(this piece, slightly modified is also available here)
Good games: board games, card games, tower defense video games, action/adventure, strategy and puzzle games, and casual games all have one thing in common, the successful and seamless integration of strategy, tactics, and in-game interaction.
Strategy when done right is derived from the story of the game combined with your interaction in the game. A good story helps you to define how you make short-term tactical choices for long-term success. A good story without the ability to change the intended outcome, or to fail, becomes filler and fails to motivate. Good games draw you in, make you a part of the story, make you care about the outcome, and make your choices affect that outcome. For kids, at least, few games do this better than Disney’s Epic Mickey, a game where your interactions – your tactical choices – determine the storyline itself. Knowing that you have multiple paths to victory isn’t just good game design, it’s good story craft that rewards players who can be flexible without losing focus and consistent without being rigid.
Tactical choices in the game, like in life, change the way you handle your next situation. A good balance of short and long-term goals enable players to embark on experimental courses of action with little long-term impact. It makes you have to “learn the game”. Where the best games succeed is in leveraging this creative tension between short-and-long term goals into a series of skills that you pick up along the way to help you in your next phase.
In-game interaction is not something I think most tower defense games have excelled at in the past. The action stops too frequently and there are breaks where there should be continuity to keep players engaged. Good in-game interaction is about knowing what information you need to know before the action starts and then acting on that information in a clear and useful manner. Casual games are actually a good model here – you play casual games because you know all the information that you need to before you start what makes this interesting for us is integrating this idea into a game where the “needed” information may not be clear. Again this gets back to beneficial in-game experimentation.
When all of these elements work together you get a great game, but when they struggle so do you.