You go to a bookshelf in a strangers home and you want to test them for communism. You rip the spine off every book, and throw all the pages in a random pile on the floor.
You have a team of workers who each pick up a page at random, and put it through a photocopier a certain number of times, then throw all the paper back into the pile at random.
Then, the workers go through the entire stack and count the occurrences of each unique sentence. This results in a list of every unique sentence in every book on the shelf occurring however many times each page got photocopied.
At the end of the day, they total up the number of times any of the following words occur: "people's", "comrade", or "bourgeois".
If this total exceeds a threshold, then the person has tested positive for communism and should be pushed out of a helicopter.
The amplification factor (or cycle threshold) is the number of photocopies the workers will make.
The more photocopies, the more sensitive the result.
If the word "people's" appears once in one book but a million photocopies are made, it would have the same result as if the entire bookshelf was full of communist propaganda and a thousand photocopies were made. That's how the false positives happen.
Or it could be compared to securely deleting data on a hard drive (in this example using ccleaner), or scanning for deleting files (with Recuva).
If you want to simple delete data from a drive, the head makes one pass over the data, and writes over top it if. If you do a very secure delete, it makes 35 passes over the data and writes X's over it 35 times. One time, and 35 times are not equal due to the amount of inexactness in the head positioning on the disk, sort of like in the old days with tape tracking.
When you run scan of a disk using Deep Scan method, all it's doing is making more passes over the same Volumn-Cylinder-Head-Sector address to see if it can pick up data using multiple passes, where as it would not using one pass.
Here's an analogy to explain PCR:
You go to a bookshelf in a strangers home and you want to test them for communism. You rip the spine off every book, and throw all the pages in a random pile on the floor.
You have a team of workers who each pick up a page at random, and put it through a photocopier a certain number of times, then throw all the paper back into the pile at random.
Then, the workers go through the entire stack and count the occurrences of each unique sentence. This results in a list of every unique sentence in every book on the shelf occurring however many times each page got photocopied.
At the end of the day, they total up the number of times any of the following words occur: "people's", "comrade", or "bourgeois".
If this total exceeds a threshold, then the person has tested positive for communism and should be pushed out of a helicopter.
The amplification factor (or cycle threshold) is the number of photocopies the workers will make.
The more photocopies, the more sensitive the result.
If the word "people's" appears once in one book but a million photocopies are made, it would have the same result as if the entire bookshelf was full of communist propaganda and a thousand photocopies were made. That's how the false positives happen.
ThT is a very sensible analogy.
Or it could be compared to securely deleting data on a hard drive (in this example using ccleaner), or scanning for deleting files (with Recuva).
If you want to simple delete data from a drive, the head makes one pass over the data, and writes over top it if. If you do a very secure delete, it makes 35 passes over the data and writes X's over it 35 times. One time, and 35 times are not equal due to the amount of inexactness in the head positioning on the disk, sort of like in the old days with tape tracking.
When you run scan of a disk using Deep Scan method, all it's doing is making more passes over the same Volumn-Cylinder-Head-Sector address to see if it can pick up data using multiple passes, where as it would not using one pass.
Disclaimer: non-technical description