Here's one:

You're playing the river heads-up and out of position, and stacks are such that there are two bets left to play for. You should be roughly aware of what the game theory optimal betting regions look like. If you solve this theoretically using a simplified poker model (see Mathematics of Poker or the poker endgame articles by Tom Ferguson), you'll find that it looks like this:

• Check-raise the top of your range
• Bet-call the next tier of hands
• Bet-fold for value with the next tier of hands
• Check-call the next tier of hands
• Check-raise bluff a small slice from the next tier of hands
• Check-fold with the next tier of hands
• Bet-fold as a bluff with some of your very worst hands

This is the solution from the so-called [0,1] uniform poker model. In real poker you have more information to take into consideration, for example the difference in strength between your range and your opponent's, and card removal effects. Still, being aware of what your river betting regions look like in the model game should help you in a qualitative way when playing the river.

For me personally, understanding the concept of bet-folding hands slightly better than the check-calling hands was a revelation.

Perhaps Janda can chime in himself with some advice about what gives most bang for the buck. Both the book and the videos are excellent.