Ok, I have a scenario I can't get my head around.
I have a system which captures information about vehicles going past a camera. 96% of vehicles will be captured correctly, so 4% will not be captured correctly.
Any particular vehicle has a 96% chance of being captured correctly.
If I have 3 capture points, the probability of any particular vehicle being MISSED is 0.04x0.04x0.04 = 0.000064 , so the chance of it being read correctly on at least one point is the inverse of that = 99.9936%
Hopefully that reasoning is correct.
What I can't grasp is how to work out the probability of it matching on any two of the three points.
Let's say we take point A to point B - 96% on each point means the probability of matching both is 92% (0.96 x 0.96). And matching at A AND B AND C is 0.96*0.96*0.96 = 88%
But what if I care about matching it at A-B OR A-C - the percentage is obviously higher than 92%, but I can't work out how to...um...work it out. My degree was an awfully long time ago.