |
DIA | 0.33 | 0.44 | 0.48 |
IWM | 0.10 | 0.46 | 0.36 |
GLD | 0.38 | 0.70 | 0.58 |
SLV | -0.01 | 0.22 | 0.21 |
TLT | 0.09 | 0.36 | 0.20 |
Two things stand out. The first is that the spread within a single fund is wide. On SPY the ten pairs run from 0.39 to 0.78 while holding sits at 0.58.
On six of the seven funds that spread is wider than the gap between the lowest pair and holding. Which pair you pick matters more than whether you trade at all. DIA is the exception, where the spread of 0.12 is narrower than the 0.16 gap to holding.
On SLV the lowest of the ten scored -0.01, meaning it lost money over the period. On DIA even the highest of the ten, at 0.44, fell short of holding at 0.48.
The second is the asymmetry. On six of the seven funds the highest of the ten beat holding on this measure. On none of the seven did the lowest.
Averaged across the funds the lowest pair fell 0.18 below holding, while the highest rose only 0.12 above it. Which pair would turn out high or low could not be known in advance.
Does the top pair on one fund work on another?
The table shows how far apart the pairs sit within a fund. It does not show whether the top pair on one fund is any good on another.
For each fund we took the pair with the highest value there, then looked up its rank among the ten on each of the other six funds. That gives 42 placements.
The average rank was 6.1 out of ten. Only 8 of the 42 reached the top three elsewhere, while 24 landed in the bottom half. On two funds, a pair that had come top somewhere else ranked last of the ten.
A pair that wins on one fund is, on average, a below-median pair everywhere else. That is why this guide does not hand you a pair to use.
Did the pairs cut the worst drawdown?
Mostly, yes. Across the seven funds, 67 of the 70 pair-and-fund combinations had a shallower worst drawdown than holding. All three exceptions sit on one fund.
That fund is SLV, where holding had a worst drawdown of 76.3%. Three of the ten pairs left it deeper, the worst of them at 85.9%.
Source: YX Insights
Chart 2 shows one fund, SLV, with holding beside the three pairs that made its worst drawdown deeper.
Silver is the reminder that the drawdown benefit is not guaranteed. The test did not measure why those three went the other way.
Did any pair beat holding on return?
Return is a different measure from the table above, which divides return by volatility. On return alone the picture is harsher.
On five of the seven funds, not one of the ten pairs beat holding.
The exceptions are two funds, with three of the ten pairs each.
On the Nasdaq-100 fund the margin comes from avoiding the dot-com fall, a drawdown of 83% for holding. On the long-term Treasury fund it comes from the 2020 to 2023 bond rout. Each rests on sitting out one episode, a different one in each case.
Why does the best moving average change from fund to fund?
A trend rule earns money when a fund keeps moving one way for longer than the rule takes to notice. It loses when the fund reverses soon after. Fast averages act early and catch more false signals. Slow averages act late on fewer.
SPY shows the trade-off in its single lengths. Its 10-day rule changed position 1,464 times and earned 0.64% a year, against 8.39% for its 200-day rule with 211 changes.
What differs most by fund is the price of being wrong. A fund that has suffered an 83% drawdown gives a rule something large to avoid. A fund with a smaller worst loss gives the same rule less to save, while each exit still costs 0.05%.
The top pair also fails to persist within a fund. When we tested this across a large set of tickers in I Tested the Golden Cross Against a Coin-Flipping Monkey, the pair that had been best did not stay best.
Appendix: the limits of this test
Cash earns nothing. A rule that sits out part of the time would have earned something on that cash, so the test understates the rules. Crediting cash with Treasury bill returns, from 2007, moved the tests beating holding on return from 7 to 11 of 105, so the finding survives this choice.
The risk column differs from a Sharpe ratio. It divides a compounded growth rate by a volatility built from daily moves.
Seven funds are not seven independent tests. SPY, QQQ, DIA and IWM are four views of one US equity market, while gold and silver move together.
Costs are simple. The 5 basis points allow no wider spread in stress and no tax.
The funds are survivors. The seven are large funds picked in 2026. The results illustrate only.
Start dates matter. Each fund runs from its own first day with a 200-day average, so the funds cover different periods. Across seven start dates from 2006 to 2015, 5 to 12 of 105 tests beat holding on return, while 82 to 100 had a shallower worst drawdown. The drawdown result leans on the 2008 crash.
What does the evidence mean for building a trend model?
The pair that tops a backtest is the pair that suited that history. On the other six funds it ranked 6.1 of ten.
So one setting does not transfer between assets. A system built on this analysis has to choose per asset. The trend model in Multi-model Signals is built that way. The rules tested here are not the ones behind any signal we publish.
Learn more with YX Insights
This explainer is part of the YX Insights Academy. Each one takes a single idea and checks it against real data.
The same approach runs through everything else we publish:
Systematic Portfolio: ready-made portfolios for Macro & Megacaps and for Commodities. We publish the holdings and every change.
Multi-model Signals: a daily long-or-flat call on every name we cover. Each call shows how strongly our four models agree.
Research: company deep dives, macro commentary and essays on how we test.
Good places to start on the website:
Academy: more explainers like this one.
Start here: get our free weekly letter by email.
Macro & Megacaps Systematic Portfolio and Commodities: the two portfolios in detail.
Track record: live results since 3 August 2026, plus the ten-year backtest.
Upgrade: get the Systematic Portfolio, Multi-model Signals and research in full.
Common questions about the best moving average
What is the best moving average?
There is no single best moving average. In a backtest on seven funds, the pair with the highest CAGR (compound annual growth rate) over annualised volatility came out as six different pairs. No pair came top on more than two. What Is a Moving Average? explains the basics.
Is the 200-day moving average better than the 50-day?
On SPY the 200-day rule did better, earning 8.39% a year against 4.03% for the 50-day rule. Holding earned 10.84%, so neither beat it. Those are single lengths. Crossover pairs are covered above. The result is one fund's.
Does the same moving average work on different stocks and funds?
No. The pair that came top on one fund ranked 6.1 out of ten on the other six, on average. On five of the seven funds, no pair beat holding on return at all. The test used seven funds, not individual shares.
Do moving averages beat buy and hold?
Rarely on return. Over full histories only two of the seven funds had any pair beat holding, three of the ten in each case. Each of those margins rests on sitting out one crash. On drawdown the answer flips: 67 of the 70 pair-and-fund combinations cut the worst fall.
DISCLAIMER: This article is strictly educational. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Any opinions, analyses, or probabilities expressed in this note are those of the author as of the note's date of publication and are subject to change without notice.