It looks like you're a public user, so you not might be able to access this link, but here's a chapter devoted to that.
Link to bonus chapter (scroll down until you see Rule Equivalency)
http://www.manhattanlsat.com/training-center.cfmIn short, I make two passes on the answer choices (you can literally take Filter #1 through all five answers and THEN take Filter #2 through all five ... or you could just apply Filter #2 on any answer choice that initially makes it past Filter #1).
Filter #1 - Answer is wrong because it's TOO RESTRICTIVE
Use previous work to eliminate these answers.
If you have a counterexample to any rule in an answer choice, then eliminate that choice.
If you don't have a counterexample, at least ask yourself, "DID this have to be true before?" If not, eliminate.
Filter #2 - Answer is wrong because it's TOO PERMISSIVE
-
Make sure the answer choice does the work of the Old Rule
-
For each answer choice, ask yourself, "could I BREAK the Old Rule while FOLLOWING the New Rule?".
There are a lot of answer choices that pass filter 1 by describing a previous implication of the game. For example, say that you're replacing a rule that "H is before K and P". This might have started off a whole Relative Ordering tree.
An answer choice might accurately say that "H can only be 1 or 2". That would probably get through Filter 1. It WAS true that H was always one of the first 2. But does that do the work of forcing H before K and P?
Not on its own. If H is 2, then K or P could still be at 1. You could BREAK the old rule while FOLLOWING the new rule.
If an answer choice passes both tests, it's correct!
There are two main trends for correct answers:
TREND 1:
the rule being substituted involves some character who shows up in ANOTHER rule, tightly wrapped up with some third character. The correct answer rewrites the original rule by swapping out the overlapping character for the third character.
For example:
we're replacing M - P
and we also have a rule that says PL
The new rule would say M - L.
TREND 2:
the correct answer uses "inverted inference syntax" to say the same thing.
(ordering game with only 5 spots)
rule being replaced: G has to be 1, 3, or 5
correct answer: G cannot be 2 or 4
(ordering game with J K M L O P)
rule being replaced: J is before K and M (which also means it's before L and O). Instead of saying that J is before K, M, L, and O ....
correct answer: only P can come before J
(in/out game with K L M N O P)
rule being replaced: M --> ~P and ~K and ~L
correct answer: Only N and O can be selected with M
Other correct answers are mean and take a lot more thinking. Be willing to skip these questions and save them for last/never.