Stop Trading Patterns, Start Stacking Edges: The Multi-Edge Method for Momentum Stocks

A simple pattern breakout, traded every time it appears, may win around 20% of the time, or even less. A trader who refuses to act on any single signal reports a win rate of 40 to 50% by waiting for several independent edges to line up at one price.

Why a lone chart pattern has almost no edge

A breakout pattern traded every time it appears is not a strategy, it is a coin toss with a stop loss. Jay Law, the trader at the centre of the source, puts the win rate of a simple pattern breakout at around 20%, and possibly lower. Many retail traders assume the opposite: that buying a classic cup and handle breakout comes with a built-in 50% win rate. Treating every occurrence of a setup the same way is trading at random, and random trading only leads to random results.

A chart pattern describes a shape, not a cause.

Stanley Druckenmiller, who learned technical analysis in detail early in his career, has said that technical analysis is about 20% as effective today as it was when almost nobody used it. His explanation is the core mechanism: when everybody uses the same tool, you no longer have a unique thing to act against.

A study of Bollinger Bands, introduced in 1983, found that trading them was very profitable before their introduction, and that their predictive power then gradually decreased, largely disappearing in most international markets since 2001. A broader study of 60 published investment strategies found that their Sharpe ratio declines by about one half after publication. The Sharpe ratio is return per unit of volatility, the standard measure of how much reward a strategy delivers for the risk it carries. The authors trace such decay to two causes: arbitrage capital moving in once an idea spreads, and overfitting, where part of the original performance only appeared because many hypotheses were tested.

That leads to the objection the source raises against itself: if a method is public, how can anyone prove it still works, and what stops large players from turning it against retail traders? A second defence is to anchor the method on what cannot be faked: in an uptrend, price keeps making higher highs, and institutions cannot manipulate that sequence.

Against that backdrop, the trader puts his own win rate typically at around 40 to 50%. The figure is self-reported, with no sample size or audited record attached, so read it as a description of the approach rather than a benchmark to hit. What it illustrates is the logic of the whole method: running the same breakouts through several independent filters should strip out many of the trades that would have failed.

The multi-edge rule: stack independent reasons, not indicators

Figure 1. Each additional independent edge raises the quality bar for a trade while reducing its frequency — the core tradeoff of confluence trading.

The governing rule fits in one line: never rely on a single indicator, and wait for several different edges to come together and align at the same place. Multi-edge trading, also called confluence trading, means taking a trade only when several of those reasons point to the same price at the same time. The source calls its way of applying this the METS methodology; this article rebuilds it from the steps shown on the charts rather than from the letters of the name.

The word that matters is independent. Quantitative researchers make the same point when they combine signals: what matters is how many genuinely distinct signals you have and how correlated they are with each other.

Stacking conditions has a trap of its own. Research on strategies built from multiple signals shows that their backtested performance is biased, often severely, by overfitting and selection. In that work, combinations of signals that individually had no true power often looked highly significant in backtests. The same authors stress that combining signals can still improve a strategy, provided the marginal contribution of each signal is evaluated on its own.

Done carefully, combination does help. One paper that combined moving-average and volume indicators in thirty-two ways found evidence that combining forecasts could benefit trading, and that exits with a stop loss beat exits without one. A study of MACD, Bollinger Bands and RSI on the LQ45 index concluded that a single indicator never gives a really accurate buy or sell signal, although the differences it measured were not statistically significant. Druckenmiller describes his own approach as multi-disciplinary, saying he will never buy something that does not have a great chart and fundamentals.

The host asks the obvious question: is there a specific number of edges before a trade can be taken? The answer is a trade-off, not a threshold. The more edges you require, the lower your trading frequency, because aligning five separate conditions at once happens rarely. When the reasons are fully solid, two edges can be more than enough to take a trade.

The host also asks whether to manage positions with the moving averages or with market structure, and again the answer depends on context rather than a fixed rule.

The method is not tied to particular tools: any indicator that has worked for you, MACD or RSI included, can sit in a multi-edge system as long as every trade has several reasons behind it. The source even argues that long-term investors run the same logic with their own filters, naming Warren Buffett and Peter Lynch. Buffett's stated filters show what that looks like: a great business with a high return on capital for a long period, a competitive position he understands, and management that will treat shareholders right. Template to adapt: write each edge on its own line next to the data it uses (price, volume, index, sector, business), then strike any line whose data repeats a line above it.

Figure 2. The three-layer filter: a trade only qualifies when the base catalyst, the sector context, and the relative strength check all confirm simultaneously.

Layer one: no catalyst, no momentum, no trade

The method stacks its edges in three layers, each answering a different question. The base asks why the stock should move at all: a fundamental catalyst that produces a momentum move. The middle asks whether the move is bigger than one company: a sector or theme rising together. The top asks whether the trend is healthy right now: relative strength against the index and a structure of higher highs.

The base layer is about cause. A price move only develops into a sustained, long-term trend when a very strong fundamental drive powers it. That drive usually arrives through a catalyst such as an earnings beat, a product launch or raised guidance.

This is why the order of operations matters. The trader needs to see a very clear momentum move happen first, and only then looks for a multiple-edge area further down the line to enter. The entry is never the breakout candle itself: it is the first orderly pause after it.

Why would a move that has already happened keep going? Finance research has documented the answer for decades as post-earnings-announcement drift: after an earnings surprise, the price tends to drift in the direction of the surprise for an extended period instead of adjusting at once. The drift can last several weeks, even several months, after the announcement. Laboratory experiments find that market prices underadjust to changes in fundamental value, and that trading strategies can profitably exploit the drift.

Figure 3. The five-step entry sequence: the breakout signals demand, the quiet pullback confirms it, and the confluence zone provides the tightest possible stop.

That persistent demand is what a momentum trader rides, and it is why the method insists on seeing the momentum move before hunting for an entry.

Mark Minervini's SEPA method runs in the order trend, fundamentals, catalyst, entry points, exit points, and requires the catalyst to explain why growth is happening and why it can continue.

The entry zone: high-volume breakout, low-volume pullback

Once the momentum move is in, the method waits for price to pull back down to the breakout level. The trader calls the ideal version a high-volume breakout paired with a low-volume pullback.

Data on chart patterns adds a twist worth knowing. In a study of over 8,000 chart patterns from July 1991 to May 2011, upward breakouts on above-average volume climbed 37.6% on average before the trend changed, against 35.8% when breakout volume was below average. The same study found price was three times as likely to throw back or pull back after an above-average volume breakout, and that failures nearly tripled. A companion page reports throwbacks 74% of the time after upward breakouts on above-average volume.

To measure the breakout side, compare the breakout day's volume with the average of the five trading days before it; above average qualifies. For the pullback side, volume must be lighter on the pullback than on the impulse that came before it.

The first place a pullback should stop is the breakout level itself, where the former swing-high resistance flips to support. The second candidate is a short moving average, the 10-period or the 20-period, either one or both.

A moving average should never be treated as one precise line, but as a range with a flexible buffer. The trader plots both the 20 EMA and the 20 MA and reads the band between them like a horizontal support range. An EMA, or exponential moving average, gives more weight to the most recent bars than a simple moving average does, so it sits closer to price and reacts faster.

Traders who specialise in 20-period pullbacks add three checks: the average must slope in the trade direction, price must have made a higher high before the pullback, and pullback volume must be lighter than the impulse. A veteran technician makes a related point: the slope of a moving average matters more than where price sits relative to it. One intraday test on the 5-minute chart of QQQ found the 20 EMA beat the 20 SMA, with a win rate of 58.2% against 55.9% and a profit factor of 1.61 against 1.44.

A third candidate is a trend line drawn across prior swing highs, which in one of the trader's examples held the first pullback after a gap. The closer the entry sits to the overlap of these supports, the tighter the stop can be, and a tighter stop allows a bigger position for the same risk. The typical plan pairs a stop of around 3% to 5% with a target of over 25%. The whole sequence, from breakout to exit, fits in one line of five steps.

Relative strength and market structure: the trend's health check

The top layer checks two constants that strong stocks share. Relative strength, a concept the source credits to William O'Neil and Stan Weinstein, measures how a stock performs against the market. Market structure is the sequence of swing highs and lows: in any uptrend, price keeps making higher highs, and that is something institutions cannot manipulate.

The cleanest way to see relative strength is a price relative line: the stock's close divided by the index's close. On StockCharts, joining two symbols with a colon builds that ratio, so AAPL:$SPX shows Apple against the S&P 500. The level of the line does not matter, only its direction: rising means the first symbol is outperforming, falling means the second one is. The ratio rises when the stock advances more than the index, or declines less than it.

Example: the index drops 2% and your stock drops 0.8% on light volume; its price relative line ticks up on a red day, and that pullback deserves a closer look. Example: the index rises 1% while your stock stays flat; the line falls even though the stock did not move, a quiet sign that money is going elsewhere.

O'Neil lists the degree of relative strength among the key factors shared by past market winners. Minervini's Trend Template asks for a relative strength ranking of no less than 70, with 80 or 90 preferred. Weinstein's stage analysis, set out in his 1988 book, organises price into a cycle of bottoms, breakouts, uptrends, tops, breakdowns and downtrends, with the uptrend as Stage 2.

Higher highs and higher lows define an intact uptrend. Each new higher low marks the place where buyers stepped in earlier than the time before, and the latest one is the natural line in the sand for a trade. When price makes a lower low after a run of higher lows, the structure has changed and the reason for the trade has to be re-examined. Structure is read on the time frame you trade, which is why time frames get their own treatment below.

This layer explains why the trader describes himself as about 70% technical. He wants momentum, relative strength and a strong sector on the chart first, and only then checks whether the fundamentals make sense. The chart is the screen, the business is the confirmation, and the entry method then locates a point with high reward for low risk.

Context: sector confirmation, liquidity and the overshoot

When the sector rises as one, it shows a real fundamental driving force behind the move. The most important thing is not the chart alone but its context: what the market and the group were doing at that specific time. CAN SLIM makes the same demand, tying winners to strong industry group moves and separating the leader from the laggard.

In the space-stock example the source discusses, earnings were irrelevant because the companies were not yet profitable. Peers such as LUNR were already showing strong momentum and clear relative strength in September, which confirmed the theme. Outside analysts saw the same gap between price and profits: one report noted that Rocket Lab's earnings losses had deepened over three years while its stock jumped 141% in six months.

Context includes the stock's own liquidity, and the trader screens on dollar volume rather than share volume. Dollar volume is price multiplied by shares traded: a million shares of a cheap stock and a million shares of an expensive one are very different sums for a fund to move. He typically wants over 100 million dollars traded a day. Once daily trading volume crosses the one billion mark, he says, institutional players arrive in force.

Figure 4. The top-down time frame routine: each chart answers exactly one question, and the lower chart never overrules the higher one.

The financial market constantly overshoots, which is why every break has to be judged in context. In one of the trader's examples, price broke below the low of the entry day, yet the level remained a multiple-edge area and the trade stayed valid. One order-flow analysis argues that a fast move through a prior swing level reflects thin trading at that moment rather than a deliberate grab of clustered stop orders. When such a break finds no follow-through and reverses, it is a failed auction, often marked by volume drying up.

Let's call it the overshoot test, a three-question checklist this article builds from the method.

The order-flow view is more sobering: the stops that story is about cannot be seen, and no public order-book study settles the question either way.

Time frames: let the higher chart give the orders

The higher time frame structure constantly dictates what happens on the lower time frames. The trader calls himself a medium-term trader whose main charts are the daily and the weekly, dropping to the hourly only to refine an entry. That pace works out, by the host's count, to just over 200 trades a year.

A level can be invisible on the daily chart and still be a multi-edge area on the weekly, so you switch time frames to suit your style.

Multi-timeframe analysis formalises this as a top-down routine in which each chart answers one question: the higher one says which way, the middle one says where, the lower one says when. As a rule of thumb, each time frame should be about three to six times the one below it, and weekly, daily and 4-hour charts suit swing trades held for days or weeks.

The classic mistake is shopping for entries. If you flick down through lower time frames until something looks like an entry, you will find one on most days, whether or not the higher chart supports it. Fix your set of charts before you open them, and let the lower chart only time a decision the higher one has already made. Example: a stock pulls back to a level that means nothing on the daily chart, but the weekly shows it is the top of a year-long base now acting as support; that weekly edge counts, and the hourly chart only places the order.

Risk first: position size, R-multiples and profit factor

A setup tells you where to enter, the stop tells you where you are wrong, and the size decides what being wrong costs. Position size is an output, not a choice: you find the trade, put the stop where the idea fails, then compute the size that makes the loss equal to a fixed slice of the account. A wider stop does not mean a bigger loss; it means a smaller position.

Worth quoting

• The breakout tells you buyers have arrived; the quiet pullback tells you sellers have not.
• A chart pattern describes a shape; the edge lives in the conditions around it.
• Let the weekly chart decide and the hourly chart execute.

Key points

• A lone chart pattern carries little edge, and widely known signals decay as more traders use them.
• No momentum, no trade: wait for a high-volume breakout, then buy the low-volume pullback into flipped resistance or the 10/20 MA band.
• The higher time frame sets direction; the lower one only times the entry.
• Fix risk before entry, for instance 0.35% of the account with a 3% to 5% stop, so a 40 to 50% win rate can still pay.
• Hold only moat companies at zero cost by selling 58.8% of the position to recover capital; sell everything else outright.

Action plan

1. Pull up your recent breakout trades and tag each one with the edges it really had, then compare the results of single-edge and multi-edge trades.
2. Write your edge checklist with the independence test: one line per edge, each fed by a different data source (price, volume, index, sector, business).
3. Plot the 10 MA, the 20 MA and the 20 EMA on your daily and weekly charts and shade the band between the two 20-period averages as one zone.
4. Filter your universe to stocks trading over 100 million dollars a day before reviewing any chart.
5. Before each trade, compute the share count as account risk divided by the entry-to-stop distance, rounded down.
6. Decide now which holdings qualify as moat companies for a zero-cost scale-out, and plan to sell the rest outright at your exit signal.

Test yourself

Q. Why do an RSI signal, a MACD cross and a stochastic reading agreeing not count as three edges?
A. They are all computed from the same price data, so they are highly correlated; what counts is how many genuinely distinct signals you have.

Q. How do you read a price relative line against the S&P 500?
A. Ignore its level and watch its direction: a rising line means the stock is outperforming, and it also rises when the stock falls less than the index.

Q. A strategy targets a 2:1 reward-to-risk ratio. What win rate does it need just to break even?
A. 33.3%, from the break-even formula 1 / (1 + R), which is why a win rate below 50% can still be profitable.

Open questions

• How can a trader tell in real time whether a break below a support zone is an overshoot or a true breakdown, when the method leaves that call to context?
• With no fixed minimum number of edges, what makes reasons solid enough for two edges to suffice remains a judgment that cannot be tested directly.
• Whether a combination of public signals keeps its edge out of sample is unsettled, since backtests of multi-signal strategies are biased by overfitting and published strategies tend to decay.
• Whether scaling out helps or hurts is debated: Van Tharp warns that top-slicing produces large losses and small profits.
• How should position size change when the stop varies between about 1.6% and 5% under a 0.35% risk cap?

Go further

1. Hard Lessons: Stan Druckenmiller — Druckenmiller in his own words on why technical analysis lost most of its effectiveness once everyone used it.
2. Backtesting Strategies Based on Multiple Signals — The research case for why backtests of strategies combining several signals are biased by overfitting, and how to judge each signal on its own.
3. Volume Study — A large study of chart patterns showing what high breakout volume really changes: slightly bigger moves, but far more pullbacks and failures.
4. Price Relative/Relative Strength | ChartSchool | StockCharts.com — How to build and read a price relative line, with a worked table of a stock against the S&P 500.
5. Multi-timeframe analysis: trend, zone and entry | CMC Funded — A clear top-down routine for multi-timeframe analysis, with the trap of shopping lower charts for entries.
6. SEPA Explained | Mark Minervini's Specific Entry Point Analysis System | ChartMill.com — Minervini's SEPA laid out element by element, including the full Trend Template and its relative strength threshold.
7. Profit Factor Calculator — Formula, Win Rate & Expectancy — Profit factor, expectancy and the break-even win rate formula, with the warning about small samples.
8. Position Sizing Formula and Methods | TradeGreen — The position sizing formula with a worked example and the tight-stop trap that inflates position value.