Depends on the intention.

If I'm planning on playing or running the game, I'll read the relevant core book(s) cover to cover. Try to make a basic character, rereading the relevant sections. Reread sections relevant to gameplay (combat, etc). Try to find a way to break the mechanics, if possible. Continue reading various sections until I have at least page numbers memorized, if not the text itself, and ensure that I'm familiar enough with the layout that I have a decent chance of finding something I don't remember off-hand.

If I'm reading for purposes of expanding my RPG horizons, I skim cover to cover, then reread sections I found particularly intriguing.

I skip ahead to the part that the author is most proud of. Usualy it's pretty clear from reading the covers and the intro what that is.
 
Then I scan that topic for sections that are unusualy lenghty or brief. A very long subsection indicates the author is obsessive over the subject, either because it's his favorite subject or because something went horribly wrong in playtesting. A very short subject tends to have errors of omission. Both are good places to look for bad writing.
 
I also like to look at rules for falling since those are a good indicator of if this is a game that simulates life, adventure, or if it's an exercise in numerology. Same applies to modeling illumination.
 
Then, I scan the whole book for madlib designing. Lists of halfbaked banialities like:
 
Fire spell
This spell shoots a ray of fire that does 1d6 damage.
 
Doom spell
This spell shoots a ray of doom that does 1d6 damage.
 
Neurosurgery spell
This spell shoots a ray of neurosurgery that does 1d6 damage.
 
If it passes those tests then I grade the game as worth looking at further. I try to establish the domain of important metrics and then read over the systems from front to back, or if its a game with an unfamiliar theme I read the fluffy bits.

Quote from: Cranewings;415421OWoD Book of Nod.

No.

RPGPundit