Scientists often want to know exactly how much energy holds a specific molecule together. They can use quantum physics equations to find this out without doing physical experiments. However, the most accurate math equations are incredibly complex and take far too long to solve.
A supercomputer might need months to run one perfect calculation for a single large molecule. In 1991, scientists named Curtiss and Pople invented the G2 method to solve this exact problem. Instead of running one impossibly long calculation, G2 runs a smart sequence of shorter calculations.
It starts with a fast, basic estimate using something called the Hartree-Fock method. Then it runs slightly harder MP2, MP4, and QCISD(T) calculations to capture more electron details. The computer adds these smaller test results together with a final empirical mathematical correction.
This clever trick mimics the result of a massive QCISD(T) calculation in much less time. The final answer usually matches real-world laboratory experiments within one or two kilocalories per mole. This makes near-perfect accuracy possible for small and medium molecules on normal laboratory computers. Chemists later created improved versions called G3 and G4 that run even faster and better.
