The traditional method for maintaining interest in successive sets is to add new mechanics (drawing on new material from the source franchise, of course) that are slightly more powerful than the old ones, while giving old mechanics a handful of new tricks so they can compete. This satisfies the casual players and collectors (for instance, Decipher's Star Wars game would introduce each expansion set around the theme of a planet, like Hoth or Dagobah). The hardcore players are driven along by the excitement of the new and interesting mechanics, as well as the promise of upgrading their current decks. Of course, over time old strategies will be almost completely replaced with the new, better cards, but to get there you have to keep a regular buying (or trading) schedule. The baseball-card distribution system means that you have to buy virtually the entire set to have enough material to trade for the deck you really want, which makes you more likely to make more decks and buy more cards. It's a tidy little system, until it goes wrong.
The venerable Magic: The Gathering TCG is still the most popular of all, with over 6 million players, and its popularity allows it a unique method of fixing problems: every year (3 sets), the slate of available cards is wiped clean. In tournament play, you can only use new cards. The design team carries over some of the most popular and iconic cards, but only the ones they're sure won't break the game. This allows a fresh infusion of craziness into each set, but turns casual play into a race between admittedly broken decks. Unfortunately, it creates a huge rift between casual and tournament players: unless you want to abandon your favorite deck on a regular basis, you can't be a pro, sorry. For less popular games, this would outright kill the scene.
Some companies simply throw up their hands and forget about balancing. For a game based on a cultural fad, the casual market is the real source of money, so the more special Johnny can feel, the more he'll want to buy, and he won't care when he finds out the game is trash. Pokemon and Yugioh do this, although Yugioh surprisingly turned out to have a hardcore following, so the designers were forced to create extensive lists of banned or heavily restricted cards. I don't expect either to outlast the competition.
Decipher, however, tends to balance its TCGs as if they were RPGs, not making any cards technically useless but providing a sliding scale of power that keeps going up until the game drowns under its own weight. Let's examine how each of those played out.
The Star Trek TCG was riddled by poor design from the start: a lengthy pre-game setup phase, counterintuitive battle mechanics, virtually nothing recognizable in the common set to induce casual players, and rares so powerful they essentially created alternate win conditions for the game. The solution was not to fix the game, but to expand the rules to be faithful to the shows and movies so they could fit in as many appeal-to-collector-and-rabid-fan cards as possible. The effectiveness of a card often had little to do with anything else. Decipher has tried several times to fix it, including one total rewrite of the rules, but the prognosis is still not good.
The Star Wars TCG, in an attempt to draw in a universe as huge as Star Trek without compounding rules, moved everything to a very abstract level. Instead of counting numbers, players counted cards in different piles; instead of scoring points or dealing damage, players discarded cards from opponent's piles to win the game. Each new mechanic was its own little piece of an infinite gameplay pie. With such a wide focus, however, it was only a matter of time before interest would wane - why should I go out and buy Jabba's minions when my turn-Luke-to-evil deck will work fine whether I buy the set or not? So Decipher took a twofold approach to rekindling interest, and with their trademark "savvy", decided to add a nice hefty power curve as soon as they ran out of movies to draw from, and to add entirely new (insanely powerful) card types to keep the spirit of the game alive. Some disagree that objective and orders cards were the death of Star Wars, essentially dictating how you could play such an open-ended game. But by the time Episode One cards rolled around, the game had become such a clamor of overpowered and unnecessarily complex new mechanics that it died too. The guys who did Magic got the license from Decipher and promptly marketed it to the too-busy-to-understand-Magic market.
Decipher's latest major effort, Lord of the Rings TCG, promised to be different. I was excited that they'd decided to keep everything simple and tightly focused, with a heavy emphasis on common versions of recognizable things (like Gandalf or the Mines of Moria) that did much the same thing as their rare versions, and in some decks were just as good. As the card sets progressed through the trilogy, the power curve seemed rather friendly, making leaps at each movie but allowing room for the trademark cards of every set to be used. New mechanics were generally quite mild, or were introduced as story elements. However, Decipher proved about as strong as Frodo, and when the end was near, gave in to the marketing dark side. The power curve reared its ugly head with the extra-expensive, blatantly overpowered, blatantly "collectors item" Reflections set. Reflections was designed to correct LOTR TCG's own design mistakes (like the uselessness of Dwarves) and to get the players used to a much more freeform game as they prepared to wipe the slate clean again, but instead showed how bad an idea it was to make a set entirely out of super-rares. Having jacked up the power curve sufficiently, Decipher made its second trademark mistake again: trying to redo everything as soon as the curve becomes untenable.
The situation of the Shadows set in LOTR is as if the people behind Magic, finding green cards needed fixing, decided to add a new color, purple, with the same theme as green. To make sure purple wasn't eclipsed, they replaced some of the key mechanics (say, card drawing) in every color with a new mechanic that you could use in every color except green. Tee hee!