There are some games that have a strategy that will win every game against every opponent, or at least have a strategy that can't lose.  Tic-tac-toe is a simple example, but even chess falls into this category, theoretically.  But there are other games where there is no such strategy.  The children's game of "Rock-paper-scisors," for example, has 3 possible "Straategies" for a given round: rock, paper, or scisors.  None of them will beat all other strategies, which adds a cirtain dynamic quality to the game.  It is a game of guessing and psychology and avoiding falling into paterns.

One of the great things about Mage Wars is that it is a dynamic game; there is no one strategy that will win against all others.  Every strategy has a "Dominant strategy" that will beat it consistantly.  This is why flexability and unpredictability are so important; if you only used one strategy, any opoonent worth their salt will adopt a dominant strategy and crush you, but if you can adopt a second strategy that dominates theirs, then you will win unless they can shift gears... and so on, and so on.

The thing is, shifting strategies has a cost.  Often, it involves losing some of the rescources you put into your first strategy.  For instance, a Beast Master might have to let some small creatures die when Mordok's Obelisk hits the table, switching from a swarm strategy (now dominated by the opponent's strategy) to, say, a beatdown strategy (which often dominates a conjuration based strategy).  The Beast Master loses some of the mana and actions he spent summoning those creatures, but the switch makes the opponent's play sub-optimal as well.  

The Beastmaster might also try to directly refute the opponent's play, destroying the Obelisk for example, but that too will have a cost in actions and mana.  Considerations in choosing between these paths boil down to 3 questions:

1) How well can my spellbook put together a dominating strategy?
2) What is the cost of changing to that strategy?
3) What is the cost of directly refuting their strategy?

When putting together a spell book, you should keep these questions in mind.  If you are playing a fairly 1 dimensional book, you force yourself to directly refute whatever they throw at you, no matter the cost, while if you play a book that has too many different strategies you will force yourself to change strategies often, even when that is quite expensive.

I will refer to these as "Overcommiting" and "Undercommiting."

When you overcommit, it means you have invested so many rescources in a single strategy, be it spell book points, mana, or actions, that it is impractical or even impossible to change strategies.  However, when you undercommit, it means that you have not invested enough in a particular strategy to win even against a dominated strategy.

A good spell book will have several strategies that it can commit to, each of which dominates a different opposing strategy.  A good player will try to find the space between overcommiting and undercommiting.
One final note.  One of the things that follows from all this is that it is a disadvantage to be the first to commit to a strategy, because you will also be the first to have to pay the price of "Changing gears."  Thus, just as it is important to put together a spell book that allows multiple strategy options, I feel it is important to play an opening that leaves multiple strategy options.

Now, a question for the community: what do you see as the major strategies in the game, and what are the strategies that dominate them?

My current thought (very rough sketch):
conjurations (temples, mana crystals, Mordok's Obelisk, etc.) are dominated by Beatdown
Beatdown is dominated by Few Big
Few Big is dominated by Control
Control is dominated by swarm
Swarm is dominated by conjurations