When I read some patents, I often don't understand their novelty, especially with CIP and CP applications.
Is it by checking backward citations and compare with present claims?
We were trained to start by reading claims, but claims are often hard to decipher. So, here's a few tips that may help you. 1 and 2 are the easiest, while 5 is probably the most authoritative.
Look at the last 1 or 2 elements in the claim. "Claim 1: Blah ... blah ... blah ... [all very ordinary standard items] WHEREIN [something special, this may be what you want]". A good thing about starting with the claim is that you can see what the examiner agreed was novel and inventive.
Look at the last paragraph of the background section. Many times this will help to focus on what problem the inventors were trying to solve, and provide useful context for understanding the claim.
Look at the lowest-numbered figure that is not labeled "prior art" and find the parts of the description that describe what's in the figure.
Now that you understand the first claim and first figure, you can go back and compare with the other prior art described in the background section.
For US patents numbered above approximately 6,800,000 the File History is available from USPTO Public PAIR. You can download the entire record of communication between Inventor and Examiner as a PDF from the "Image File Wrapper" tab. This will show all the points Examiner raised why the patent application should not be allowed, and also all the arguments and amendments made by the Inventor to get the claims allowed.
Please note, among people working in the patent field, "novelty" and "inventiveness" refer to two distinct things. Roughly speaking "novelty" means what is claimed has not been previously described as a whole. Roughly speaking "inventiveness" means what is claimed was not obvious from the prior art. Many times, claims can pass a novelty test, but fail an inventiveness (obviousness) test.
By the way, I am not a lawyer (IANAL), and this is not legal advice!