Measurement
One shot gives back one ordinary bit, picked at random using the state's chances. Real counts wobble around the perfect chances. The typical wobble shrinks like 1/√N, where N is the number of shots. Measuring also changes the qubit: afterwards it sits in the state you just saw.
What does one measurement give you?
Why care? Measurement is the only way to get an answer out of a quantum computer. If you misread it, you misread everything.
A shot is one full run of a circuit, ending in one measurement. It gives back one ordinary bit string. Nothing more.
Take a qubit in the state α|0⟩ + β|1⟩. Here α and β are its amplitudes, the numbers that set its chances. A shot reads 0 with a chance of α squared. It reads 1 with a chance of β squared. (If the amplitudes are complex, square their sizes.) The shot does not give back α or β.
Run many shots and you get counts, such as 487 zeros and 513 ones. Counts are samples from the perfect chances. Think of tasting soup with a spoon. Each spoonful tells you something about the pot, but it is not the whole pot. In the same way, counts are evidence about the state, not the state itself.
- Why don't 1,000 shots of a 50/50 state give exactly 500 each?
- Run it: sample a balanced stateINTERACTIVE
- Try this: remove the randomnessINTERACTIVE
- What happens to the qubit after measurement?
- What do people often get wrong about measurement?
- What changes on a real device?
You’ve read the opening of chapter 4. Pro unlocks the other 6 sections — plus every chapter of every course, with circuits you can run right on the page. That’s $11.99 a month, about the price of a coffee, or $99.99 a year (save 30%). The first chapter of every course, and the whole math course, stay free.