Moving Average Strategy for Swing Trading Tested: On H4 the 50/200 Cross Fired Once in 12 Weeks, and 0 of 19 Testable H1 and H4 Runs Passed

Swing trading with moving averages is usually sold with one argument: on H1 and H4 the spread is small next to the move, so a simple crossover finally has room to work. The advice that follows is always the same few settings. 50 and 200 for the big picture, 20 and 50 for the swing, 9 and 21 for the entry.

I already measured fast averages on the 5-minute chart in moving average strategy for scalping. This time I took the same price data up to H1 and H4 and asked two questions. Does the cheaper spread turn the crossover into a profit? And how many signals does a swing trader actually get?

What is a moving average swing strategy?

A swing trader holds for days, not minutes. The moving average version has three common forms:

• Always-in crossover: buy when the fast average crosses above the slow one, sell and reverse when it crosses back. You are always in the market, and the average decides the side.
• Crossover with a fixed exit: take the cross, but close at a stop or a target instead of waiting for the opposite cross.
• Pullback: wait until the averages are stacked (20 above 50 above 200), then buy when price dips to the 20 EMA and closes back above it.

All three are in every second swing trading guide. None of the guides I read said how many trades a quarter of H4 data gives you.

How was it tested?

• Data: EURUSD M5 from June 3 to August 26, 2026, 16,926 bars. Gold M5 from June 16 to August 26, 13,736 bars. H1 and H4 bars were built from the same M5 bars on the UTC clock: 1,417 and 345 for EURUSD, 1,146 and 292 for gold. M5 and M15 were run too, for contrast.
• Averages: EMA and SMA, pairs 9/21, 20/50 and 50/200, on the close. A rule starts trading only once its slowest average has a value.
• Always-in: entry at the close of the cross bar, exit and reverse at the close of the next opposite cross.
• Fixed exit: stop 1.5 ATR, target 3 ATR, 48 bars at most, one trade at a time. The same exit as in forex trading systems tested.
• Cost: 1 pip a trade on EURUSD, $0.30 an ounce on gold.
• Pass: profit after cost, profit in both halves (split July 15), above the 95th percentile of 300 random runs, and at least 20 trades. For the always-in rule the random runs kept the same holding periods and flipped a coin for the side.

How many trades does a swing trader actually get?

One golden cross in twelve weeks of EURUSD H4, and one in ten weeks of gold. Both happened to win, +153 pips and +$488 an ounce. One trade is an anecdote, not a test.

The warm-up is the hidden part. A 200-period average on H4 needs 200 bars, about 33 trading days, before it draws at all. Of the 345 EURUSD H4 bars only 144 could trade the 50/200 rule, and on gold 91 of 292. In MetaTrader you never notice this because the chart has years of history behind it. In a test it decides how much you can learn.

Across H1 and H4, 33 of 52 runs ended with fewer than 20 trades. Every H4 run except one was in that group. The one exception, SMA 9/21 always-in on EURUSD, had exactly 20 trades and lost 5.6 pips each.

Did any swing rule pass?

Pips on EURUSD, dollars an ounce on gold, trades in brackets. Lines with fewer than 20 trades were not judged.

EURUSD H1 lost on all 9 judged lines, from 0.50 to 5.99 pips a trade. Gold H1 made money on 6 of 9. Three of those were positive in both halves: SMA 9/21 both ways and SMA 20/50 with the fixed exit. None cleared the random benchmark. The closest was SMA 9/21 always-in at +$14.23 an ounce against a random 95th percentile of +$14.43.

That near miss needs context. Gold went from $4,351 to $4,675 in this sample. An average that is long most of the time looks good in a rising market, and so does a coin flip held for the same stretches, which is why the random runs scored so high.

The 50/200 lines are where the swing guides put their faith, and they had 5 to 8 trades on H1. Gold EMA 50/200 always-in made +$47.74 a trade, after -$70.50 in the first half and +$126.57 in the second. That is two different markets, not one result.

Did the cheaper spread help?

Yes, on paper. I measured the average move over 6 bars on each timeframe and divided the cost by it.

The swing argument is right about the arithmetic. On EURUSD H1 the spread is a quarter of the burden it is on M5. It still did not help, because 8 of the 9 judged EURUSD H1 lines lost money before the spread was charged. EMA 9/21 always-in, for example, lost 3.06 pips a trade gross. A cheaper toll does not fix a road that goes the wrong way.

For comparison, the fast lines on M5 and M15 were tested on the same terms, 52 runs, and two passed, both EMA 9/21 with the fixed exit: EURUSD M15 at +0.40 pips over 179 trades and gold M5 at +$1.10 over 371. The lower timeframes had enough trades to judge at all, which the swing ones did not.

Did the first-half winner keep winning?

If you pick a swing setting from the past, you are picking the first half and trading the second. I ranked the judged H1 lines on the first half and checked where the leader finished.

• EURUSD H1: EMA 20/50 with the fixed exit led the first half at +1.25 pips. In the second half it lost 6.62, last of 9.
• Gold H1: SMA 9/21 with the fixed exit led at +$11.76. In the second half it made +$1.39, 6th of 9.

The same pattern showed up in the RSI version of this test, RSI settings for swing trading: the higher timeframe gave fewer trades, and the best setting changed with the half.

So what is worth keeping?

• Count your crosses before you trust a setting. On H4 the 50/200 gave one trade a quarter. To judge it you need years of history, not a few months.
• On EURUSD H1 the crossover lost before cost. Moving up a timeframe cut the spread share from 28.3% to 7.6% and changed nothing about the sign.
• Gold H1 looked better, and so did random entries. 6 of 9 lines were positive while gold rose 7.4%, and none beat the coin with the same holding periods.
• The pullback to the 20 EMA lost on both markets on H1, -3.51 pips and -$7.78 a trade. On M5 it lost too, which matches the EMA pullback setup on M5.

Is a 50/200 on H1 really different from a long average on the 5-minute chart? I compared them hour by hour in swing vs scalping moving averages. How to set the swing averages up in MetaTrader, with the full grid, is in moving average swing settings for MT4. The EMA and SMA difference on its own is in EMA vs SMA.

Frequently asked questions

What is the best moving average for swing trading?

On this data none passed. EURUSD H1 lost on all 9 judged lines, and gold H1 made money on 6 of 9 without beating random entries held for the same periods.

Does the 50/200 golden cross work on H4?

It crossed once in 12 weeks of EURUSD and once in 10 weeks of gold. Both trades won, but one trade cannot tell you whether a rule works.

Is the spread less of a problem on H1 and H4?

Yes. On EURUSD it fell from 28.3% of a 6-bar move on M5 to 7.6% on H1 and 3.8% on H4. But 8 of 9 EURUSD H1 lines lost before cost.

Is the 20 EMA pullback better than a crossover for swing trading?

Not here. It lost 3.51 pips a trade on EURUSD H1 and $7.78 an ounce on gold H1.

How much data do I need to test a swing moving average?

Enough for at least 20 trades after the slow average has warmed up. For a 50/200 cross on H4 that means years, not a quarter.