You can't look at the measurements and conclude anything about the other member of the pair. The key point is, though, that you really only can detect anything strange happening after the fact. That is a bit less than a 1 in a million probability, so it's very likely you have entangled pairs. If you find that your partner measured the opposite value for all 20 measurements, you now are fairly confident that you were measuring entangled pairs since there is only a (1/2)^20 probability of that occurring by chance. You would not notice anything until your partner gave you his/her values. If you measured 20 of them, for example, you'd get some string of + and - values, occurring randomly. To really detect anything strange, you'd need to measure a large number of electron spins. (50% probability of that occurring in fact). You could measure a pair of non-entangled electrons and get opposite results by chance. Even if you communicated your result with your partner, (who would say he/she got the opposite value), that would not really be conclusive evidence that your particles were really entangled. Nothing about that measurement indicates entanglement. For example, if you measure the spin of an electron about some axis, you get either +1/2 or -1/2. If you have one member of an entangled pair and you measure some attribute of it, you would notice absolutely nothing out of the ordinary. However, I don't believe that there is any way for an observer working with one member of an entangled particle pair to know that the other member of the pair has been measured short of actually communicating with the person doing the measuring on the other particle. I must confess that QM is confusing to me, so I may be wrong here.

Ftl superluminal 2.0