Advertisement IC Markets

Combining Trend, Pullback, and Momentum: A Layered Rule Set

Discuss 1-minute to 15-minute price action setups, fading intraday momentum, key support/resistance zones, and proven short-term trading methodologies.
Post Reply
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Combining Trend, Pullback, and Momentum: A Layered Rule Set

Post by Fairman »

Combining Trend, Pullback, and Momentum: A Layered Rule Set

Some of the most robust approaches layer simple filters instead of relying on one signal. A common structure combines trend, pullback, and momentum.
diagram.png
diagram.png (37.54 KiB) Viewed 215 times
Layer one is trend: define it objectively, for example, price above the 100-period average on the four-hour chart, with the average sloping upward. Trade only long in this condition, and only short in the opposite.

Layer two is the pullback: wait until price retraces to a zone, such as the area between the 20 and 50 averages, or a prior support level, on the one-hour chart.

Layer three is momentum confirmation: require an indicator like RSI to dip below 40 during the pullback and then turn upward, or require a bullish candle pattern at the zone.

Entry occurs when all three align, with a stop below the pullback low and a target set at the previous high or at two times risk.

The benefit is that each layer addresses a different failure mode. The trend filter avoids counter-trend trades, the pullback improves entry price and stop distance, and the momentum condition avoids catching falling knives.

The cost is fewer trades, and the risk of over-complication. Each added condition cuts the sample and raises the risk of overfitting.

Limit yourself to three or four conditions, and test each one's contribution separately.

Practical step: test the strategy removing one layer at a time, and see which layer adds the most value.
It’s Fairman :geek:
Recommended broker for automated trading & scalping IC Markets
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Re: Combining Trend, Pullback, and Momentum: A Layered Rule Set

Post by Fairman »

Removing one layer at a time is a great test. Often you find one layer adds nothing, and dropping it gives you more trades with the same quality. Run the full system first, then the same test without the momentum condition, then without the pullback zone, and compare expectancy and trade count each time. If dropping a layer barely changes expectancy but adds forty percent more trades, it was just filtering out good trades along with bad ones. If dropping it makes results much worse, keep it. It's an hour of work that tells you which rules are actually earning their place in the checklist.
It’s Fairman :geek:
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Re: Combining Trend, Pullback, and Momentum: A Layered Rule Set

Post by Fairman »

Three conditions is about my limit too. After that every extra filter seems to help the backtest and hurt the live results. The reason is that each extra filter removes trades, and with fewer trades the backtest results get noisier and easier to fit by accident. A system with three sensible conditions and two hundred test trades is easier to trust than one with six conditions and forty trades, even if the second looks prettier. It's also easier to follow live. When there are six things to check, some will be judgment calls in the moment, and that's where discipline starts to slip.
It’s Fairman :geek:
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Combining Trend, Pullback, and Momentum: A Layered Rule Set

Post by Shadow Trader »

Some layers overlap more than they look. If price pulls back from above the 20 average into the zone near the 50 on H1, RSI dipping under 40 on the way is close to guaranteed, so the momentum layer might mostly be repeating the pullback layer.

A quick count shows whether that is true. Say there are 120 zone touches in the test period and 105 of them also see RSI under 40, then the real filter is the turn back upward, and the dip itself adds almost nothing.

Sloping upward also needs a number before the removal test means anything. One option is the 100 average on H4 being higher than it was 30 bars earlier, which is about a week of trading.

With one layer removed, comparing total R over the same period sits well next to expectancy per trade. A layer that lowers expectancy slightly but doubles the trade count can still add more over a year, as long as costs stay small relative to the stop.

How is a setup handled when the H4 slope flips while the H1 pullback is still forming? Deciding that in advance keeps it from becoming one of the judgement calls mentioned above.
FXS Prop Desk
Posts: 1693
Joined: Wed Oct 07, 2026 9:52 am

Re: Combining Trend, Pullback, and Momentum: A Layered Rule Set

Post by FXS Prop Desk »

Each layer filters out trades, so the final number of signals may be small. Checking that there are enough trades in the backtest (at least 50–100) before drawing conclusions is important, since layered systems with few trades can easily look good by chance.

Testing each layer separately before combining them helps understand which layer adds the most value.

If the final rule set produces only a handful of trades a year on one pair, applying it to several uncorrelated pairs can increase the sample without loosening the rules.
FXS Official Team | The Idea Lab & Strategy
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Combining Trend, Pullback, and Momentum: A Layered Rule Set

Post by Shadow Trader »

Adding pairs to grow the sample, as suggested above, works best when the pairs really are different. EURUSD and GBPUSD often give the same trend and pullback signal on the same dollar move, so 60 trades on each can behave more like 90 independent ones than 120.

A quick check is to count how many signals on the two pairs land within the same hour. If half of them do, the second pair is adding about 30 fresh trades rather than 60.

The removal test also has more variants than it first seems. With three layers there are three single removals and three ways to remove two at once, so six variants next to the full system, and picking the best of seven runs is a small dose of curve fitting in itself.

Keeping the last year of data untouched until the layers are chosen limits that. If the chosen version holds up there with roughly the same expectancy, the layer decision was probably real.

Has anyone found a layer that only helps on one pair? That usually says more about how the pair behaves than about the rule.
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Combining Trend, Pullback, and Momentum: A Layered Rule Set

Post by Shadow Trader »

A simple pattern: a strong move, followed by three small counter-trend candles, followed by a resumption. The idea is to buy the dip in an uptrend after a shallow, orderly pullback.

Its appeal lies in simplicity. The pattern is easy to spot, the stop sits below the pullback low, and the reward-to-risk is usually favourable. It's a pullback entry in its most basic form.

The challenges are in the details. How strong must the first move be? How small must the pullback candles be? What if the pullback has two bars or five? A rigid three-bar rule filters out many valid pullbacks, while a loose rule catches a lot of failing ones.

Context drives results. In a clear trend with healthy momentum, the pattern often works. In a choppy range, it produces a string of small losses.

Combining it with a trend filter, such as price above a rising 50-period average on the five-minute chart, and a volatility threshold, improves the odds in many tests.

A common refinement ties the pullback depth to the first leg. If the pullback retraces more than 50% of the impulse, the pattern is considered failed, and no trade is taken. That single rule filters out a lot of the pullbacks that turn into reversals, at the cost of missing some deep-but-valid ones.

Has the three-bar pullback survived a proper test, and what filters made the difference?
Post Reply