by Dave Arnold
This, then, is the second part of the mega cocktail science post. For the science of ice, temperature and dilution, see Part 1.
Here Iâ€™ll deal with proper temperature, proper dilution, and the different qualities of shaken and stirred drinks. I will also talk about good batching techniques, and then youâ€™ll get, at no extra charge, the Thomas Waugh section promised last week.
Many folks think all drinks should be diluted some standard amount; 25% is often bandied about. Not true — there is no single ideal dilution. In Part 1 I showed that the average stirred cocktail is both warmer than, andÂ less diluted than, the average shaken cocktail â€“so at a minimum there are two different ranges of good dilutions: one for shaken and one for stirred drinks.Â But it’s more complicated still. In a series of blind taste tests with bartenders Kenta Goto, Scott Teague, Eben Klemm, Don Lee, Chad Solomon and Christy Pope, we made Pegu Club cocktails and sidecars with different known dilutions, hoping to find the ideal. No luck.Â Preferences depended on a number of factors including palate fatigue (is this your first drink, or your third?), and what had been consumed prior and after. A drink tasted balanced at a low ABV (alcohol by volume), went out of balance as the ABV went up, but then come back into balance again at an even higher ABV. The more components there are in a drink â€“spirits, acid, sugar, bitters, etc. the more complicated dilution becomes, because each component may respond to dilution differently. For more details see Ideal Dilution Through %ABV.
Many people believe drinks should be as cold as possible. Not true. No drink should be served colder than -20Â°CÂ or itâ€™ll be painful. Straight shots of certain spirits, like 80 proof vodka and aquavit, start to crystallize at about -23Â°C, so if they are clear they probably wonâ€™t hurt. Those same straight shots are really enjoyable between -16 and -18Â°C, where the temperatures accentuateÂ sweetness, minimize the perception of alcohol, and produce a viscous body that is supremely enjoyable. Lucky for us, this is the temperature range of most home freezers. Vodka in the freezer is a good idea â€“ as you already knew. On the other hand, minus 18Â°C is waaaay too cold for most booze and for most mixed drinks. Most over-chilled drinks are dull and lack flavor and aroma. Shaken drinks usually taste best between -5Â°C and -10Â°C. Thatâ€™s lucky, because properly shaken drinks are right in that range. Most stirred drinks taste best between -0.5Â°C and -7Â°C. Lucky again;Â most stirred drinks are in that range.Â These observations are my general rules of thumb derived from side by side taste tests over the years. Many people enjoy over-chilled drinks because they like the sensation of biting cold.Â Thatâ€™s fine, of course â€“but if they analyze the overall drink, theyâ€™ll notice that flavor is compromised.
Temperature can also radically affect the balance of a drink. Iâ€™ll give you an example:
Bottle Strength GnTâ€™s:
A while back I was in the habit of making bottle strength gin and tonics.Â Iâ€™d re-distill gin to bring it above bottle proof without changing the flavor profile, then add quinine sulfate, sugar, clarified lime juice, a pinch of salt, and enough water to bring me back to bottle proof.Â Iâ€™d chill the whole batch down to -20Â°C and carbonate it.Â I kept the bottle in a -20Â°C chiller and poured shots as needed.Â By the time the customer drank the shot, itâ€™d be up to about -18Â°C.Â Remember when I said most drinks arenâ€™t good this cold? Well, I balanced this drink specifically to taste good at that temp.Â They were delicious â€“carbonated, stinging cold, just the right acidity and sugar, not overly alcoholic tasting, with a syrupy body.Â Problem was, those shots only tasted balanced between -18Â°C and -15Â°C.Â Any warmer and they started to go out of balance â€“way out of balance.Â The taste just flew apart.Â The problem was so bad that I would get visibly angered if people didnâ€™t drink the shot right away. Since I canâ€™t guarantee people will not sit about idling chatting instead of drinking, I no longer serve the drink. Bottle strength GnTâ€™s are an extreme example, but the balance of any drink is affected by temperature.
Postulate of Classic Cocktails
I have distilled many spirits and made many drinks that, like the bottle strength GnT, need precise temperatures and dilutions to be delicious.Â Â I have realized that these spirits and drinks will never become popular, and they don’t have what it takes to become classics.Â My postulate of classic cocktails:Â Â classic drinks are those that maintain deliciousness over a wide range of temperatures and dilutions â€“ they can withstand bartender abuse and interpretation, and they don’t wither in front of a lazy drinker chatting up their date.
Temperature and dilution are fairly easy to measure.Â Texture is much more difficult.Â After our first Science of Shaking seminar at least yearâ€™s Tales of the Cocktail, everyone wanted to know how we would measure texture, and whether a bartenderâ€™s shaking style had a significant influence on drink texture.Â We set out to find the answers.
In our first test we had Alex Day, Kenta Goto, Don Lee, and Chad Solomon all shake daiquiris. Eben Klemm, Audrey Saunders, Nils and IÂ tasted the drinks blind.Â Apart from varyingÂ ice crystal quantities floating on the surface, which we ascribe to straining technique, the drinks tasted fairly similar.Â See the post here.
That different bartenders using different shakes produced similar drinks, in conjunction with our findings that shaking style and ice types didnâ€™t affect temperature and dilution, led me to believe that unless a drink contained a foaming agent like egg white, the texturizing effects of shaking were fleeting.Â I figured the air bubbles and ice crystals in a shaken drink just floated to the top and were gone after the first sip. Wrong.
Shaking versus Stirring â€“the blind-fold taste test:
We gathered another group of bartenders: Eben, Thomas, Joaquin Simo of Death and Co, Dana Cory of Lady Jay’s, Karen Jarman of Painkiller, and Nils, Nastassia, and me.Â We chose two classic stirred cocktails: the Manhattan and the Negroni.Â We made each drink four ways: stirred fast, stirred slow, shaken and poured through a single strainer, and shaken and poured through a double strainer.Â The drinks looked so visibly different that we tasted them wearing blindfolds –a real blind taste test. To my surprise, every one of us was able to distinguish between the shaken and stirred drinks.Â Almost always we liked the stirred ones and dis-liked the shaken ones.Â The only exception was the double-strained Negroni, which some people thought was good –but not as good as the stirred. Although these tests didn’t control for dilution, It was clear that dilution wasn’t the only thing making the cocktails different. Our ability to tell the difference between the drinks was maintained for at least 5 minutes, but disappeared by 8 minutes. Contrary to what I had believed, a textural elementÂ introduced by shaking was maintained for at least several minutes. Also –turns out we like our stirred drinks stirred, not shaken.
Fake Shake 1:
If shaking really does introduce some lasting texture to the drink, how do you emulate thattaxtureÂ in large demos where it is impossible to shake to order?Â We blind tasted 4 daiquiris: fresh shaken, diluted and chilled in a freezer, diluted and chilled in a freezer then spun in a blender for service, and chilled with Liquid Nitrogen.Â Everyone liked the fresh shaken best.Â The Liquid Nitrogen was a close second.Â The blender was bad –the bubbles were too big and frothy (this is no indictment of blender drinks –this drink had no ice, just liquid).Â The merely chilled cocktail was what it was –rather flat.
We made drinks, poured them on the white table, and took pictures of them with the Eyeclops.
Under 200x magnification, stirred drinks appear blank, but both strained and unstrained shaken drinks had lots of tiny bubbles — straining a shaken drink removes ice crystals, but not air bubbles.Â Drinks chilled with liquid nitrogen, which aerates as it chills, also had bubbles, but not quite as many as the shaken drink –lending credence my theory that LN is the best chilling mechanism for shaken drinks inÂ super-high-volume scenarios.
Fake Shake 2:
Not everyone has liquid nitrogen, so we tested some easier ways to fake the shake. I made a batch of whiskey sours, some of which I pre-diluted and put in the freezer.Â At tasting time I shook one drink to order, poured one straight out of the freezer, shook one of the pre-chilled guys in a quart container, and shook the last one in a quart container filled with the springs from hawthorn strainers, for extra aeration. The drinks were noticeably different.Â One taster liked the drink straight from the freezer –a texture-hater apparently, because he liked the drink shaken with the springs the least. Most people liked the freshly shaken drink best, with the drink shaken in the quart container a close second.Â The one shaken with springs was similar to the other two shaken drinks, just a little more airy.
Stir Texture = No Texture,Â and the best stirred drink:
I have been saying for a while that stirring is a technique that chills without providing texture –it adds nothing extra.Â To test this theory we blind-tasted drinks stirred live versus drinks that were pre-diluted and chilled.Â They were indistinguishable.Â The best stirred drinks, therefore, can be made by pre-batching the drink with water and chilling it in a freezer.Â Just make sure to use a freezer that isn’t too cold.Â You don’t want to serve the drink much below minus 7C.Â Drinks made this way are effortless and consistently perfect.
The last question to tackle: how to figure out the amount of water to add to your batch recipes.
Scientific Batch Recipes:
If you want to pre-batch a cocktail and chill it in the freezer, it is helpful to know how much water to add. Diluting to taste isnâ€™t a good idea. Itâ€™s likely you wonâ€™t be diluting at the proper serving temperature, and temperature has an effect on your perception of dilution.Â Hereâ€™s the best way to come up with a batch recipe:
Gather enough ingredients to make a cocktail and mix them together without ice. Better yet,Â mix enough for two cocktails and your measurement inaccuracies will be proportionally less;Â if you make more than two drinks at once the mix might not shake or stir properly.Â Make sure you measure by volume (most cocktail recipes are by volume, and the density of cocktail ingredients differs widely). Make sure you measure accurately â€“ no â€œhalf a lemonâ€ nonsense.Â Bitters are tough. I tend to ignore them in my calculations unless a lot is included. Weigh the undiluted cocktail as accurately as possible. Now chill the drink exactly as you normally would.Â If you normally stir for 20 seconds, do that.Â If you normally shake for 15 seconds, do that.Â When you are done, strain the cocktail off the ice. It is important to leave as little of the drink in the shaker or mixing glass as possible.Â Now weigh your drink.Â Â Taste the drink.Â Do you like it? Then, bang â€“ you know how much dilution that drink should have.Â If you donâ€™t like it, go back and repeat â€˜til you do.
While the above technique is a good start, it introduces major inaccuracies because some of the cocktail remains trapped in the ice after you strain it — we call this trapped liquid ‘holdback’, and you need to correct for it.Â Hereâ€™s how: after you strain your cocktail, dump the ice into a tray, recover the last bit of liquid from the ice, weigh it, determine what percentage drink you think it is (this is a non-scientific, best guess kind of thing â€“like hey that tastes like 75 percent water). Do the math (see below) and re-calculate the cocktail based on the estimated holdback.
Two Example Batch Calculations (With and Without Holdback Correction):
Whiskey Sour Basic Recipe
2 parts whiskey 45% abv
1 part strained lemon juice
0.5 parts 1:1simple syrup
Tiny pinch salt
Total Batch: 3.5 parts
We jiggered up one cocktail and it weighed 101 grams. Eben Klemm shook it using his normal technique.Â When we strained the drinkÂ it weighed 157 grams (it was -4Â°C).Â We then dumped the ice into a tray and recovered 18 grams of liquid.Â We tasted it and it seemed to be about 25% initial mix. This is unscientific, but the best we could do without good equipment.
Water calculation with holdback (more accurate):
Water calculation with holdback:
Manhattan Basic Recipe with proper dilution:
2 parts rye whiskey 45%abv
1 part sweet vermouth
some bitters (sorry, I said bitters were tough)
1.5 parts water
Total batch: 4.5 parts. Total dilution (percent of initial mix volume in water added to drink): 50% !Â Approximate Final ABV: 20%
Secret Bonus for Reading this Far â€“Salt:
The secret ingredient in our cocktails isnâ€™t love, itâ€™s salt. A pinch of salt added to most cocktails brightens them up and rounds out the flavor tremendously. You donâ€™t need a lot.Â Add so little you donâ€™t even perceive saltiness â€“I call it “sub-threshold salting”.Â Sugar and vanilla also do interesting things in sub-threshold quantities.Â When I distill liquor, I usually add a small amount of salt and sugarÂ – not enough sugar to add sweetness.
Bonus â€“Thomas Waugh from Death and Co tells us how all this business has affected the way he tends bar.
by Thomas Waugh
Upon completing our investigation on the variables of stirring, here are some humble suggestions for the everyday bartender:
1.Â Â Chill your mixing (stirring) glass — ice works, as does a fridge or freezer.
2.Â Â If making multiple drinks, build the shaken drinks first, but without ice, and set aside.
3.Â Â Continue to build the stirred drinksÂ in the cold mixing glass.
4.Â Â Stir without breaks until the stirred drink is cold (we found that 45 seconds is an ideal stirring period).
5. Crack some ice to maximize surface area. Note: too much cracked ice will hinder the control of your dilution rate.
6. Pour stirred drinks first.
5.Â Â Finish the order by adding ice to your shaken drinks and shaking, as shaking is a much more efficient technique.