1. Is someone else about to win the game? Reserve their card if possible. 2. Can you buy a card? Buy the highest point value card you can. 3. Will taking two chips win the game? If no, take 3.
Splendor is all about efficiency (Which is why taking 2 chips is baaaad. Reserving is situationally good, but usually bad.) At higher player counts, the most efficient player almost always wins. At 2 players, reserving becomes much more strategic and isn't such a generally terrible action.
Side strategic note: Aim for the nobles with the most overlap with each other.
Then there's the management of liquidity of meeples through the lifecycle of the game. You want to invest a lot, early; but you need to try and get them back towards the endgame, so you can steal fields and place in micro-fields.
There's a surprising amount of subtlety that emerges from the gameplay. In particular, learning how different growth patterns make your resources easier or harder to steal, or worse, block from completion.
All that said, I almost always use an expansion or two (or three) to keep things varied and add depth.
Learning all the tiles, and tile counting, both just seemed too dull and annoying to actually do.
It is a fun game if nobody learns the tiles :)