The Pullback Strategy That Helped Martin Luk Win the U.S. Investing Championship

ELI5/TLDR

Martin Luk won the 2025 US Investing Championship with a 969% return. His trick is buying strong, fast-rising stocks during a brief dip, but only at a spot where the price is likely to bounce — so he can set a stop-loss just below his entry and risk almost nothing. He is wrong about 85% of the time, but the rare winners run so far that an average win is roughly seven times an average loss. The whole game is making the loss tiny so the math works.

The Full Story

The obsession with tight stops

Most traders think about where a stock might go. Martin starts at the other end: how little can I lose if I’m wrong. The reason is arithmetic. “Risk multiple” just means profit measured in units of the risk you took. If you risk 3% and make 25%, that’s about 8 multiples. Halve the risk to 1.5% and the same 25% gain becomes 16 multiples. Same trade, double the score, purely from a tighter stop.

Tightening a stop-loss from 3% to 1.5% on a trade that eventually gains 25% can add roughly eight extra risk multiples. On a 50% winner, that same reduction in stop width can add nearly 17 additional risk multiples.

This is what he calls a parabolic relationship: shaving the stop doesn’t help a little, it helps disproportionately. So his entire energy goes into finding entry points where the stop can sit unusually close.

Buy weakness in strength, not weakness in weakness

The setup is a momentum stock taking a breather. Martin only buys dips in stocks that are clearly in an uptrend — making higher highs and higher lows, with rising short-term moving averages underneath. He avoids the opposite: stocks drifting below falling averages and making lower lows. A dip in a strong stock is a discount. A dip in a weak stock is just the next leg down.

He hunts with two kinds of support. First, moving averages — the 9, 21, and 50 EMA (exponential moving averages, which weight recent prices more heavily). When a rising stock keeps bouncing off its 9 or 21 EMA, those lines become a floor. Second, the anchored VWAP, or AVWAP — the average price paid since a chosen date, weighted by volume. Anchor it to a major high or low and it often acts as a magnet: resistance on the way up, support on the way back down. He layers price-action clues on top — old highs and lows, and unfilled gaps that tend to attract price.

Finding the candidates

The scan is mechanical. Pull the best performers over 1, 3, and 6 months — stocks up more than 30% — and keep the top 20-30% by performance. Add filters to drop the junk: market cap above $100 million, cash volume above $1 million, so you avoid thinly traded penny names. Then eyeball each survivor for a long consolidation, ideally over eight weeks, that has a clean history of respecting AVWAP and EMA levels.

What the entry actually looks like

The pattern repeats across his trades. A stock consolidates, bottoms, starts making higher lows, reclaims its EMAs and AVWAP, then tightens into a narrow range. The trigger is an intraday flush — a quick dip that pierces a support level and snaps back the same day, often touching two supports stacked together (say the 9 EMA and the AVWAP). He buys as the stock turns positive on the day or closes back above the levels, with the stop just under the day’s low.

The examples are a parade of the same outcome. Tesla flushed to its 9 EMA, recovered intraday, and offered an entry with under 1% risk — it ran 24% before closing below the 9 EMA, a 24x trade. SanDisk gave a 4% risk entry and ran 126% (31x). RMLD: 1.5% risk, then up 45% (30x). IonQ, AMPX, SYRE — all the same shape: a tiny stop, a long run, exits taken on a close below the 9 EMA (faster) or the 21 EMA (slower, more room).

The uncomfortable truth

The pullback buys with such tight stop losses fail a lot more than they work. However, the system still has super positive expectancy because the wins are massive multiples of the risk.

That is the catch with tight stops. Squeezing the risk to 1% means normal market noise will knock you out constantly. Martin is wrong 85% of the time. He survives — thrives — because his average win is close to 7x his average loss. The strategy is unforgiving of anyone who can’t stomach a long string of small losses while waiting for the home run.

Key Takeaways

Risk multiple = profit divided by the risk taken. Halving stop width roughly doubles the multiple on the same percentage gain — a parabolic, not linear, payoff.
Setup direction: buy dips only in stocks in clear uptrends (higher highs/lows, rising 9/21 EMA). Never buy dips in stocks below declining EMAs making lower lows.
Scan: top performers over 1/3/6 months, up >30%; keep top 20-30% by performance; filter for market cap >$100M and cash volume >$1M.
Best consolidations run longer than 8 weeks and have a track record of respecting AVWAP and EMA levels.
Support tools: AVWAP anchored to a major swing high/low; the 9, 21, 50 EMA; plus prior highs/lows and unfilled gaps.
Confluence matters: strongest entries are where two supports stack (e.g. 9 EMA + AVWAP, or AVWAP + horizontal support) on the same intraday flush.
Entry trigger: an intraday flush that pierces support and reclaims it; buy as the stock turns positive on the day, or on a close above the support. Stop goes at the day’s low.
Exits: trail with a close below the 9 EMA (tighter, locks gains faster) or the 21 EMA (looser, captures bigger runs).
Expectancy: ~85% of these trades fail, but average win ≈ 7x average loss, giving strong positive expectancy. The edge is entirely in keeping losses tiny.
An AVWAP that acted as resistance on the way up tends to act as support on a later pullback — and vice versa.

Claude’s Take

This is a clean, honest distillation of a real winner’s method, and the central insight holds up: when your win rate is low, the only lever that matters is the ratio of average win to average loss, and the cheapest way to move that ratio is a tighter stop. The parabolic-payoff framing is correct and worth internalizing.

The catch is survivorship and hindsight. Every example here is a winner shown after the fact, with support levels drawn once we already know the bounce happened. The video is upfront that 85% of these trades fail, which is to its credit, but it can’t convey how brutal that feels in practice — entering on a 1% stop means you are paying a small toll over and over, and the discipline to keep doing it through a drought is the actual skill, not the chart-reading. The mechanics are teachable; the temperament is not. There’s also a soft sell at the end (their own weekly-chart version), which is fine but worth noting.

Score: 7. Specific, concrete, and intellectually honest about expectancy and failure rate. It loses points for the inevitable cherry-picked examples and for hand-waving the hardest part — execution and psychology under a constant trickle of losses.