At the risk of resurrecting a dead thread, I thought I'd make a comment or two.
First, the question of how you know you are looking at the correct particle.
We have devices that will emit a single particle at a time that then either decays or is transformed into two entangled particles. We can then do any measurements on those particles *before* another particle (and pair) are even produced.
Second, when there are 'trillions' of particles, the entanglement tends to be destroyed by interactions with those other particles.
Third, in quantum mechanics, we predict *probabilities* and not specific events. For entangled particles the probabilities are correlated.
So, for example, suppose I make a particle that decays into a spin up/spin down pair. The two particles in the pair go off in different directions.
Some things are important:
1. We cannot tell ahead of time which direction the spin up particle will go or which direction the spin down particle will go. There is a 50/50 split in the possibilities.
2. Over many such pairs, both sides see 50% spin up and 50% spin down. Neither side can predict what the next particle will be.
3. When the measurements are brought together, it is found that whenever one side detects a spin up, the other side detects a spin down. So the results on the two sides are completely (anti-)correlated.
4. If we do something at either end that affects the spin, the spin of the other end will be found to stay opposite of the changed one.
5. A subtlety: since we don't know ahead of time whether the particle at our end will be up or down, there is no way to use this to send a signal to the other end (they don't know either).
Fourth, the question arose as to how we know measurements affect the results.
The classical example is the double-slit experiment. Send a stream of electrons through a device with two thin slits close together. Detect the electrons on a screen past the slits.
1. If that is all you do, there will be an 'interference pattern' detected on the screen.
2. If, instead, you attempt to do a measurement to determine which slit the electrons go through (say, by shining a light on them), the interference pattern disappears. Instead, we get two 'lumps' of detection.
What happens is that the light that would show which slit the electrons go through interacts with those electrons and that interaction destroys the pattern.
3. If we 'turn down' the light to the place where we can no longer determine which slit the electron goes through, the interference pattern re-emerges.
Fifth, this was described as purely theoretical. This is false. These effects have been observed, tested, re-tested, and some current technology is based on exactly these phenomena. Lasers, for example, depend on entanglement of photons to work.