Skip to the main content
Salty & Clever

Salty & Clever

The Rescue

One hundred and thirteen ways a dish goes wrong, what actually happened in the pan, and what to do about it now rather than next time. Eight of them cannot be saved, and it says so.

This night

The Rescue

Something has gone wrong and dinner is in forty minutes. This is the page for that.

One hundred and thirteen ways a dish goes wrong, what actually happened in the pan, and what to do about it now rather than next time. Every entry says the mechanism first, because a fix you understand is one you can adapt and a fix you memorized is one that only works on the day you read it.

This page is the emergency half. The other half is why any of it works — six stages of mechanism, read when the kitchen is calm, so that fewer evenings end up here.

Some of these cannot be rescued. Those say so, in the same words, and tell you what the thing can become instead. A page that promises every dish is recoverable is a page that will lie to you at the worst possible moment.

The one rule under all of this

Almost every failure on this page is heat, water or timing — usually heat. A broken sauce, a scrambled custard, a seized chocolate and a tough steak are four names for the same mistake made in four different pans. Once you can see which of the three went wrong, the fix stops being a trick and starts being obvious.

Sauces and emulsions

An emulsion is fat and water held together by something. Every failure here is that something letting go, and almost all of them are heat.

The pan sauce broke and there is a slick of butter on top

What happened The pan was too hot when the butter went in. Butter is an emulsion already, and above about 130°F it stops being one — the fat separates out and floats.

Right now Pull the pan off the heat entirely. Put a tablespoon of cold water or stock in a clean warm bowl, then whisk the broken sauce into it a spoonful at a time. You are building a new emulsion around the water rather than trying to repair the old one. It comes back almost every time.

Next time Butter goes in off the heat, in cold cubes, one or two at a time, swirling. The pan should be hot enough to melt them slowly and not hot enough to sizzle. Every broken sauce is the same accident — one move fixes all six of them.

Hollandaise or beurre blanc split

What happened Same mechanism, one step earlier. Either the butter went in faster than the yolks could take it, or the bowl got hot enough to cook the egg.

Right now If it is grainy as well as split, the egg has cooked and it is finished — see below. If it is smooth but separated, start a fresh yolk with a teaspoon of cold water in a clean bowl and whisk the split sauce into that, slowly, as if it were the butter. A tablespoon of ice water whisked hard into the original sometimes does it on its own, and takes ten seconds to try first.

Next time Off the heat more often than feels necessary, and butter warm rather than hot. If the bowl is too hot to rest your hand against, it is too hot for a yolk.

Mayonnaise will not come together

What happened The oil went in too fast at the start. Before there is enough emulsion to hold it, a yolk can only take oil a few drops at a time.

Right now Put a fresh yolk, or a teaspoon of already-made mayonnaise, in a clean bowl. Whisk the failed batch into it drop by drop until it thickens, then in a thin stream. Nothing is wasted and it takes two minutes.

Next time Everything at room temperature, and the first quarter cup of oil goes in slower than you believe is necessary. A cold yolk straight from the fridge is the most common reason a mayonnaise fails.

The vinaigrette separates before it reaches the salad

What happened There is nothing in it holding the oil and the acid together. Oil and vinegar shaken alone will always separate, usually within a minute.

Right now Whisk in half a teaspoon of mustard, or a spoonful of honey, or a mashed anchovy, or an egg yolk. Any of them will hold it for the length of a meal. Mustard is the one most kitchens have and the one that changes the flavor least.

Next time Build it in the order that works: acid, salt and the emulsifier first, whisked until the salt dissolves, then the oil in a stream. Salt does not dissolve in oil, so a dressing built the other way is both separated and underseasoned.

The cheese sauce went grainy and stringy

What happened The cheese boiled. Above a simmer the proteins tighten into strands and squeeze the fat out, which is exactly what you are looking at.

Right now Off the heat, then whisk in a splash of evaporated milk or a teaspoon of cornstarch slurried in cold water. Both work by giving the protein something else to hold onto. A pinch of sodium citrate, if you happen to own it, fixes this instantly and completely.

Next time The cheese goes in off the heat, grated rather than cubed, in handfuls, once the base is thick. Aged cheeses break more readily than young ones, which is why a mac and cheese made entirely of sharp cheddar is harder than one cut with something young and mild.

The gravy is lumpy

What happened Flour hit hot liquid faster than it could hydrate, so each lump is dry flour wearing a cooked skin.

Right now Pour it through a sieve. That is the whole answer and it takes twenty seconds. A blender works too and gives a smoother result, but it also knocks the air in and can go pasty, so the sieve is the better move if the gravy is thin enough to pass.

Next time Either cook the flour into the fat first and add the liquid in stages, whisking, or shake the flour with cold water in a jar and pour that in. What never works is flour straight into hot stock.

The sauce is too thin and the food is ready

What happened Not enough time, or too much liquid, or both. Reduction is the only thickener that also concentrates flavor, and it is the one that takes the longest.

Right now If you have five minutes, move the sauce to a wider pan and boil hard — surface area, not heat, is what reduces a sauce. If you have one minute, whisk in a teaspoon of cornstarch slurried in cold water and bring it to a boil for thirty seconds to lose the raw taste. Cold butter whisked in off the heat thickens a small sauce and makes it glossy at the same time.

Next time Reduce before the protein comes out of the pan, not after, so the resting juices go into a sauce that is already the right thickness.

The sauce is too salty

What happened Reduction. Salt does not evaporate, so a sauce that was correctly seasoned at the start is over-seasoned by the time it has halved.

Right now Add volume, not potatoes. The potato trick does not work — a potato absorbs salted liquid at the same concentration and removes nothing. What works is more of the unsalted thing: cream, stock, water plus a little more of everything else, or the starchy water the pasta is sitting in. Acid and a little sugar will not remove salt but they will change what your tongue notices first, which sometimes is enough.

Next time Season a sauce at the end. If it has to be seasoned early, season it to about two thirds and finish it when the reduction stops.

The sauce tastes of raw flour

What happened The starch finished before the flour did. Flour granules swell and set well below a simmer, so a sauce turns thick and glossy minutes before the raw cereal taste has gone anywhere.

Right now Put it back on low heat at a bare simmer for another 10 to 20 minutes, scraping the corners of the pan. It will not get appreciably thicker; the starch has already done its work. If there is no time, butter and a spoonful of cream stirred in off the heat cover a good deal of it, and so does anything assertive: mustard, grated cheese, hard-reduced stock.

Next time Cook the roux until it smells like toast. A well browned roux thickens roughly half as much as a pale one, so a dark roux needs about twice the flour.

A skin formed on the sauce

What happened Water evaporated off the surface and left protein behind. Concentrated enough, it coagulates into a film — dairy skins fastest, but a gelatin-rich stock sauce does the same left still on the warm end of the stove.

Right now Lift it off with a fork or skim it with a spoon and throw it away. Whisking it back in looks like the tidier move and is not: the film is already set, so it breaks into specks that stay visible in a pale sauce. If it has already been whisked in, pour the sauce through a fine sieve and it will come out clean.

Next time Press plastic wrap flat onto the surface so it touches, rather than covering the pan. A thin film of melted butter floated over the top holds a sauce for an hour and gets stirred in at the end.

The tomato sauce tastes tinny

What happened Part of it is the fruit’s own acidity. The rest is metal: that acid pulls flavor out of anything reactive it touches — bare cast iron, unlined aluminum, the opened can it stood in.

Right now Get it into stainless or enamel now, or it will keep going. Fat is the most reliable tool: 2 tablespoons of butter or a splash of heavy cream per quart, then five minutes of gentle cooking. Salt suppresses the sharp edges the metal rides on. A ⅛ teaspoon of baking soda per quart, stirred in off the heat, takes the acid down a step; past that it tastes flat and soapy, and there is no undoing it.

Next time Taste the tomatoes straight from the can. Brands differ more than anything you do to them afterward, and one that tastes metallic cold will taste metallic in three hours.

The reduction turned bitter

What happened Either the pan scorched, because a thick sauce stops circulating and the bottom overheats, or the wine reduced past what the rest of the sauce can carry.

Right now Tilt the pan and look. A dark film welded to the bottom means scorch, which runs through the whole liquid. Nothing takes it out; pour off the top without disturbing the bottom, and start again with fresh stock if it is still bitter. If the pan is clean and the bitterness is sharp rather than burnt, it is over-reduction: add unsalted stock or water to bring the volume back, then whisk in cold butter off the heat. Salt blunts bitterness better than sugar.

Next time Reduce the wine by itself to about half before the stock goes in, and stop while the sauce still moves freely. The last two minutes are when a reduction burns: by then it is thick enough to insulate its own bottom.

Eggs and custard

Egg protein sets in a window about fifteen degrees wide. Nearly everything below is that window being missed on the high side.

The custard scrambled

What happened The egg passed about 180°F and set into curds instead of thickening smoothly. It is cooked, and cooked protein does not uncook.

Right now If it is faintly grainy, an immersion blender and a fine sieve will rescue it — blend, strain, chill. If it looks like sweet scrambled eggs, it is past saving as custard. Strain out the curds, sweeten the liquid and use it as a thin sauce, or fold the whole thing into a bread pudding, where it is going to be baked anyway and nobody will know.

Next time Custard is done at 175 to 180°F, which is well below a simmer and looks like almost nothing is happening. Off the heat at the first sign of thickening, and a thermometer costs less than the ingredients.

The scrambled eggs went watery on the plate

What happened Overcooked. The proteins tightened and squeezed out the water they were holding — the puddle on the plate came out of the eggs.

Right now Nothing brings it back, but a spoonful of cream cheese, creme fraiche or grated cheese stirred in off the heat will bind the free water and the plate will look and eat considerably better.

Next time Lower heat, constant movement, and off the pan while they still look slightly underdone. They carry on cooking in their own heat for about thirty seconds after they leave it. Salt at the start rather than the end, which sounds like the opposite of the usual advice and produces a more tender curd.

The omelette stuck and tore

What happened Either the pan was not hot enough when the eggs went in, or there was not enough fat, or the pan is one that eggs stick to and always will.

Right now Stop trying to fold it. Push it into a rough heap, call it a scramble, and put the cheese on top. Nobody at a table has ever been disappointed to be handed a scramble.

Next time Heat the pan first, then the fat, then the eggs — in that order, and give the fat ten seconds to get hot. A drop of egg should set at the edge on contact without browning.

The meringue is weeping or beading amber drops

What happened Undissolved sugar. Beads on the surface are syrup drawn out of the meringue by sugar that never went into solution; weeping underneath is usually an under-beaten meringue, or one that met a warm filling.

Right now Beading is cosmetic and the meringue is fine to eat. Weeping under a pie is not fixable after baking. Serve it cut rather than whole so the layer is not the first thing anyone sees.

Next time Add the sugar slowly to whites that are already at soft peak, and beat until a smear between two fingers feels smooth rather than gritty. Spread meringue onto a filling that is still hot, so the underside cooks rather than steaming.

The lemon curd did not set

What happened It did not get hot enough. Curd thickens at around 170°F and not before, and it looks alarmingly thin right up until it does.

Right now Put it back on gentle heat and keep stirring until it coats the back of a spoon and holds a line drawn through it. If it has already been chilled, reheat it slowly in a bowl over water rather than in a pan, and strain it afterward.

Next time Trust the thermometer over the clock. Curd also firms considerably as it chills, so the pan should look like a pourable sauce, not a set one.

The poached egg shredded into the water

What happened The thin outer white dispersed before it could set. Every egg has a looser white outside the thick one that gets looser with age, and water at a real boil pulls it apart faster than heat can set it.

Right now It will not come back together. Lift out what did set and trim the ragged edges against the side of the spoon; it eats perfectly well. For one more attempt, crack the egg into a fine-mesh sieve, let the thin white drain for thirty seconds, and slide what is left into water at 180 to 190°F, where the surface only shivers.

Next time The sieve and the age of the egg do nearly all the work; vinegar and a whirlpool do less than their reputation suggests. The fresher the egg, the better it holds — the exact opposite of what hard boiling wants.

The hard-boiled eggs will not peel

What happened The white is bonded to the membrane lining the shell, which is what a fresh egg does. An egg loses carbon dioxide as it sits and the white turns alkaline; an alkaline white lets go cleanly.

Right now Nothing unsticks a white that is already bonded. Crack the shell all over and peel under water, starting at the fat end where the air pocket sits; water gets under the membrane and does most of the lifting. Expect to lose the first one or two. Pitted eggs are still good eggs: chop them for egg salad.

Next time Lower them into already-boiling water, or steam them, so the outside of the white sets hard in the first minute and shrinks off the shell: 12 minutes steamed for a firm yolk, then ice water for 15. Buy the eggs a week ahead — the one dish where an older egg is the better egg.

The frittata is rubbery

What happened The egg went past about 180°F. The proteins tightened, squeezed out the water they were holding, and what is left is dense and bouncy. Cooked protein does not loosen again.

Right now Nothing brings the texture back. Serve it warm rather than hot or straight from the fridge, cut it into smaller pieces than you meant to, and put something wet with it: crème fraîche, salsa, dressed arugula. Cold is where a rubbery frittata is at its worst, so leftovers should sit out twenty minutes first.

Next time Bake at 325°F and take it out when the center reads about 160°F and still wobbles. A cast iron pan carries it ten degrees further on the counter, so slide it onto a plate rather than resting it in the pan. About ¼ cup of cream or ricotta per six eggs makes it harder to overcook.

The quiche filling is watery

What happened Either the custard overbaked and wept, or water came out of the filling. Raw spinach, mushrooms, zucchini and tomatoes give up a great deal of it in the oven.

Right now Tilt the pan and blot the pooled liquid off with a folded paper towel. Weeping does not reverse, and more oven time makes it worse. The exception is a quiche whose center is genuinely liquid rather than merely soft: that one is underbaked, and goes back at 325°F until the center reads about 170°F. Either way, give it fifteen minutes before cutting — custard firms as it cools, and a hot quiche always looks looser than it is.

Next time Cook the wet ingredients down first and squeeze them dry; a pound of raw spinach carries most of a cup of water into a pie. Bake at 325°F rather than 375°F, and use one large egg per ½ cup of dairy.

The egg wash dried streaky

What happened The wash was never one liquid. A whole egg out of the shell is thick white, thin white and yolk in ropes beside each other, and the brush drags them across the pastry, where they take color at different rates.

Right now If it has not gone in the oven yet, beat the egg hard with a pinch of salt, leave it five minutes while the salt loosens the white, then strain it and lay one thin coat over the streaks, brushing in one direction. If it is already baked, the streaks are set and will stay: powdered sugar or a brush of warmed, strained apricot jam evens out a sweet pastry, and a savory one is fine as it is.

Next time Use two thin coats rather than one heavy one. Keep it off the cut edges of puff pastry, where it glues the layers together and stops that side from rising.

Bread and dough

Three variables: yeast, water and gluten. Nearly every problem here is one of them, and the dough will tell you which if you press it.

The dough did not rise

What happened Dead yeast, a cold kitchen, or salt dumped directly onto the yeast at the start. In winter, a cold kitchen is the answer far more often than dead yeast.

Right now Give it warmth and time before you give up on it — an oven with only the light on sits at about 80°F, which is right. If there is genuinely no activity after another two hours, dissolve a fresh packet of yeast in a little warm water, knead it through, and let it go again. The dough is not wasted.

Next time Proof the yeast in warm water with a pinch of sugar and watch it foam before anything else happens. Water hotter than about 140°F kills it, and hot tap water can be hotter than that. The longer version, in the order the causes are actually the problem: your dough is probably not dead, it is probably just cold.

The bread is dense and gummy in the middle

What happened Under-baked, almost always. A loaf that looks brown outside can be twenty degrees short in the middle, and the starch has not set.

Right now If it is still warm, put it back in a 350°F oven for another ten to fifteen minutes. If it has cooled, the crumb will not recover — slice it thick and toast it, or turn it into croutons or breadcrumbs, both of which it is now better suited to than it ever was to sandwiches.

Next time Bake to internal temperature, not to color: 190 to 200°F for a soft loaf, 205 to 210°F for a lean crusty one. And let it cool completely before cutting — bread finishes cooking as it cools, and a loaf cut hot goes gummy on the knife.

The dough is too sticky to handle

What happened Usually nothing. Most enriched and high-hydration doughs are supposed to be sticky, and the instinct to fix it with flour is what makes the bread dry.

Right now Wet your hands instead of flouring the counter, and use a bench scraper. Give it a ten minute rest before kneading — the flour keeps absorbing water for a while after mixing, and a dough that was unmanageable often behaves after a rest without a gram of extra flour.

Next time Weigh the flour. A cup of flour scooped from the bag can be twenty percent heavier than a cup spooned in, which is the single largest source of variation in home baking.

The pizza dough springs back and will not stretch

What happened The gluten is tight. It is doing exactly what gluten does, and it is not a fault.

Right now Walk away for fifteen minutes, covered. It will stretch without a fight. Fighting it now tears the dough and makes it tighter still.

Next time Take it out of the fridge an hour before it is needed. Cold dough is tight dough, and a cold ball will not relax in the ten minutes people usually give it.

The crust is pale and soft

What happened Not enough heat where it matters, or the oven was opened. Bread browns from radiant heat and from steam early in the bake, and both leave when the door does.

Right now Straight back in, on the highest rack, at 425°F, for five to ten minutes. It will not be worse and it is usually enough.

Next time Preheat longer than the oven’s own beep — a stone or a heavy tray wants forty-five minutes, not fifteen. Steam for the first ten minutes, then let it dry out and color.

The pastry shrank down the side of the pan

What happened The dough was stretched into the pan rather than laid into it, and it pulled itself back during the bake. Gluten remembers.

Right now Patch the wall with a little raw dough or a paste of butter and flour and bake another five minutes. A tart with a short wall still works; a tart with a hole does not.

Next time Ease it in, never stretch, and chill the lined pan for thirty minutes before it goes near the oven. Cold fat and rested gluten between them account for nearly all of this.

There is a hollow gap under the crust

What happened The crust set while the crumb was still expanding, and the two parted company. Usually the loaf went in under-proofed, so most of its rise happened in the first ten minutes against a top that had already dried and hardened.

Right now Nothing closes it. The loaf is baked and the gap is where it will stay. It is a cosmetic fault and not a flavor one, so slice it and serve it. The slices from the domed end will be fragile and want to tear along the void — cut those thicker, about ¾ inch, and send the ragged ones to toast, croutons or stuffing.

Next time Keep the dough covered through the final rise so the surface never dries into a skin, and proof it further than feels right: a floured fingertip should leave a dent that fills back only about halfway in ten seconds.

The loaf over-proofed and collapsed in the oven

What happened The gluten stretched past what it could hold. The bubble walls thinned and ruptured into one another, so when the gas expanded in the oven heat there was no structure left to push against.

Right now It is past saving as a loaf, and the dense band near the base will stay dense. It is still good for anything where bread is soaked or re-dried: cube it and dry the cubes at 300°F for about 20 minutes for croutons, or slice it for French toast. Long fermentation is good for flavor, which is why this failure tastes better than it looks.

Next time Proof by the dough, not the clock. Recipe times assume a kitchen around 70°F; in a warmer one the same dough is usually ready sooner, by an amount that depends on the yeast, the sugar, the salt and the dough itself. Under-proofed bread is still good bread and over-proofed bread is not, so err short.

The loaf tore open down one side

What happened The crust set before the loaf had finished expanding, so the gas pushed out at the weakest point. A shallow score, or dough that went in slightly under-proofed, sends it out the side.

Right now Leave it in the oven. The tear is cosmetic, and pulling the loaf early trades a scar for a gummy center. Bake until the middle of a lean loaf reads 205 to 210°F, and tent the torn edge with foil if it is browning ahead of the rest. Serve it tear side up and slice from the middle out; nothing closes it, and nobody minds.

Next time Score deeper than feels right, ¼ to ½ inch, one confident cut held at a shallow angle, and cover the loaf or add steam for the first ten minutes. A timid score is what makes a loaf burst somewhere else.

The pie crust came out tough

What happened Gluten. Water is what lets the proteins in flour link into an elastic network, and every extra spoonful of it and every extra pass of the rolling pin builds more.

Right now Baked gluten does not relax, so nothing will make the crust tender again. Warmth helps more than anything else: 15 minutes at 300°F softens the fat and makes the difference between a crust that cuts and one that shatters and drags. Cut it with a serrated knife, sawing rather than pressing down, serve it warm, and put cream or ice cream alongside.

Next time Use less water than looks right, then rest the dough in the refrigerator overnight. Flour goes on absorbing water for hours after mixing, so a dough that seems too dry and crumbly at the counter is usually correct and will roll out cleanly once rested.

The pie has a soggy bottom

What happened Water from the filling soaked into the bottom crust before its starch had set. The bottom of a pie is its coldest part, shielded by an inch of cool filling and sitting on a pan that heats slowly, so it steams while everything above it bakes.

Right now It is often retrievable: the one thing that drives water back out is heat from underneath. Put the pie on the lowest rack at 425°F for 10 to 15 minutes, with foil laid loosely over the top if the surface is dark. Lift an edge with a spatula and look at the underside rather than guessing. A bottom that is genuinely raw paste will improve without ever becoming crisp.

Next time Take the blind-bake further than looks comfortable: pale gold is not baked. Metal and glass pans pass heat into the base; thick ceramic insulates against it.

Cakes and cookies

Baking fails at a distance: the mistake was made twenty minutes before the thing that looks wrong. These entries point back to where it actually happened.

The cake sank in the middle

What happened The structure had not set when the lift ran out. Usually an oven door opened early, sometimes too much leavening, occasionally an oven that is thirty degrees cooler than it claims.

Right now Cut the crater out with a round cutter and call it a ring cake, or level the whole top off and turn it into a layer cake where the frosting is doing the work. Sunken cake is also the correct raw material for a trifle, and a trifle looks deliberate.

Next time No opening the door in the first two thirds of the bake. Buy an oven thermometer — home ovens are routinely off by twenty-five degrees and yours has never been calibrated. Five things make a cake sink and they leave five different craters — read the crater to find out which was yours.

The cake domed and cracked down the middle

What happened The oven was too hot. The edges set while the middle was still rising, so the middle had nowhere to go but up and out.

Right now Level it with a serrated knife once it is cool. This is what every bakery does and nobody photographs it.

Next time Twenty-five degrees lower and longer. If it keeps happening, wrap a damp strip of towel around the outside of the pan so the edges heat at the same rate as the middle.

The cake is dry

What happened Too long in the oven, too much flour, or both. Five extra minutes matters more in a cake than in almost anything else you bake.

Right now Brush the layers with simple syrup — equal sugar and water, boiled and cooled, a tablespoon or two per layer. This is standard practice in professional kitchens and it is not cheating. A dry cake under syrup and frosting is a good cake by the time it reaches the table.

Next time Weigh the flour, and start checking ten minutes before the recipe says. A skewer should come out with a few damp crumbs, not clean — clean means you are already past it.

The cookies spread into one continuous sheet

What happened The butter was too soft or the dough was too warm when it went in, so the fat melted and ran before the structure set. A hot tray reused straight from the last batch does it every time.

Right now While they are still hot, push the edges back in with a spoon or cut them apart with a knife. For the rest of the dough, chill it for thirty minutes and use a cold tray.

Next time Softened butter means it holds a fingerprint, not that it is glossy. Chill the shaped dough. Never put dough on a tray that is still warm. Four levers decide how far a cookie spreads, and only one of them is the butter.

The cookies did not spread at all

What happened Too much flour, or the butter and sugar were creamed until they were very aerated, or the oven is cool.

Right now Take the tray out at the halfway mark and flatten each one with the back of a spoon, then finish the bake. It works and nobody can tell afterward.

Next time Weigh the flour. Cream the butter and sugar for two or three minutes, not eight — a very aerated dough sets before it has time to spread.

The brownies came out cakey and I wanted fudgy

What happened Too much air, too much flour, or too long. Fudgy is a ratio question — more fat, less flour, fewer eggs, and no creaming.

Right now Nothing changes a baked brownie. Chill the tray thoroughly and cut small: cold makes a cakey brownie eat denser than it is, which is most of the difference.

Next time Melt the butter rather than creaming it, stir rather than beat, and pull them out when the middle still looks underdone. A brownie that looks ready in the oven is a cake by the time it cools.

The cheesecake cracked

What happened It cooked too fast or cooled too fast. The surface set and then the middle shrank underneath it.

Right now Sour cream topping, fruit, or whipped cream. This is what a topping is for and it is the reason so many classic cheesecakes have one.

Next time Water bath, low oven, and then leave it in the switched-off oven with the door ajar for an hour. Most cracks happen in the twenty minutes after the cake comes out, not while it is in.

The cake is stuck in the pan

What happened Sugar at the surface caramelized against hot metal and bonded to it, and as the pan cooled that layer set into a hard glue.

Right now Heat first, then leverage. Set the pan back over low heat for 30 to 60 seconds to soften that sugar layer. Run a thin offset spatula around the side, then invert onto a rack and knock the far edge once, firmly. A cake out of the oven less than an hour usually releases. A detailed bundt will leave pieces behind whatever you do, and a broken cake is still a cake: layer it with cream and fruit as a trifle.

Next time Turn cakes out at 10 minutes. Sooner and they are too fragile to hold their own weight; later and the sugar has set against the metal. Grease with a paste of equal parts flour, softened shortening and oil — spray on its own runs down the vertical walls before the batter arrives.

The cake layers will not stack level

What happened The layers came out of the oven at different thicknesses, or domed. A dome is thickest at the center, so each layer tilts the one above it further.

Right now Chill the layers twenty minutes first; cold cake cuts clean and does not slide under the knife. Level each with a long serrated knife held parallel to the counter, sawing lightly rather than pressing down. Stack the thickest layer on the bottom and set the top layer cut side down, so no cut surface faces up under the frosting. Pipe a ring of stiff frosting around each layer to dam the filling, which is what usually lets a stack slump.

Next time Weigh the batter into the pans instead of eyeballing it, within half an ounce. Domes come from an oven running hot, so 325°F for longer gives flatter layers than 350°F.

The buttercream split into curds

What happened Temperature, almost always. Buttercream is an emulsion of fat and water that holds only in a narrow band around 65 to 68°F. Below that the butter is too firm to take up the liquid and it breaks into grains; above it the fat melts and separates.

Right now It comes back almost every time. If it is cold and curdled, scoop out about ½ cup, melt it, and drizzle it back into the bowl with the mixer running. If it is warm and soupy, refrigerate the whole bowl for 15 to 20 minutes and beat it again. Either way, keep the mixer going a full 5 minutes before deciding it has failed.

Next time Bring everything to the same temperature first: butter at about 65 to 68°F, the base cooled to match, liquids at room temperature. A meringue buttercream that curdles partway through the butter needs more beating, not more butter.

The frosting is grainy

What happened Sugar crystals that never dissolved. Sugar dissolves in water and not in fat, and a frosting that is mostly butter holds very little water, so a crystal that goes in whole tends to stay whole.

Right now Add liquid and warmth, in small amounts. Beat in a tablespoon of heavy cream, then another, and run the mixer 5 minutes; that alone clears most cases. If it is still gritty, warm the bowl gently until the frosting is around 100°F, stir until it looks smooth, then cool it back to room temperature and re-whip. Grit that came from granulated sugar will soften but is unlikely to vanish altogether.

Next time Sift the confectioners’ sugar, and dissolve it in the cream before the two go near the butter — crystals dissolve readily in liquid on its own and hardly at all once fat has coated them.

The cookies came out cakey instead of chewy

What happened Chew comes from water held in the dough by sugar and from fat that lets the cookie spread before it sets. A cookie that stands tall has usually had extra flour, baking powder in place of baking soda, or butter creamed pale and fluffy.

Right now The baked ones will not become chewy; that texture is fixed once the starch has set. Sealed overnight with a slice of bread they will soften, which is not the same thing but is an improvement. If there is dough left, the rest of the batch is savable: flatten each ball to about ½ inch, bake at 375°F, and pull the sheet at 9 to 11 minutes while the centers still look underdone. They finish on the hot pan.

Next time Melt the butter rather than creaming it, and shift the sugar toward brown — molasses is hygroscopic and holds water in the finished cookie in a way white sugar does not.

Rice, grains and pasta

Starch and water, and a narrow window between the two.

The rice is mushy

What happened Too much water, or the lid came off and it was stirred. Stirring rice releases starch, which is what you want in a risotto and the last thing you want in a pilaf.

Right now Spread it on a tray and put it in a low oven for ten minutes to dry the surface. Or lean into it: mushy rice makes excellent congee, rice pudding or fried rice, and fried rice is the answer people are happiest to be given.

Next time Rinse until the water runs clear, measure the water, and do not lift the lid. Most long grain rice wants less water than the package says once it has been rinsed.

The rice is crunchy and the water is gone

What happened The heat was too high, so the water boiled off before the grain absorbed it.

Right now Add a quarter cup of boiling water, lid on, lowest possible heat, five to ten minutes. It comes back. Do not stir.

Next time Bring it to a boil once, then drop to the lowest setting the burner has. If the pot is audibly bubbling for the whole cook, the heat is too high.

The rice stuck and the bottom burned

What happened Heat too high, pan too thin, or both.

Right now Do not stir — that is how the burned flavor gets into the rest of it. Take the pot off the heat, lay a clean towel under the lid for five minutes to pull the steam and the smell out, then lift the good rice off the top and leave the bottom in the pan.

Next time Heavier pot, lower heat. In several cuisines that crust is the best part and is made on purpose, which is worth remembering the next time it happens.

The risotto went gluey

What happened Too much stirring, too little liquid at the end, or it sat. Risotto begins to set the moment it stops moving.

Right now A ladle of hot stock beaten in vigorously off the heat, then the butter and cheese. It should move when you shake the pan — the Italian test is that it spreads slowly on a flat plate.

Next time Serve it immediately, loosen it a little more than looks right, and stir less than the internet tells you. Constant stirring is a myth that produces exactly this.

The pasta clumped into a block

What happened Not enough water, no stir in the first minute, or it drained and sat.

Right now Back into hot water for thirty seconds, or straight into the hot sauce with a splash of the cooking water and a hard toss. Both work. Oil on drained pasta does not fix this, it just stops the sauce sticking later.

Next time Plenty of water, stir for the first minute while the surface starch is releasing, and drain it into the sauce rather than into a colander to wait.

The pasta is overcooked and mushy

What happened The starch granules took on more water than they could hold and burst, and the strands have gone slack all the way through. Cooked starch does not firm back up.

Right now Drain it and spread it in one layer on an oiled sheet pan, because a hot drained pile keeps cooking itself for several more minutes. Then stop trying to serve it as pasta. Fry it in a skillet with a film of oil until the edges crisp, or fold it with sauce and cheese into a baking dish and bake at 400°F until the top browns, where soft pasta is the point. It will not come back al dente.

Next time Start tasting two minutes before the lowest number on the box. Pasta that will finish in the sauce has to leave the water a full minute before it tastes right, because the sauce goes on cooking it.

The sauce slides off the pasta

What happened The sauce and the pasta never became one thing. A sauce clings when its fat and its water are bound together by starch and by agitation; ladled over drained pasta, it is two separate things in a bowl.

Right now Everything back into the widest pan you have, over medium heat. Add ¼ to ½ cup of the pasta cooking water per pound and toss hard for thirty to sixty seconds; the sauce goes glossy and tightens onto the noodles as the starch does its work. If the pasta water is already down the drain, plain hot water and a tablespoon of cold butter gets you most of the way, though not all of it.

Next time Cook the pasta in less water than feels right. The less water, the starchier it is, and starchy water is the thing doing the binding.

The polenta is lumpy

What happened The cornmeal went into hot liquid faster than it could take up water. The outside of each clump gelatinized on contact and sealed dry meal inside, which is why the lumps will not whisk out.

Right now An immersion blender straight into the pot breaks them up in about thirty seconds, while it is hot and still loose. It reads slightly pastier than polenta that never lumped. It still needs its full time on the heat: coarse cornmeal wants thirty to forty-five minutes at a bare simmer, and the grain inside those lumps has not had it yet.

Next time Whisk the cornmeal into cold liquid and bring the two up together, one cup of coarse cornmeal to four cups of liquid. Lumps cannot form in cold water.

The quinoa tastes bitter and soapy

What happened Saponins, the natural coating on the seed. They taste soapy because they behave like soap, and cooking does not destroy them. Boxes labeled pre-rinsed can still carry enough to taste.

Right now Into a fine-mesh sieve under hot running water, rubbing the grains between your fingers, until the foam stops coming, about a minute. Then dry the surface on a sheet pan in a 300°F oven for ten minutes. That takes off most of what is left, not all of it, and it waterlogs the grain a little on the way. Dress it warm with olive oil, salt and lemon, because fat and salt blunt bitterness better than anything else you have.

Next time Rinse the raw grain thirty to sixty seconds under cold water, rubbing it, until the water runs clear and stops foaming. The foam is the test, not the clock.

Meat and fish

Water on the surface stops browning. Heat past a threshold squeezes water out of the middle. Almost everything here is one of those two.

The steak is gray with no crust

What happened The surface was wet, or the pan was not hot enough, or the pan was crowded. Water has to leave before browning starts, and a wet steak in a moderate pan spends its whole cook boiling.

Right now Get the pan properly hot, dry the steak hard with paper towels, and give it another minute a side. If it is already cooked through, slice it and sear the slices — cut faces brown fast and it reads as deliberate.

Next time Dry surface, hot pan, one steak at a time. Salting an hour ahead and leaving it uncovered in the fridge dries the surface for you and seasons it properly at the same time.

The chicken breast is dry

What happened Past about 165°F the muscle fibers contract and force out the water they were holding. Chicken breast has very little fat to disguise it, so it is the cut where five degrees is visible.

Right now Slice it thin against the grain and put it in something wet — a pan sauce, a dressing, mayonnaise for a sandwich, a bowl of broth. Thin slices in liquid eat moist even when the meat is not.

Next time Start checking early, but keep cooking until the thickest part reaches 165°F before the chicken leaves the heat. Carryover may push it higher during the rest, so the practical fix is earlier checking rather than using the counter to cross the safety minimum. Brining can still buy some margin on juiciness.

The roast is overcooked

What happened Carryover. A large roast keeps climbing three to twelve degrees after it leaves the oven, and the bigger it is the more it climbs.

Right now Carve it thin, on a warm platter, and spoon over the juices and a good pan sauce. Do not carve it thick — thickness is what makes overcooked meat feel overcooked in the mouth.

Next time Take it out three to twelve degrees below the number you actually want to finish at, never below the safety minimum you are cooking to. The Roast Clock does the arithmetic per cut and per weight, because carryover scales with mass and nothing prints it.

The roast is underdone and people are sitting down

What happened A cold start, a heavier bird or cut than the recipe assumed, or an oven running cool. It is the most common timing failure at a holiday table.

Right now Carve it into thick slices or sections and finish them in a hot oven or a hot pan for five to ten minutes. Surface area is time — a carved roast finishes in a fraction of what a whole one needs, and this is the standard restaurant answer rather than an emergency one. Poultry still has to reach 165°F whatever the clock says, and beef and lamb still go to the number you wanted; USDA’s minimum for whole cuts is 145°F with a 3-minute rest.

Next time Build a buffer into the plan and rest longer if it finishes early. A roast holds well for an hour, loosely tented, and holds badly for zero minutes when it is not done.

The fish stuck to the pan

What happened It was moved too early. Fish releases itself when the crust forms; until then, the protein is bonded to the metal.

Right now Leave it. Genuinely — another thirty to sixty seconds and it will loosen on its own. If it is properly welded, slide a thin metal spatula under it and accept that some skin stays behind; skin left in the pan is a garnish once it crisps.

Next time Hot pan, dry fish, enough fat, skin side down, and do not touch it for three minutes. Press it flat for the first twenty seconds so it does not curl away from the heat.

The fish fell apart

What happened Overcooked, or turned too often, or it is a flaky species being asked to do a firm species’ job. Cod flakes; halibut holds.

Right now Stop trying to keep it whole. Flake it deliberately over a grain bowl, into a chowder, or into fish cakes. Broken fish plated as broken fish looks like a choice.

Next time One turn. Many cooks pull most white fish at 125 to 130°F for texture, because it keeps cooking on the plate; that is a preference, and USDA’s safe minimum for fish is 145°F.

The meatballs or burgers came out tough and bouncy

What happened The mixture was overworked. Handling ground meat develops the proteins into a springy network, which is what a sausage wants and what a burger does not.

Right now If they are in sauce, more time in the sauce will soften them. If they are grilled, slice and serve them in something wet.

Next time Mix with your fingertips until it just comes together and stop. A panade — bread soaked in milk — physically blocks the proteins from tightening and is the single biggest improvement most meatball recipes are missing.

The pork chop is tough and dry

What happened Modern pork is bred lean, and lean means the window is narrow. The old advice to cook it through is a leftover from a parasite risk that has not applied for decades.

Right now Slice thin against the grain, into a pan sauce or a bowl of noodles. Cold, thinly sliced, it makes a much better sandwich than it does a dinner.

Next time 145°F and a rest, which leaves it faintly pink and is the USDA number. Brine a lean chop for an hour if you have it.

The burgers domed into balls on the grill

What happened The proteins at the surface and the rim contract as they heat, so the patty pulls inward and the meat in the middle has nowhere to go but up.

Right now Turn them and cook the domed side down for a minute or two. One firm press with a wide spatula, while the center is still soft, will level it, and this is the single moment in a burger’s life where pressing is worth what it costs in juice. Then take the temperature at the top of the dome, now the thickest part: ground beef is done at 160°F. It will not sit as flat as a burger that never domed, and the bun forgives the rest.

Next time Press a dimple ½ inch deep into the center of each raw patty, and make the patty ½ inch wider than the bun. The dimple fills in as the edges contract.

The chicken skin did not crisp

What happened Skin is about forty percent water and nearly as much fat, and it cannot brown until that water has left and the fat beneath has rendered. For as long as water is evaporating, the surface is held at 212°F.

Right now Blot the skin hard with paper towels and put it under the broiler, about six inches from the element, for two to four minutes. Watch it, because it goes from pale to black quickly. Skin that has been sitting in sauce is not coming back while it is attached to the meat; pull it off in sheets, crisp it flat on a sheet pan and serve it broken over the top.

Next time Salt the bird and leave it uncovered in the refrigerator overnight, about 1.6 teaspoons of Diamond Crystal per pound, 1 teaspoon of Morton kosher or ¾ teaspoon of table salt, which is the 1 percent the rest of this site dry-brines at, with 1 tablespoon of baking powder per 4 pounds mixed in. That makes the skin blister as well as brown; baking powder, not baking soda.

The bacon curled and cooked unevenly

What happened A strip is alternating bands of lean and fat. The lean contracts sharply as it heats while the fat renders and gives way, so the two halves change size at different rates and the strip buckles.

Right now Turn the heat down to medium-low and flatten the strips with a spatula. Warm bacon relaxes; the same pressure on cold bacon does nothing. If the pan itself is the problem, move everything to a sheet pan in a 400°F oven. A strip that has crisped in a curl will not straighten, but cut in half it lies flat enough for a sandwich.

Next time Start bacon in a cold pan over medium-low. Adding a few tablespoons of water sounds wrong and is the whole trick: it holds the surface at 212°F while the fat renders, and by the time the water has boiled away the strips are limp and flat.

The sausages split open

What happened The heat was too high. Water and fat inside turned to steam faster than the casing could stretch, and the casing tore. What ran out was what would have made them juicy.

Right now Drop the heat to medium-low and finish them gently. The split ones will not close, so make the split the point: lay the cut face down for a minute to brown it, or slice them on the diagonal and serve them in gravy or a pot of beans. Cook them to 160°F for pork or beef sausage, 165°F for chicken or turkey. Nothing puts back what leaked out, so the sauce is doing the covering.

Next time Twelve to fifteen minutes over medium-low beats eight over high every time. Do not prick them; the hole does not release pressure, it only lets the fat out early.

The fish stuck to the grill and tore

What happened The fish went on too early, or moved too early. Fish protein bonds directly to hot bare metal and only lets go once a crust has formed underneath, and grates make that worse than a pan does, because the fillet’s whole weight rests on a few narrow lines of contact.

Right now Stop pulling. Slide a thin metal spatula flat along the grate and under the fish, and take what comes; more force only means a bigger tear. If it is already in pieces, lay a sheet of foil on the grill, move the pieces onto it and finish them there. Serve it flaked, where broken fish reads as a decision.

Next time Preheat fully, scrub the grates while they are hot, and oil the fish rather than the grates. Oil on bare grates burns off before you walk back with the platter; oil on the fish is still there when the fish lands.

The breast is done and the thigh is still raw

What happened Two different muscles were asked to finish at the same moment. Breast is lean and dries out past about 165°F, while thigh only goes tender near 175°F and sits deeper in the bird.

Right now Separate them. Carve the breasts off whole, tent them on a warm platter, and put the legs and thighs back into a 425°F oven until the thickest part of the thigh, away from the bone, reads at least 165°F and closer to 175°F for texture. That is another ten to twenty minutes. Do not serve poultry short of 165°F because the rest of the bird is ready; carved breast under foil holds its heat well.

Next time Spatchcock the bird: backbone out, pressed flat, breast and thigh the same distance from the heat. A 4-pound bird takes about fifty minutes at 425°F that way, against the hour and five minutes The Roast Clock gives for the same bird left whole.

Braises, soups and stews

The long-cooked things are the most forgiving on this page. Almost nothing here is unrecoverable, and the usual fix is more time.

The stew meat is still tough after two hours

What happened It is not done. Collagen turns to gelatin somewhere between 160 and 180°F and it takes hours, and there is a stage in the middle where the meat is tighter than it was raw.

Right now Keep going. Another hour, lid on, at a bare simmer. If it has been going four hours and is still tough, the cut was lean — a stew made from a lean cut never gets there, because there was no collagen to convert.

Next time Chuck, shin, shoulder, cheek. Never a lean roasting cut. And a braise should barely move — a hard boil tightens the muscle faster than the collagen can soften.

The soup is too salty

What happened Salted early and then reduced, or a stock cube on top of salted stock.

Right now Dilute with more unsalted liquid and rebalance. If there is no room in the pot, add starch and volume — potato, rice, beans, bread — which does not remove salt but does put more food against the same amount of it. Acid and fat both blunt what the tongue reads as salt, so a squeeze of lemon or a spoon of cream buys real ground.

Next time Salt at the end of a reduction, never at the start of one. The Salt Calculator gives the percentage by weight, which is the only version of this number that survives a change of pot.

The soup tastes like nothing

What happened Almost always underseasoned rather than under-flavored, and the second most likely cause is no acid.

Right now Salt first, in stages, tasting. Then acid — vinegar, lemon, a splash of wine. Then something savory: soy, fish sauce, miso, parmesan rind, a spoon of tomato paste cooked in fat. Then fat, which carries flavor to the tongue. In that order, and stop when it is right rather than doing all four.

Next time Brown the aromatics properly at the start. Most bland soup is a soup whose onions were sweated for four minutes instead of fifteen.

It is far too spicy

What happened Capsaicin is fat soluble and it does not cook off. Time will not help.

Right now Fat and dairy: coconut milk, cream, yogurt, sour cream, cheese. They physically carry the capsaicin off the receptors, which is why a raita works and a glass of water does not. Sugar and acid both help at the edges. More volume of everything else is the reliable fix if there is room in the pot.

Next time Add heat in layers with tasting between them, and remember that heat builds in a pot over an hour — a curry that is right at the start is too hot at the end.

The braise is greasy

What happened Nothing went wrong. Fat rendered out of the meat, which is the point, and now it is sitting on top.

Right now Skim with a ladle, or lay paper towels flat on the surface and lift them off. If there is time, chill it — the fat sets in a sheet and lifts off in one piece, and a braise is better the next day anyway.

Next time Make it a day ahead on purpose. This is the one dish where the reheated version is unambiguously the better one.

The beans are still hard after hours

What happened Old beans, hard water, or acid added early. Acid and calcium both stop the skins softening, which is why beans cooked in tomato can go all afternoon and stay stubborn.

Right now A pinch of baking soda — an eighth of a teaspoon per cup of beans — and another half hour. It raises the pH and works, though too much makes them slippery and soapy, so use a pinch and not a spoon.

Next time Salt from the start, which contrary to the old advice tenderizes the skins. Tomato, vinegar and wine go in once the beans are already soft. And beans older than a couple of years may simply never soften.

The stew is watery and tastes of nothing

What happened The meat went into the pot without browning, so there is no Maillard layer under the sauce, and there was too much liquid, so what flavor there is has been spread thin.

Right now Lift the solids out with a slotted spoon and keep them warm. Boil the liquid hard and uncovered until it has reduced by about a third, ten to twenty minutes, which does more than any addition. Now salt, a teaspoon of Diamond Crystal per quart at a time, tasting between additions. Last, off the heat, 2 teaspoons of red wine vinegar. Salt and acid account for most of what people mean by tastes of nothing.

Next time Brown the meat dry, in batches with space around every piece; it takes twenty minutes, not five. Braising liquid should come only halfway up the meat, which gives up plenty of its own.

The braised meat is tender but dry and stringy

What happened It went past tender. The collagen melted, then the muscle fibers kept shrinking and squeezed out the water they were holding, so the meat shreds instead of yielding.

Right now Take it off the heat; more time makes this worse, which is the opposite of what a tough braise needs. Pull or slice it, return it to the liquid, and leave it off the heat twenty minutes to take some back. Reduce the sauce until it coats a spoon, whisk in a tablespoon of butter, and serve it wetter than looks right. Stringy meat in a glossy sauce reads as moist; in thin broth it does not.

Next time Braise in a 300°F oven with the liquid at a bare shiver, and start checking an hour early. Cool it in its liquid rather than on a board, since it reabsorbs liquid on the way down.

The soup curdled when the cream went in

What happened Acid and heat broke the dairy: the casein clumped into visible grains, usually because the base was acidic with tomato or wine, or because it boiled after the cream went in.

Right now Off the heat. An immersion blender brings back most of these — run it a full minute, then strain through a fine sieve, which takes the curds below the size the tongue reads. If it is still grainy, whisk a tablespoon of cornstarch into ¼ cup cold cream, stir it in, and rewarm gently, well short of a simmer. Soup split into curds and clear liquid will not go silky again, though blended it is a decent rustic soup.

Next time Temper the cream by ladling hot soup into it, then return it all and hold below a simmer. Heavy cream curdles far less than milk or half-and-half; the fat shields the protein, so the richer is safer.

The stock came out cloudy

What happened It boiled. Rolling agitation emulsifies fat into the water and lifts fine protein and bone particles into suspension, where they stay.

Right now Cloudy stock tastes like clear stock, so the first question is whether anyone will see it. If clarity matters, chill it, lift off the fat, ladle from the top and stop an inch above the bottom so the sediment stays behind, then strain through damp cheesecloth. That makes most stock respectable. An egg-white raft, one white per quart, does better and costs an hour and some of the body you spent the day building.

Next time Start the bones in cold water and skim hard for the first fifteen minutes, then hold it between 180 and 190°F, where a bubble breaks the surface every second or two. It is the boil, not the hours, that ruins the look of it.

The dumplings are gummy in the middle

What happened Dumplings on a stew cook in steam, and every look under the lid drops the temperature and stalls the tops. Batter stirred past the point where the flour disappears develops gluten, and gluten sets as paste.

Right now Lid back on, heat up until the stew simmers steadily rather than barely moving, and give it five to eight minutes without lifting the lid again. A finished dumpling is about 200°F at the center. Undercooked comes back; overmixed does not, because gluten that has set gummy only tightens with more cooking, and the most you can do is split them, spoon the stew over and let them be biscuits.

Next time Stir until no dry flour is visible and then stop, usually ten or twelve strokes, which always feels too soon. Drop the batter in 2-tablespoon mounds onto pieces of meat that stand above the liquid.

Sugar and chocolate

The least forgiving section on this page. Sugar and chocolate both have failure states that are genuinely final, and both are said so here.

The caramel crystallized into a gritty mass

What happened One sugar crystal — off the side of the pan, off a spoon, out of the air — landed in the syrup and the whole thing seeded around it. That is what crystallization is, and it happens in about a second.

Right now Add a little water, put it back on the heat and let it dissolve completely, then start the boil again without stirring. It works, and it costs you the time rather than the sugar.

Next time Do not stir once it boils — swirl the pan. Wash the sides down with a wet brush. A spoonful of corn syrup or a squeeze of lemon in the pot interferes with crystal formation and makes the whole thing far more forgiving.

The caramel burned

What happened Sugar goes from amber to black in about fifteen seconds at the end, and the pan carries on cooking after the heat is off.

Right now It is gone. Burned caramel is bitter all the way through and no amount of cream or salt covers it, and adding cream to it only makes a larger quantity of something bitter. Fill the pan with water and boil it to clean it, and start again — the ingredient cost is pennies and the failed attempt taught you where your stove’s threshold is.

Next time Have the cream or the butter measured and standing beside the pan before the sugar goes on. Pull it a shade lighter than you want and let the pan finish it.

The chocolate seized into a stiff paste

What happened A tiny amount of water got into it. Melted chocolate is a suspension of solids in fat, and a drop of water makes the sugar and cocoa particles clump together and stick. A wet spoon or steam from underneath does it.

Right now Counterintuitively, add more liquid, not less: a tablespoon of boiling water or warm cream at a time, stirred hard, until it comes smooth. It will not go back to being couverture for dipping or tempering, but it is now a perfectly good sauce, ganache or frosting base.

Next time Everything bone dry. Melt in short bursts and stop while there are still lumps — the residual heat finishes it, and overheated chocolate is the other way this goes wrong.

The chocolate set dull, streaky or with a white film

What happened It was out of temper, so the cocoa butter set in the wrong crystal form. The white film is fat or sugar that has migrated to the surface. It is a texture and appearance problem, not spoilage.

Right now It is perfectly good to eat and to bake with. If it needs to look right, melt and temper it again — the chocolate has not been damaged and can be re-tempered indefinitely.

Next time Temper properly, or use compound chocolate where appearance matters more than flavor. Store it away from heat swings, which is what causes bloom on a bar that was fine last week.

The ganache split into oil and grit

What happened The emulsion broke — cream too hot, chocolate too cold, or the ratio was off. Same family as a broken butter sauce.

Right now Warm it gently to about 100°F and blend with an immersion blender, or whisk in a tablespoon of warm milk at a time until it comes back. It nearly always comes back and it takes a minute.

Next time Pour hot cream over chopped chocolate, wait two minutes without touching it, then stir from the middle outward in small circles until the center emulsifies before you widen.

The icing is too runny

What happened Either there is more liquid than the sugar can carry, or the fat is too warm. Butter holds a frosting up only in a narrow band around 65 to 68°F.

Right now Put a hand on the bowl. If it is warm, refrigerate it for 20 to 30 minutes, then whip it for thirty seconds; that alone fixes most runny buttercream, and sugar added to warm frosting only makes a sweeter puddle. If the bowl is cold and the icing is still thin, it is a ratio problem: sift in confectioners’ sugar 2 tablespoons at a time, or beat in 2 to 3 tablespoons of cool butter if it is sweet enough already. Cream cheese frosting beaten to liquid firms in the refrigerator but never comes all the way back.

Next time Take the butter out an hour ahead and use it at 65 to 68°F, cool to the touch and dented by a finger. Butter that feels pleasantly soft is already too warm.

The meringue will not whip

What happened Fat, nine times in ten: a trace of yolk or an oily film in a plastic bowl. Fat crowds the egg proteins off the surface of each air bubble, and a speck is enough.

Right now If fat got in, nothing you add brings that bowl back; start again. Use metal or glass rather than plastic, wipe the bowl and the whisk with vinegar and dry them, and crack each white into a small cup first, so one broken yolk costs one egg. If the whites are clean and merely sluggish, let them come to about 70°F, add ⅛ teaspoon of cream of tartar per white, and keep going; the acid helps the proteins unfold. Hold the sugar until the whites take a soft peak.

Next time Separate the eggs straight out of the refrigerator, when the yolk membrane is firm, then let the whites stand for half an hour. Cold eggs separate cleanly and warm whites whip fast.

The toffee turned grainy

What happened The sugar crystallized. Boiling syrup is supersaturated, and a single crystal off the side of the pan seeds the whole batch in seconds. It is not burnt, only set in the wrong form.

Right now Break it up, return it to the pan with ¼ cup of water per cup of sugar in the recipe, and warm it over low heat, stirring, until the liquid runs clear and no grit drags on the spoon. One crystal left behind starts it over. Wash the sides down with a wet pastry brush, stir in 1 tablespoon of corn syrup, then cook back up to hard crack, 300 to 310°F, and take it to the top of that if you want it to snap. The second batch runs a shade darker and smaller.

Next time Toffee cannot be protected by leaving the pan alone the way a plain sugar caramel can, because the butter has to be stirred to stay emulsified. Use 1 tablespoon of corn syrup per cup of sugar, and a lid on for the first two minutes so condensing steam rinses the sides for you.

The whipped cream turned to butter

What happened Whipped cream is air held in a mesh of partly clumped fat globules; the beating crushed them together until the water squeezed out, leaving fat in a pool of buttermilk. That is churning, and it does not run backward.

Right now This cream is not coming back, so finish the job on purpose. Beat it another minute until the fat gathers into one lump, pour off the buttermilk and keep it for pancakes, then knead the butter in a bowl of ice water, changing the water until it runs clear. You have made butter, which is worth having.

Next time Cream whips best below 40°F, so chill the bowl and the beaters, and stop the mixer while the peaks still slump: the window between soft peaks and grainy is about fifteen seconds. Grainy is the last stage that can still be saved, with 2 tablespoons of cold unwhipped cream folded in by hand.

Vegetables

Vegetables are mostly water, and the whole game is deciding where you want that water to go.

The roasted vegetables steamed instead of browning

What happened Crowded tray, cool oven, or wet vegetables. Browning cannot start until the surface is above the boiling point of water, and a tray packed edge to edge traps steam that keeps it below.

Right now Split them over two trays, turn the oven up to 450°F, and give them ten more minutes. They will still get there. The broiler for the last two minutes finishes it if the oven cannot.

Next time Preheat the tray, dry the vegetables, more oil than feels right, and space between the pieces. Two trays beat one every single time.

The greens are bitter

What happened That is what they are. Brassicas and chicories are bitter by design, and long cooking concentrates it rather than removing it.

Right now Fat, salt, acid and a little sweetness, in that order — olive oil, a real pinch of salt, lemon or vinegar, then a touch of honey if it still needs it. Bitterness is not removed, it is balanced against, and all four of those do it.

Next time Blanch the very bitter ones in salted water first and discard that water. Or cook them faster and hotter, so char replaces bitterness with something people read as sweetness.

The garlic burned

What happened Garlic has almost no water and a lot of sugar, so it goes from gold to black in seconds. It also usually happened because it went into the pan first.

Right now Start over with the garlic. Burned garlic is bitter and it flavors everything it touches, so a dish built on it will taste of it all the way through. If it is only the pieces that are dark, fish them out and taste the fat — if the fat is bitter, that has to go too.

Next time Garlic goes in late, or into cold oil that comes up to heat slowly. Sliced burns slower than minced, and a whole crushed clove slower still.

The onions still are not caramelized and it has been ten minutes

What happened Nothing. Caramelized onions take forty-five minutes and every recipe that says ten is wrong. This is probably the most widely repeated lie in recipe writing.

Right now If there is time, keep going on medium-low with a pinch of salt and a splash of water whenever the pan goes dry. If there is not, accept softened onions — they are a good ingredient with a different name — and add a teaspoon of balsamic or a pinch of sugar to buy some of the sweetness back.

Next time Budget the forty-five minutes, or make a large batch and freeze it in tablespoons. Caramelized onions are the single best thing to have in the freezer.

The mashed potatoes turned gluey

What happened A potato is cells packed with starch granules. Mashing that goes too far, in a processor or a mixer or with a hard hand, ruptures them and spills free starch that hydrates into paste.

Right now It does not come back as mash; more cream and butter make looser glue, not lighter mash. Spread it in a buttered dish, comb the top with a fork, dot it with butter and bake at 425°F for 20 to 25 minutes: sticky potato holds a shape and browns, which is the one thing it is now good at. The other honest exit is potato cakes, beaten with an egg and a spoonful of flour and fried in butter.

Next time Cook them in large pieces, put them through a ricer and never near a blade, and add the butter before the milk. What the starch cannot reach it cannot turn into paste.

The roast potatoes will not crisp

What happened Crowded pan, wet potatoes or cold fat, and the surface sits in its own steam at 212°F instead of drying out, which is below the temperature where browning begins.

Right now Split them between two pans so there is bare metal around every piece, add 2 to 3 tablespoons of fat per pan, and give them 15 to 20 minutes at 450°F. If they are cooked through and still pale, press each one until it cracks open, spoon hot fat over the broken faces and give them 3 minutes under the broiler. Ones that sat covered in a warm dish have gone soft through and only ever come back partway.

Next time Parboil until the edges crumble, about 10 minutes at a full boil, then shake them hard in the dry pot until the outsides look fuzzy. That fuzz is the crust, and the fat should already be hot in the oven when the potatoes meet it.

The green beans went army green

What happened Heat drives the magnesium out of the chlorophyll and hydrogen takes its place. What is left is pheophytin, which is olive drab, and the reaction runs one way only.

Right now The color does not come back, but the beans are usually fine, so take them somewhere drab reads as deliberate. Drain them, dry them, and finish them hot in a skillet with brown butter and toasted almonds. Squeeze the lemon over the platter at the table rather than into the pan, since acid speeds up the reaction that got you here.

Next time Blanch in at least a gallon of hard-boiling water per pound, salted with a tablespoon of Diamond Crystal per quart, three to four minutes, then into ice water. Leave the lid off: a covered pot rains the beans’ own volatile acids straight back into the water.

The dressed salad wilted before the guests sat down

What happened Salt and acid pull water out through the leaf’s cell membranes; the cells lose pressure, the leaf goes limp, and that water is the puddle in the bowl. Tender leaves have about five minutes; sturdier ones fifteen or twenty.

Right now Pour the liquid off and lift the driest top leaves into a dry bowl. Rebuild the crunch instead of trying to restore it: radish, shaved fennel, toasted nuts and croutons, tossed with a spoonful of plain oil and nothing else, since the bowl already holds all the salt and vinegar it was given. Ice water revives a leaf that wilted from drying out and does nothing for one that wilted from salt.

Next time Toss the leaves in the oil alone first, then add the vinegar and the salt. The oil slows the salt reaching the leaf, which buys fifteen minutes that a shaken jar of dressing does not.

The mushrooms are rubbery and pale

What happened A mushroom is about ninety percent water. In a crowded or cool pan it gives that water up faster than the pan can evaporate it and simmers at 212°F, where browning cannot start and the proteins tighten into that squeak.

Right now This one comes back. Turn the heat to high, spread them over two pans, and let the liquid boil away completely. The moment the pan goes dry the temperature climbs past 212°F and they brown in three to five minutes. Do not add fat until the pan is dry; oil sitting in that liquid makes a greasy broth and speeds nothing up.

Next time Start them in a hot dry pan with no fat, and salt them only after they brown. Washing is not the culprit it is said to be: a rinsed mushroom picks up very little water next to what it already holds, and crowding does the damage.

Drinks

A cocktail is three variables and one of them is water. Most bad drinks are the water.

The cocktail is too sweet, or too sour

What happened The ratio drifted, usually because a lime was small or a syrup was heavier than the last one.

Right now Correct in quarter-ounce steps, tasting between each. Too sweet takes citrus, or a dash of bitters, or a pinch of salt — salt suppresses sweetness measurably and is the most underused correction behind a bar. Too sour takes syrup, a quarter ounce at a time. Fix it before adding more spirit, because more spirit only makes a bigger drink with the same problem.

Next time Measure. A sour is two parts spirit, one part citrus, three quarters syrup, and every part of that shifts with the fruit.

The batched cocktail tastes flat and watery

What happened Either it was batched at full strength and then shaken as well, so it took its dilution twice, or it was batched with no water at all and served over melting ice, which does the same thing more slowly.

Right now If it is watery, add spirit and a little citrus rather than only spirit. If it is flat but strong, it needs its acid and its bitters, not its sugar.

Next time A batched stirred drink wants 22 percent added water and a batched shaken one wants 28, because shaking is the more violent of the two. Take about six points off either number if it is going over fresh ice, since the dilution carries on in the glass. Then it is poured, not shaken again. The Bar Calculator does that arithmetic and the ice separately, because ice in the glass and ice around the bottle are two different purchases.

The coffee is bitter, or thin and sour

What happened Bitter is over-extraction: ground too fine, brewed too long, or water too hot. Thin and sour is the opposite — under-extraction, ground too coarse or brewed too fast.

Right now Neither is fixable in the cup, but both tell you exactly what to change. Milk and sugar mask bitterness; nothing masks sour.

Next time Change one variable at a time, and change the grind before anything else. Water off the boil for thirty seconds — about 200°F — and the same weight of coffee every time.

The cocktail will not foam

What happened The foam is egg white or aquafaba protein unfolding around air bubbles. Ice in the tin on the first shake leaves the white too cold to stretch, and any trace of fat, from yolk or an oily film, collapses the bubble walls as fast as they form.

Right now Strain the drink off the ice back into the empty tin and shake it dry, no ice, as hard as you can for 15 seconds. Then add ice, shake another 15 seconds to chill and dilute, and double strain into the glass. If it still will not hold there is fat in the tin: wash it in hot soapy water, add a fresh ½ fl oz of white and shake again, accepting a drink more dilute than you meant.

Next time Dry shake first, every time, and use ½ fl oz of white per drink rather than a whole one, which is more protein than the drink needs and most of what people smell when they call a sour eggy.

There is not enough ice and people are still arriving

What happened Ice is a rate problem, not a storage one. A home ice maker makes two or three pounds a day and a tray takes four hours, so nothing started now arrives in time.

Right now Send someone for two ten pound bags; until then spend what is left on chilling, not serving. Fill a sink with the ice, add cold water and a few handfuls of salt, and stand bottles in it; salted ice water goes below 32°F and chills a bottle in eight to ten minutes, faster than the freezer. Stir batched drinks and strain into small glasses rather than shaking each one to order. The Bar Calculator gives pounds by headcount.

Next time Buy six ounces of ice for every drink poured over it and a pound for every bottle you need cold, which lands near 2 pounds a head, plus a cheap cooler. Most of it goes to holding bottles rather than into glasses, which is why an absurd-sounding amount still runs out.

The wine is warm and there is no time

What happened Warmth pushes alcohol into the air ahead of everything else, so a red at 72°F smells hot and tastes flat. Reds want 60 to 65°F and whites 45 to 50°F.

Right now Ice water, not the freezer. Fill a bucket with half ice and half water, throw in a handful of salt so the bath drops below 32°F, and stand the bottle in it up to the neck. Water pulls heat out many times faster than air: eight to ten minutes with salt in the bath, twelve to fifteen without. An ice cube in a glass of warm red is not a crime; the dilution costs less than the heat does.

Next time Take the red out of the refrigerator half an hour before pouring and the white out fifteen minutes before. Most reds reach the table too warm and most whites too cold; coming at both from the cold side is the only way to land in between.

Frying

Frying is a race between water leaving the food and oil moving in. Temperature is what decides which one wins, and almost every failure here is the oil being cooler than the cook thinks.

The oil temperature crashed when the food went in

What happened The food took the heat out of the oil. Warming a cold batch and boiling the water off it costs a great deal of energy: a pound of fries into 2 quarts of oil pulls it down 50 to 100°F, and below about 300°F the crust never sets.

Right now Turn the burner to high the moment it happens and lift out half the batch if the pot is crowded, since less food in the oil means it climbs back faster. Anything that came out pale and limp goes back in for a minute once the oil reads 375°F again. Turn the burner down as the oil recovers, or it will overshoot.

Next time Preheat to 375°F even when the recipe asks for 350°F; the drop is expected and you are budgeting for it. Depth beats width: 3 inches of oil in a heavy pot holds temperature far better than 1 inch in a wide one.

The batter slid off in the oil

What happened There was water between the food and the batter. A damp surface flashes to steam the second it hits the fat and lifts the coating away before it can set.

Right now Lift the loose shreds out with a spider before they darken and turn the whole pot bitter. For everything still waiting, change the method: dry each piece hard with paper towels, dust it in flour or cornstarch and shake off all but a thin film, then batter and fry, because that film is what the batter grips. The bare pieces will not take a new coat, so salt them and serve them with a sharp dipping sauce.

Next time The invisible flour dusting under the batter is the whole trick, and it works better the less of it there is. Batter thick enough to coat the back of a spoon and stay there is thick enough.

The fried food is greasy rather than crisp

What happened The oil was too cool. At 350 to 375°F the water at the surface boils outward hard enough to keep oil from moving in; below about 300°F that push is weak, the crust never dries, and the food drinks.

Right now Bring the oil to 375°F and fry it a second time for 1 to 2 minutes. This works reliably, even on food that has been out of the oil for half an hour. If the pot is already put away, a 400°F oven on a wire rack for 5 to 10 minutes does most of the same job. Salt it as it comes out, while the surface is still tacky enough to hold the grains.

Next time Most of the oil goes in on the way out rather than in the pan: as fried food cools, the steam inside condenses, the pressure drops, and it pulls surface oil into the crust. That is why it drains in a single layer on a rack, and why a stack on a paper-lined plate is the greasiest finish.

The breading fell off the cutlets

What happened The flour never turned into glue. Breading holds because flour hydrates against a damp surface and egg sets into it; a wet cutlet or an early turn sheds the coat in one sheet.

Right now Stop turning them: a cutlet wants three minutes face down untouched, and most breading comes off in the first minute of fidgeting. Lift the loose crumbs out before they scorch and turn everything after them bitter, and check the fat depth: breading needs at least ¼ inch to fry in. Coating that has come off will not go back on: fry the bare cutlet, toast those crumbs in butter with lemon zest and scatter them over the top.

Next time Bread them, then leave them on a rack in the refrigerator 15 to 30 minutes before frying; the coat sets into something adhesive. Shake nearly all the flour back off first, because a heavy flour layer is a slip plane, not a foundation.

The oil is foaming and everything tastes of fish

What happened The oil has broken down. Water and food particles have split some of the fat into free fatty acids, which act as surfactants, lowering the surface tension so the steam bubbles stop breaking and the foam stands.

Right now The oil is finished, and straining does not bring it back, because what you can taste is in solution rather than in suspension. Strain it warm, label it, keep it for fish alone, and fry the rest of tonight’s food in fresh oil. Food already cooked in it tastes of what it was cooked in, and a vinegary sauce covers some of that but not all.

Next time Fry fish last in a session, never first, and store the strained oil covered in the dark, where it is generally good for six to eight fries. Foam that rises while food cooks is ordinary steam; foam that stands on empty oil is the signal to change it.

Reheating and leftovers

Reheating is a second cooking, and it is the one nobody budgets for. Everything here is about arriving at hot without arriving at cooked twice.

The sauce split when it was reheated

What happened It got too hot on the way back up. Heat makes the suspended fat droplets in a cream or butter sauce merge, while the milk proteins tighten and squeeze out what they were holding.

Right now Off the heat first. Put a tablespoon of cream or water in a clean warm pan and whisk the broken sauce into it a spoonful at a time, building a new emulsion around the fresh liquid rather than repairing the old one. It comes back most of the time, and an immersion blender does it in ten seconds. If it is grainy as well as separated, the proteins have curdled: whisk in a teaspoon of cornstarch slurried in cold water.

Next time Reheat sauces over the lowest heat you have, stirring the whole time, and take them off the moment they reach 165°F, the USDA figure for reheated leftovers; slow heat and constant stirring are what keep a cream sauce whole on the way there. A cream sauce built with a teaspoon of cornstarch per cup reheats without breaking.

The reheated chicken is dry and stringy

What happened It is being cooked a second time. Fibers that already contracted and shed water the first time contract further on the way back up, and a microwave on full power makes it worse by boiling the outer half while the center is still cool.

Right now Water does not go back into meat, so change what surrounds it instead. Slice it thin across the grain and bring it up to temperature in liquid: broth, gravy, tomato sauce, anything with body. Leftovers need to reach 165°F to be safe, and in sauce they get there without the surface drying out on the way. A knob of butter in the pan does the rest.

Next time Reheat covered, with a spoonful of liquid in the dish, in a 300°F oven or at 50 percent power. Slow to 165°F comes out wetter than fast to 165°F, because the outside spends less time far above the target while the middle catches up.

The pasta bake is brown on top and cold in the middle

What happened The oven was too hot for the depth of the dish. A 9 by 13 pan straight out of the refrigerator starts at around 40°F in the center, and the surface reaches browning temperature in twenty minutes with the middle nowhere near done.

Right now Cover it tightly with foil, drop the oven to 325°F, and put it back for 20 to 30 minutes. Under foil the top cannot darken further, so the only question is the center, and a thermometer answers it: 165°F. A dark brown top will taste fine; a genuinely black one is bitter and worth scraping off before it goes to the table.

Next time Covered for most of the bake, uncovered for the last 15 minutes. A refrigerator-cold 9 by 13 wants roughly an hour at 350°F covered, and an hour on the counter beforehand does more for the center than raising the oven ever will.

The rice from the fridge is hard and separate

What happened Starch retrogradation: the starch that gelled during cooking recrystallizes as it cools and pushes the water back out, the same thing that stales bread.

Right now Add 1 to 2 tablespoons of water per cup, cover it, and heat until it is steaming all the way through; two minutes covered in the microwave or a covered pan over low heat both work, and it comes back almost completely. One caution worth knowing: cooked rice carries spores that survive cooking and produce a toxin at room temperature that reheating does not destroy, so rice refrigerated within a couple of hours is fine and rice that sat out overnight goes in the trash.

Next time Spread it in a shallow layer to cool, then refrigerate it within the hour. Hard, dried-out grains are exactly what fried rice wants, so cooking rice a day ahead is a technique rather than a compromise.

The fried food went soft in the refrigerator

What happened A crust is a dry porous shell sitting on wet food, and water moves toward dry. Overnight, moisture from the interior migrates out into the crust, and steam trapped under a lid condenses straight back onto it.

Right now Dry heat, uncovered, on a wire rack: 400°F in the oven or an air fryer for 5 to 10 minutes, or back into 375°F oil for about a minute. Both work on fried chicken, fries, croquettes and cutlets. Thin tempura and delicate batters come back only partway. The microwave is the one method that guarantees it stays soft, because it heats the water first and pushes it into the crust.

Next time Cool it on a rack, uncovered, before it goes into the refrigerator, and store it loosely covered rather than sealed. Wrapping fried food while it is still hot is what softens it: the steam has nowhere to go and condenses under the crust.

Where this comes from

Every number on this page is the same number the calculators here run on. There is one canon and it lives on The Numbers — pull temperatures, salt percentages, hold times, pan volumes, bar dilution. If a figure here ever disagrees with one there, the one there is right and this page is wrong.

This page is for after the mistake. The Substitutions is the one for the moment before it, when the thing you are missing is an ingredient rather than a rescue: 79 of them, what each is actually doing in the recipe, and what you give up by swapping. Four of those have no substitute either.

Most of what goes wrong in a kitchen is not a technique problem, it is an arithmetic problem that arrived too late to fix. That is what the rest of the tools are for: the Roast Clock for the temperature to take it out at, the Salt Calculator for how much salt in the box you actually own, the Pan Calculator for the swap that ruins the cake, and the Oven Traffic Calculator for the dish that was never going to make it.

See all the tools

Where this goes next

You have the number. Now use it.

Put the number to work

The next tool

When the free answer is not enough