Most trader write-ups paraphrase. Linda Bradford Raschke's numbers are still sitting on her own website, in the original magazine PDFs, printed in numbered boxes. This guide walks the Holy Grail setup exactly as those four rules are written, then her indicator checklist — the 3-10 oscillator, the 14-period ADX, Keltner Channels at 2.5 ATRs, the 2-period rate of change and three market-internals readings — linking the source for each number, and saying plainly which of it a daily-bar backtest can reproduce.
A note on integrity before we go further: InvestingPaths is an independent research site and is not affiliated with, endorsed by, or connected to Linda Raschke or LBRGroup. Everything below is our summary of material she publishes for free on her own site, attributed and linked to the source PDF. Trading involves substantial risk of loss, and nothing here is financial advice or a promise of any result. We do not reproduce anything from her book Street Smarts (copyright M. Gordon Publishing Group), and where a setup exists only there we say so and leave it out.
Where these numbers come from
Everything here comes from four PDFs Raschke hosts free at lindaraschke.net/articles:
| Source | What it contains |
|---|---|
| AIQ Opening Bell, August 1997 (Vomund) | Boxed numbered rule sets: Holy Grail Pattern (4 rules), Three-Day Unfilled Gap Reversals (5), Wolfe Wave Construction (6) |
| Active Trader, March 2004 (Etzkorn) | The "Indicator checklist" sidebar — seven indicators with exact settings — plus the Grail's intraday cascade |
| Stocks & Commodities 11:9, 1993 (Hartle) | Her ADX rules of thumb and the 20-period EMA, eleven years earlier |
| Active Trader, August 2000 | "The rituals of trading" — what she does when a Grail buy fails |
We link to her copies rather than rehosting them. If a figure below matters to you, open the PDF and read the box.
The Holy Grail setup: the four rules as published
The 1997 AIQ interview prints the pattern in a box headed "Holy Grail Pattern", noting "Buy rules (sells are reversed)". Here it is, with what each condition asks for.
| # | The rule as printed | What it tests |
|---|---|---|
| 1 | The 14-period ADX must initially be greater than 30 and rising | A real trend, not chop, is in force. ADX measures strength and says nothing about direction |
| 2 | Look for a retracement in price to the 20-period exponential moving average, usually accompanied by a turndown in the ADX | The correction inside that trend. The ADX turndown is that correction showing up in the indicator, not a separate condition |
| 3 | When price touches the 20-period moving average, put a buy stop above the high of the previous bar | The trigger. You buy the first bar that takes out the prior bar's high, not the touch — so a dip that keeps going never fills you |
| 4 | Once filled, enter a protective sell stop at the newly formed swing low, and trail stops as profits grow | Risk defined by structure, not a percentage |
One textual detail, because people quote around it. Rule 2 says exponential; rule 3 says only "the 20 period moving average"; the body of the same interview says "a 20 day moving average". The same interview answers it a few paragraphs on, in her own words: "the 20 period exponential moving average often acts as support for stocks." So the shorthand in rule 3 is shorthand, not a second indicator. The 1993 Stocks & Commodities interview agrees — "I personally use a 20-period exponential moving average on all my markets" — though she adds in the same breath that "there really is no right or wrong parameter for a moving average", which makes the EMA her habit rather than a constant of the setup. Read the rules as a 20-period EMA.
The logic she gives is a probability claim, not a chart-pattern one: when a security makes a new momentum high and pulls back, the odds are high it will retest that high, so any pullback from a new momentum high can be bought. She puts the odds a top is already in place at "probably only 5% to 10%" — an estimate offered in a 1997 interview without a study behind it. Treat it as her conviction, not a measured base rate.
The part almost every write-up drops: it also cascades intraday
The Holy Grail is usually presented as a daily-chart setup, and on the 1997 evidence that is fair: every example in the AIQ interview is a daily stock chart, and her own gloss there is that we want the stock "to fall back to a 20 day moving average" once ADX has crossed 30. What the usual presentation drops is the 2004 Active Trader interview, where she describes the same setup running as a multi-timeframe intraday sequence — a layer above the four rules rather than a correction to them.
She describes it as a ladder. A market clears a long range and prints new momentum highs; the first retracement to the 20-bar EMA shows up on the 15-minute chart, sometimes on something as fine as the five-minute. The same pullback then repeats a step slower each time — half-hourly, hourly, two-hourly — and by the time it reaches the last rung, in her words, "you should be looking for a top."
And the condition that keeps it honest across those timeframes: "each retracement has to be preceded by a 14-bar ADX reading above 30 on that time frame." The ADX test is not inherited from the daily chart. It has to be true on the chart you are trading.
Her December 2003 S&P example runs through exactly that: the ADX rose above 30 on the 30-minute chart, so she bought the first retracement to the 20-period EMA looking to exit on a push to new highs; another Grail buy set up on the hourly on 23 December, and another on the 120-minute on 2 January. Her stated best confirmation is the 3-10 oscillator making new momentum highs alongside it.
The consequence, stated precisely: the cascade is a layer a daily-bar test cannot reach. It needs several timeframes at once and we have one, so the sequencing is out of range. The four printed rules are not — they were printed against daily stock charts to begin with. What a daily test gives you is the shape without the ladder: trend in force, dip to the 20 EMA, buy the resumption, stop under the swing.
Her indicator checklist, with the exact settings
The March 2004 sidebar is headed "Indicator checklist" and lists seven indicators she uses most often. Worth knowing whose sentences these are, because the sidebar mixes two voices. Items 1, 5 and 7 carry her own words in quotation marks; the settings for the ADX, the Keltner Channels, the rate of change and the closing-breadth line are the interviewer Mark Etzkorn's summary of what she uses. All seven are published sourcing rather than a community reconstruction — but only three of them are her phrasing, and the ADX entry gives no threshold at all.
| # | Indicator | The setting as published |
|---|---|---|
| 1 | 3-10 oscillator | The difference between a 3-day simple moving average and a 10-day simple moving average, plus a second line that is a 16-period simple moving average of the 3-10 line |
| 2 | ADX | 14-period, used to gauge trend strength |
| 3 | Keltner Channels | Bands placed 2.5 average true ranges either side of a 20-period exponential moving average |
| 4 | Rate of change | Two-period ROC, on daily charts |
| 5 | TICK / TIKI | Highs and lows. A TIKI of +26 or −26 is, in her words, usually a good indication a basket or program was fired off. She built her own TICK on the top 600 stocks but finds the standard one best |
| 6 | Closing breadth | A 10-period simple moving average of NYSE advancers minus decliners |
| 7 | Put-call ratio | A five-period simple moving average. Her comment on it: "The put-call ratio has been the No. 1 sentiment indicator. It used to be the TRIN" — a reading of what the market watches, not explicitly a ranking of her own tools |
Three of these deserve a note.
The 3-10 is not proprietary, and she tells you how to draw it. In the same sidebar: on a chart she usually just alters a MACD's settings, changing the moving average type from exponential to simple and the lengths to 3, 10 and 16. That makes it reproducible in any package that lets you set MACD's lengths — but only if it also lets you switch the averages to simple. Most do not, and a MACD(3, 10, 16) left on exponential averages is a related line, not the 3-10 oscillator.
She dates her own use of it to 1981. That figure comes from the 2004 checklist and is hers — the 3-10 is "something I've been using since 1981". The 1993 Stocks & Commodities interview, eleven years earlier, already describes the same three-minus-10 oscillator with a 16-period moving average, but the age it gives is the tool's, not hers: "that one's been around for 20 or 30 years."
TICK, TIKI, breadth and put/call are market-internals readings, not stock indicators. They describe the whole US market at that moment; none can be computed from a single stock's bars.
Her ADX rules of thumb (1993)
Separate from the Holy Grail, the 1993 interview gives two ADX thresholds she calls rules of thumb: she avoids markets where the ADX falls below 20 and is declining (not enough movement between swings), and calls it "something of a golden rule" that she will not trade against a market where the ADX is above 32 and rising. Those are 1993 figures on futures markets, stated as personal filters — not part of the Holy Grail rule set, where 30 is the number.
What we are deliberately not reproducing
Raschke's best-known setups — the ones with names — are published in Street Smarts, copyright M. Gordon Publishing Group. We are not reproducing rules from that book in any form, including paraphrase. So the entry, exit and filter rules for Raschke and Connors' Turtle Soup and Turtle Soup Plus One, the Anti, Momentum Pinball, 80-20's and the rest of the book's pattern set are not here, and nothing in the tables above comes from it. (If you arrived looking for "Turtle Soup", the term now overwhelmingly refers to an unrelated liquidity-sweep concept from the Inner Circle Trader material — same name, different thing.)
The Holy Grail also appears in Street Smarts, and the August 2000 Active Trader interview refers to it that way. Everything in our table is transcribed from the 1997 AIQ box she hosts herself, which is why we can print it. That interview adds one thing about the setup rather than from the book: when a Grail buy fails to hold the moving average, she stops looking to buy pullbacks in that market at all and looks at the short side.
What our free tools can and cannot do with this
We built the rule vocabulary behind our backtester, so we can be exact rather than promotional about the fit.
| Holy Grail element | In our rule set | Verdict |
|---|---|---|
| 14-period ADX above 30 | ADX trend strength: ADX (14) over 30 | Exact. Length and level are both free numbers |
| …and rising | — | Not expressible. Our ADX rule's operator is over/under only, with no slope option; Moving average slope reads moving averages, not indicators |
| Retracement to the 20-period EMA | Distance from a moving average: Price is less than 1% above its 20 EMA, or Corrective pullback | Approximate. "Coiled on the average" is a band; "touches" is not a daily-bar statement. The 1% is ours, not hers |
| Buy stop above the high of the previous bar | Entry order kind stop, referenced to the signal bar's high | Exact in spirit. The order rests above that high and fills on the first bar trading through it |
| Protective stop at the newly formed swing low, trailed | Trailing stop kind low: the lowest low of the last N bars, ratcheted upward only | Close, not identical. An N-bar low is not specifically the swing low that just formed |
| The 15 → 30 → 60 → 120-minute cascade | — | Impossible here. Our engine reads daily bars only, so the sequence cannot be tested at all |
And the indicator checklist:
| Indicator | In our rule set |
|---|---|
| 3-10 oscillator | MACD at fast 3, slow 10, signal 16 — but our MACD is built on exponential averages, so this is her shape with different weights, not her line. The simple-MA version she specifies is not available |
| 14-period ADX | ADX trend strength, length 14. Exact |
| Keltner, 2.5 ATR / 20 EMA | Price vs Keltner channel, length 20, multiplier 2.5. Exact — our channel's middle is an EMA of the close and the bands are ± multiplier × ATR |
| 2-period ROC | Rate of change, length 2. Exact |
| TICK / TIKI ±26 | Not available — we hold no market-internals feed |
| NYSE advancers minus decliners | Not available |
| 5-period SMA of the put/call ratio | Not available |
Three of the seven indicators are exactly reproducible, one approximable, three not there at all. On the setup, three of four rules translate and the ADX-rising clause does not.
Two numbers in that first table are ours, not hers, and are knobs to tune rather than settings to copy: the 1% band standing in for "price touches the 20 EMA", and the N-bar low the trailing stop follows. She says "touches" and "the newly formed swing low" — chart judgements a daily-bar rule has to approximate with a number.
One more limit, and it is the one most likely to waste your time if nobody says it: none of this runs as a market-wide scan here. Our free universe screen answers fundamental questions — valuation, growth, quality, sector, liquidity — and a chart condition handed to it comes back as "not screenable yet" rather than quietly passing every company. Chart rules like ADX and the 20 EMA are evaluated bar by bar inside a backtest instead. So the Grail is something you can test on daily bars today, not something you can scan the market for on this site.
An honest look at what this is
- It is discretionary in origin. Raschke says outright that she trades with discretion and uses systems as triggers, and that much of the time she does not know what her oscillators are telling her. The four printed rules are a skeleton run with judgment on top.
- It is a continuation trade, so it needs a trend to continue. Rule 1 is not decoration. Without ADX above 30 there is no momentum high for price to retest, and the rest is just buying a dip.
- Nothing here was tested by us. We quote no hit rate for the Holy Grail. Doing that honestly needs a defined daily-bar approximation and a survivorship-free universe — at which point you are measuring our approximation, not her setup.
- The market she described is not this one. The Grail rules are from 1997, the checklist from 2004, the ADX rules of thumb from 1993, on S&P futures and large-cap US stocks of that era.
Frequently asked questions
What is Linda Raschke's Holy Grail setup? A pullback continuation trade. The 14-period ADX must be above 30 and rising, price retraces to the 20-period EMA, a buy stop goes above the previous bar's high, and once filled a protective stop sits at the newly formed swing low and is trailed. Those four rules are printed in her hosted August 1997 AIQ Opening Bell interview.
What are the 3-10 oscillator settings? A 3-day simple moving average minus a 10-day simple moving average, with a 16-period simple moving average of that line as the second plot. She draws it by setting a MACD to 3, 10 and 16 and switching its average type from exponential to simple.
What timeframe is the Holy Grail traded on? In her 2004 description it is intraday and sequential: the retracement to the 20-bar EMA appears first on the 15-minute (or even five-minute) chart, then the 30-, 60- and 120-minute — each one preceded by a 14-bar ADX reading above 30 on that timeframe.
What Keltner Channel settings does she use? Bands 2.5 average true ranges either side of a 20-period exponential moving average, per her March 2004 indicator checklist.
Is the Holy Grail the same as Turtle Soup? No. Turtle Soup is a separate setup published in Street Smarts, which we do not reproduce here, and the term has since been taken over by an unrelated Inner Circle Trader concept. The Holy Grail is the ADX-plus-20-EMA pullback above.
Where to go from here
For the idea underneath the setup, see our guides to pullback trading, the impulse-pullback pattern that Raschke's momentum-high-then-retest framing belongs to, and moving averages in technical analysis. The swing trading playbook puts a pullback entry in the context of a whole trade.
If you would rather test a Grail-shaped rule set than read about one, our backtester is free and needs no signup — that link opens with a 0.1% fee and 0.05% slippage set, since the tool defaults both to zero. Build ADX (14) over 30 plus Price is less than 1% above its 20 EMA, set the entry order to a stop above the signal bar's high, and trail the stop on the 20-bar low. Hold in mind what that test is: three of four rules, two of them with a number we chose, on one timeframe, of a method built on four. Other traders' published criteria get the same treatment in our write-ups of the Minervini Trend Template and the Qullamaggie scan settings.
The primary sources are one click away, hosted by Raschke herself. Read the boxes.