Linda Raschke — Turtle Soup and the Holy Grail
Linda Raschke · 1980s–2010s (Street Smarts published 1995, featured in The New Market Wizards, retired from money management in 2015)
Published 2026.07.08
Linda Raschke is known as a trader who, instead of riding breakouts, buys back the moment a breakout "fails." The setups she published with Laurence Connors in Street Smarts (1995) are built on reaction, not prediction — a stop order is left to fill only once the market first proves the setup's direction, and the instant it fills, a structural stop is placed on the opposite side. This article breaks down the "structure" of that method for educational purposes; it is not individual entry instructions or a claim about future prices.
- Not breakout-chasing but reactive trading that "buys back failed breakouts" — the stop order fills only when the market proves it.
- Classify the regime first: swap in Holy Grail (EMA pullback) for strong, high-ADX trends and Turtle Soup (false-breakout reversal) for ranges.
- The stop is a hard stop sent to the exchange the moment you enter, sizing is volatility-weighted — and performance is judged by reward-to-risk (expected value), not win rate.
- Because it relies on lagging indicators and discretionary tape-reading, it is hard to replicate from rules alone — and Raschke herself noted that the more famous a pattern becomes, the more its edge erodes.
① Who Raschke Is, and Why She Classifies "the Regime First"
Linda Raschke is a short-term trader who managed institutional money as a registered CTA (Commodity Trading Advisor) before retiring in 2015, and she is widely known for Street Smarts (1995), co-authored with Laurence Connors, and for her interview in Jack Schwager's The New Market Wizards. The one sentence that runs through her method is "don't predict, react."
The key is that regime classification comes before the setup. The same breakout is more likely to "fail and reverse back" in a range, and more likely to "pull back, then resume" in a strong trend. So before memorizing any entry rule, you first use indicators to sort whether the moment is a trend regime or a range regime.
Read the direction and strength of momentum with the 3/10 oscillator (the difference of the 3- and 10-period simple moving averages plus a 16-period signal line), and gauge trend strength with the 14-period ADX. When ADX is above 30, treat it as a trend regime and scan for Holy Grail candidates; when it is low, treat it as a range or reversal regime and scan for Turtle Soup candidates. These indicators are lagging observational tools that forecast nothing about the future.
② Turtle Soup — A Reactive Entry That Buys Back the False Breakout
Turtle Soup is a mean-reversion setup that takes the opposite side of "Turtle-style breakout following." A new low (or new high) is broken, but you step in only when that breakout fails and reverses back. As the name suggests, it targets the zone where the participants who entered chasing the breakout become the "soup ingredients."
- Confirm the prior 20-bar low was formed at least about 4 trading days earlier (exclude lows that are too recent).
- Watch whether price briefly breaks below by taking out that 20-bar low — up to here it is only a "failed-breakout candidate," not an entry.
- Place a buy-stop order a few ticks above the prior 20-bar low so it fills only when price returns and reclaims that level.
- The moment it fills, send a structural stop below the breakout low to the exchange (no mental stops).
- Exit where momentum is exhausted (typically before the next day's close, at the range midpoint, etc.). Apply symmetrically to a new high on the upside.
Turtle Soup is not averaging down because price has fallen a lot; it is a reactive entry where the stop order fills only when the market proves the breakout structurally failed. If price never returns and breaks straight down, the setup is void, and stepping in there becomes the opposite risk.
③ Holy Grail — Re-Entering on a Strong Trend's First Pullback
Despite the name, the Holy Grail is no "holy grail" at all — it is a simple pullback setup that re-boards a strong trend's first pullback. Raschke and Connors named it "Holy Grail" ironically, precisely because of that simplicity.
- Confirm the 14-period ADX is above 30 — a candidate only when the trend is strong enough.
- Wait for the "first" pullback where price returns and presses into the 20 EMA (the second and third pullbacks are less reliable).
- Re-enter in the trend's direction when price breaks up through the high of the pullback bar that touched the 20 EMA (in a downtrend, a break down through the low).
- Place the stop below the prior swing low.
- ADX naturally easing during the pullback is normal, but if ADX collapses below 30 the trend premise is broken, so abandon the setup.
Raschke later noted that many of Street Smarts' setups had grown so well known that their edge had weakened, yet she is said to have regarded the Holy Grail as still relatively durable. Pullback re-entry, coupled with the principle of "not fighting the trend," gives it a comparatively solid reward-to-risk structure.
④ Risk and Sizing — Hard Stops and Volatility Weighting
The real engine of Raschke's method is not the entry rule but risk management. The instant an entry fills, the opposite-side structural stop is sent to the exchange, and if the setup breaks you exit automatically — cutting off mental stops and mid-trade hesitation at the source.
Risk per trade stays within roughly 1–2% of managed capital. Because volatility differs across markets, use volatility-weighted sizing that scales contracts and quantity to the "average daily dollar range" to standardize per-trade risk. When things turn against you, don't freeze — cut size first, down to a level where you can think clearly again.
Performance is measured by reward-to-risk and expected value, not win rate. By selecting only spots with good reward-to-risk, you aim for a structure where a few large gains offset many small losses. "Filling the boat" with a big bet is used sparingly, only in the exceptional regimes when volume and volatility explode, and size is otherwise kept restrained.
⑤ Invalidation Points and the Limits of This Method
Invalidation comes in three forms. (1) When the false breakout never returns but accepts the level and continues into a trend (Turtle Soup void); (2) when ADX collapses below its threshold and the trend premise disappears (Holy Grail void); (3) when the structural stop placed at entry fills. If any one of the three occurs, the thesis is immediately void.
These setups are built on lagging indicators (ADX, EMA, the oscillator), so they are vulnerable to confirmation bias, and the more famous they become the more competition wears them down (as Raschke herself noted). Above all, her real performance leaned heavily on discretionary tape-reading across more than 24 markets tracked by hand — a few lines of rules do not replicate that judgment. No dramatic "downfall case" is known, but that in no way means an individual's discretionary results reproduce as-is for the average trader.
Cut losses fast, and when things turn against you, cut size first to get back to a state where you can think clearly — the gist of the discipline Raschke repeatedly stressed (a paraphrase of the intent, not a direct quote).
⑥ When Porting to Crypto (Perpetual Futures)
The concepts can be ported, but adjustments are essential. Crypto is 24-hour, ultra-high-volatility, and funding-based, so its nature differs from the original index-futures and options markets.
- Redefine the session: with no open or close, setups that depend on "the first hour / the daily close," such as the 80-20 bar and Momentum Pinball, must have their session arbitrarily redefined using, e.g., UTC daily bars, or be excluded.
- Stop width: instead of a fixed tick stop, widen it with an ATR-based stop, and recalibrate the ADX and EMA periods to the coin's timeframe.
- Beware liquidation hunts: Turtle Soup's "false-breakout reversal" overlaps with the liquidation hunts of perpetual futures and reacts quickly, but at high leverage the wider retracements and slippage make it likely that a stop pinned to the boundary is swept first → lower the leverage.
- Add a context filter: layering funding rates and open interest (OI) as a regime filter improves the reliability of identifying a "pullback regime."
This article is an educational breakdown of the structure of a method Raschke has publicly documented and explained. It is not a buy or sell instruction for any specific coin, and it makes no claim about future prices or profits. High leverage in crypto can wipe out capital through forced liquidation, so no adjustment removes the risk.
Viewing Turtle Soup (a downside reversal) in an R frame (an observational example, not a buy call): say an asset briefly breaks below its prior 20-bar low of 60,000, sweeps down to 59,700, then returns above 60,000 and fills a stop order at 60,100. Placing the stop below the breakout low at 59,600 gives 1R = 60,100 − 59,600 = 500. If you set the reversal target at the range midpoint of 61,600, that is +1,500 = a 3R spot. With a 10,000 account and 1% risk per trade (=100), since 1R=500 the notional quantity back-solves to 100 ÷ 500 = 0.2 units. Nothing is said about win rate — with 3R reward-to-risk the break-even threshold for a positive expected value drops, so even being wrong often, you aim for a structure where a few +3R outcomes offset many −1R ones. If the reward-to-risk isn't there, the spot is excluded from consideration in the first place.
- Did I first classify whether this is a trend regime (ADX>30) or a range regime, or did I pick the setup first?
- Am I about to enter by "chasing" the breakout, or have I "waited" for the breakout to fail and reverse back and left a stop order?
- Did I place a structural stop on the exchange the moment I entered, or am I relying on a mental stop?
- Did I calculate this spot's reward-to-risk (target as a multiple of risk) before entering, and does it clear the minimum threshold?
- Am I not using session-dependent setups (the 80-20, Momentum Pinball) as-is in 24-hour crypto?
FAQ
Turtle Soup is "counter-trend trading" — isn't it dangerous?
It can be, so this is not a suggestion to copy it as-is. The point is not averaging down because price fell, but that the stop order fills "only when the market proves the breakout failed and reversed back," with a stop placed immediately below the breakout low. If price never returns and breaks straight through, the setup is void, and holding on there becomes the opposite risk.
Does "Holy Grail" mean it has a high win rate?
No. "Holy Grail" is just an ironic name pointing at its simplicity; performance is measured by reward-to-risk and expected value, not win rate. It aims for a structure that re-enters on a strong trend's first pullback, cuts losses short, and rides gains until the trend is exhausted — and no rule guarantees a profit.
Can I use it as-is on crypto perpetual futures?
Not as-is. Session-dependent setups (the 80-20, Momentum Pinball) lose their premise in a 24-hour market and must be redefined or excluded, and at high leverage the wider retracements and slippage put a stop pinned to the boundary at high risk of being swept first by a liquidation hunt. You need adjustments: widen the stop with an ATR-based stop, lower the leverage, and add funding and OI as context filters.
Why the advice not to chase breakouts?
Raschke's method is premised on "reaction, not prediction." Chasing a breakout that has already run far with a market order pushes the stop farther away, worsens the reward-to-risk, and leaves you repeatedly caught by false breakouts (whipsaws). So you react with stop orders only where risk is structurally defined — such as a breakout's "failure" or a trend's "pullback."