17g to 50g of Grated Cheese: The Three Axes Through Rome's Pastas

The traditional Roman schools build their sauces from cheese, fat, and pasta water alone, with no heavy cream. There are exceptions worth flagging: Tomomi Ogura adds a small amount of heavy cream in his fava-bean version, and Fabio takes the pan off the heat, adds 50 g of cream cheese, and lets residual heat melt it before the Pecorino goes in for his cacio e pepe (see the Fabio cacio e pepe write-up; his own explanation of why this resists separating hasn’t been confirmed in the post text) — both are alternate routes to a stable emulsion. As the editors read it, this region’s “big four pastas” (some groupings count gricia as a fourth; this article sticks to the three whose ingredients overlap most — the gricia breakdown is on the gricia comparison page) become easier to map once you treat them as a product of three variables.

Cacio e pepe’s constraint — Pecorino and black pepper, nothing else — turns the technique for melting the cheese into the whole point of the dish. The original Armando al Pantheon version that Italia Squisita profiles uses 150 g of Pecorino with no added fat, and mantecas it always off the heat (sempre a fuoco spento). Vincenzo’s Plate’s 2026 version, by contrast, cooks the pasta unsalted in a reduced amount of water, builds a Pecorino paste in a separate bowl with pasta water, then tosses it in a saucepan set over the pasta pot, using the steam’s indirect heat (though the same page’s write-up also says the paste goes into the pan before it’s set on top of the pot, so the source itself disagrees on the order of steps). Both agree on one thing — never put the cheese directly over a flame — which lines up with the general cooking-science point that Parmigiano and Pecorino separate easily at high heat.

Pork fat is the axis carbonara and amatriciana share. Guanciale is cured pork jowl — higher in fat than pancetta (cured pork belly) and without the smoke that bacon carries — and the choice of cut and heat changes how heavy the finished sauce feels. (Saltiness varies a lot by brand, so it’s hard to generalize.)

The biggest difference between the three dishes is what triggers the sauce to thicken. Carbonara relies on the egg’s coagulation temperature; cacio e pepe relies on starch concentration in the pasta water. Amatriciana draws on moisture from the tomatoes plus pasta water, and splits into two camps: the Perilli version, Pasta Grammar, and Yoshimi Hidaka all take the pan off the heat before stirring in Pecorino to finish it, while Vincenzo’s Plate saves the cheese for a topping instead. (White wine varies too — from a glass in Vincenzo’s Plate to a tablespoon in Italia Squisita’s Perilli version, while Ryuji’s version skips it and Fabio’s ingredient list doesn’t mention it, based on a partial reading of the video description.) These three triggers are independent phenomena, but they all come back to the same question: where does the heat actually get built. Carbonara’s split between direct heat, bain-marie, and residual heat is covered in Working Backward From Egg Coagulation in Carbonara; the general principles behind emulsification itself are in What Emulsification Actually Is.

For a fuller breakdown of each school, see the carbonara comparison page, the cacio e pepe comparison page, and the amatriciana comparison page. If you want to pick a chef for each step and build your own version of the dish, each page’s recipe builder can do that.

• Cheese clumps: Putting cheese directly over a flame makes it separate easily, which is why Italia Squisita and Vincenzo’s Plate both manteca off the heat or with indirect steam. Adding the pasta water gradually helps.
• Egg overcooks (carbonara): The standard fix is to pull the pan off direct heat and finish with residual warmth or a bain-marie. Tomomi Ogura uses a pan that’s already off the heat; Vincenzo’s Plate uses steam to keep the egg from setting too fast.
• Guanciale burns, or the fat doesn’t render: Vincenzo’s Plate goes low-to-medium and slow; Italia Squisita’s Perilli version sears hard and fast for color. These are opposite strategies — picking based on whether you’re after rendered fat or a seared, aromatic edge is the shortest route to fewer failures.

Carbonara, cacio e pepe, and amatriciana use different ingredients, but they split along the same three axes: how much cheese goes in, how the pork fat is handled, and what triggers the emulsion. Read side by side, Roman cooking looks less like three separate dishes and more like one system competing to control temperature and quantity precisely with a short ingredient list. A good way in is to make one dish exactly as a source specifies, then make a second dish and watch how the same three axes — cheese quantity, pork fat, what starts the emulsion — shift.