The Savoy Cocktail Book

Why Everyone Still Recommends This Book (And What It Actually Gets Wrong)

The Savoy Cocktail Book came out in 1930. Harry Craddock put it together when he was head bartender at the Savoy in London. It has roughly 750 recipes organized into categories like Long Drinks, Short Drinks, Punsch, Champagne Cocktails, and so on. People treat it like scripture. It is not scripture. It is useful, but it is also a product of its time and it has some genuinely annoying gaps if you try to use it seriously. The original text is public domain. You can find free digital copies on Project Gutenberg and the Internet Archive. The 1930 first edition scan on archive.org is the cleanest version I've seen, though the page numbering runs oddly because of how the typesetter handled the recipe illustrations. There are also modern reprints from various publishers that add commentary, but you do not need those. The raw text is what matters. If you want the physical thing, any bookstore will have a reprint. They vary in paper quality. The Dover edition from 1988 is decent and cheap. Do not pay more than twenty dollars for it unless it has annotations you actually find useful.

How to Actually Use This Book Without Frustrating Yourself

The recipes are written in a style that assumes you already know what you are doing. A typical entry reads something like: "One dash Curacao, one dash Maraschino, two dashes Port, two dashes Old Tom Gin, juice of half a lemon, two dashes Angostura bitters. Shake with ice and strain into a cocktail glass." That is it. No measurements in ounces or milliliters. No guidance on which type of Curacao to use beyond the dash framework. No note about whether to use a coupe or a nick and nora glass. The dash system is the first thing that trips people up. A dash is not a standardized unit. In practice, one dash from a standard bitters bottle is somewhere between one-sixteenth and one-eighth of a fluid ounce depending on your pour. Modern bartenders usually convert a dash to about 0.06 fluid ounces or roughly 1.8 milliliters. The book itself never says this because the assumption was that every reader worked behind a bar and already knew. If you are making these at home, you will want a proper jigger and you will want to do the conversion yourself rather than guessing. Another thing nobody warns you about: the gin specified throughout is Old Tom gin. This is a slightly sweeter style than modern London dry. If you substitute a dry gin directly without adjusting the sugar content, several recipes will taste thin and harsh. The Bijou, for instance, falls apart with Tanqueray. Use a proper Old Tom like Ransom or Monkey 47 Schwarzwald Gold or even a vintage-style Plymouth. This is not optional.

Recipes That Work and Recipes That Need Fixes

Some formulas hold up remarkably well. The Savoy Cocktail (the drink itself, not the book) is essentially a perfectly balanced sour with Curacao replacing simple syrup. The quantity is roughly one-and-a-half ounces gin, three-quarters ounce sweet vermouth, three-quarters ounce Curaçao, and a dash of bitters. Shake hard. Strain into a chilled coupe. It is elegant and not overly sweet. I made this for a group of three people who had never had a pre-Prohibition cocktail and every single one of them asked for the recipe. That does not happen often. The Golden Dream is another one that works straight from the page. One part gin, one part dry vermouth, one part Curacao, one part egg yolk. The egg yolk was not listed in the original but many practitioners add it for texture. The book itself lists it as a straight spirit-forward drink and it is drinkable that way, but the mouthfeel is thin. I started adding the yolk around 2019 and have not gone back. Now for the ones that cause problems. The sidecar formula in the book calls for two parts brandy, one part Curacao, and one part lemon juice. That is a very strong drink by modern standards and the Curacao in 1930 was a different product than what you buy today. Modern triple sec is thinner and less orange-forward. I had a customer once who ordered a sidecar from my menu and sent it back saying it tasted like cough syrup. I tracked it down to the fact that I was using a cheap orange liqueur and the recipe ratio was pushing the alcohol to around forty percent ABV in the glass. I reformulated it to use Cointreau at a two-to-one-to-one ratio with a quarter-ounce simple syrup adjustment and it stopped being an issue. The original book recipe is technically correct for its era but it is not friendly to someone using contemporary ingredients without modification.

The Organizational Problem

The book is organized by drink type, not by base spirit. This seems minor but it is a genuine pain point if you are trying to build a bar program around it. There is no index by spirit. If you want every recipe that uses rye whiskey, you have to flip through the entire book and cross-reference manually. I spent about forty-five minutes once just trying to catalog which recipes were rye-based for a restaurant menu I was developing. A spirit index would have taken thirty seconds. There are modern companion books that add this kind of reference material. The Death & Co. Modern Cocktail book has good appendices. I keep both the Savoy and a modern reference on the bar because they serve different purposes. The Savoy is good for historical authenticity and classic technique. The modern books are good for when you actually need to find something quickly on a Friday night.

What the Book Does Not Cover That You Will Need

There are no beer cocktails. No highballs beyond the most basic rum and gin varieties. No discussion of irrigation or shaking technique beyond the bare instruction to shake with ice. No guidance on ice quality, which is probably the single most impactful variable in cocktail making and the book treats it as irrelevant. It also has zero recipes for drinks that require blending or muddling beyond the most trivial cases. If you are looking for a comprehensive guide to cocktail technique, this is not it. It is a recipe collection from a specific hotel at a specific moment in time. The champagne section is also narrow. It has about a dozen champagne cocktail recipes but none of them account for modern champagne that is significantly drier than what was available in 1930. The Sugar Bowl, for example, calls for a quarter ounce of sugar syrup per glass of champagne. Most contemporary Brut champagnes will make that drink cloying. I usually reduce the sugar to a bar spoon and taste before committing.

A Practical Workflow

Here is what I actually do when I pull recipes from this book. I open the archive.org PDF on one screen and my notes on the other. I scan the category I need, write down the recipe in standardized measurements, note any ingredient substitutions I think are necessary based on what is actually available, and then make a test batch at half strength before putting it on a menu. This usually takes me about ten to fifteen minutes per recipe. If a recipe feels off after the test batch, I adjust the balance rather than forcing it to match the original exactly. The original is a historical document, not a mandate. The book is worth having on your shelf. Just do not pretend it is the final word on anything. It is one reference among many, and it shows its age in the places that matter most.