There is an interesting analogy about micro-transaction-based gameply over at Killed in a Smiling Accident. Melmoth likens micro-transactions to arcade machines that tempt players to 'insert coin/s to continue', not allowing further access to content without ensuring the game gets your money. I find his argument to be compelling, resonating with me. Brian 'Psychochild' Green replies to Melmoth making a good counterpoint, showing that a game relying on micro-transaction payments can be as good value for money as a subscription-based game, but I believe the two arguments are heading in different directions.
I think it is important to realise the changes that have been made to games in general since the early days of arcade gaming, moving to home computers and consoles, and the trends that are continuing. Early arcade games had characters that were fragile and environments that were lethal. It only took one bump with a ghost, knock by a barrel, or misstep off a ledge for your character to die, losing a life and often being put back at the start of the level. This quick loss of lives made good sense for machines that needed to be fed continually with coins for new games, but when the games were translated to home computers and consoles there was less reason to make the gameplay quite so frustrating.
Having bought the game outright, rather than renting time on a public machine, the notion of killing of the player's character quickly and often made less sense. Computer games no longer needed to be designed to have promote many short sessions of the same content, and instead could focus on presenting as much of the available content to the average gamer. Understanding the differences between coin-ops and home gaming led to a sea change. Where main characters were once killed quickly and brutally they became resilient to attacks, withstanding several bullet hits before dropping, swimming and holding their breath instead of drowning instantly. The health bar made all this possible, soon followed by the ability to replenish lost health through pick-ups. There was little point in killing off the character if the player behind them was not motivated to pour more money in to playing.
Being able to continue long games from where you left off underwent a change too. Arcade games offered a strictly time-limited option to continue within a matter of seconds, but home computers introduced the ability to store the game's state and reload that state. The player could either to continue at pre-determined points, which first had to be reached, or had to decide when to save the game manually, introducing its own problems of how often to save, to reduce the risk of future actions. Just as characters became more resilient so did the save game feature, becoming more automatic and regular, and even seeing some games, notably the Monkey Island series, remove the concept of character death altogether.
The divergence from arcade game lethality to home computing entertainment was gradual but had a profound effect on gaming. Understanding that games didn't need to be limited to five-minute slices of action led to new ideas and genres. Of course, the health bar and its relations were also used in coin-ops, where its application gave the player motivation to pump money in to the machine to boost dwindling stamina. And I think this is the crux of Melmoth's concern, that micro-transactions could be a worrying trend backwards. Comparing free-to-play games relying on micro-transactions for revenue to coin-op games seems apt, but is perhaps not as much of a concern as a game that also has a retail price, where development costs can be recovered more readily and thus be a more suitable platform for cosmetic improvements from micro-transactions. Brian Green's argument that gamers can find their own relative comfort levels of payment seems quite reasonable in this respect.
However, it is important to note that Green's perspective is less about 'micro-transactions' and more about paying reasonable sums for small expansion packs, whereas Melmoth seems focussed towards the prospect of many almost-negligible transactions that have the possibility to accumulate in to a significant expenditure over a period of time, in much the same way that ten pence per credit doesn't seem like much money until they are all added up at the end of a gaming month. Being able to pay for content as and when we want to gain permanent access to it could be a real benefit to the casual gamer, but having the true cost of a game masked by many small and individual charges is a legitimate concern, particularly if the gains are effectively temporary.
There is plenty of experience to show that companies rarely make decisions that benefit the consumer over the concerns of the company itself. Coin-ops are still around today, and their nature has evolved to adapt to the market. Understanding that some players are quite skilled and used to be able to play for hours on a single credit, coin-op games now tend more towards time-limited gameplay, offering a definite amount of time instead of a number of lives. And just as coin-ops still favour the games companies it is reasonable to expect companies basing revenue around micro-transactions to be as shrewd.
Companies may begin by offering additional and premium content for micro-, or macro-, transactions, but the payments can also be used in the same way that some coin-ops allow players to gain direct benefits, or extra longevity out of games by inserting coins. I sympathise with Melmoth's concern that micro-transactions will be used for consumable items or temporary gains that either let players twink their characters or almost vital in overcoming the more difficult challenges in the game. The worry is that we'll be find ourselves playing the MMORPG equivalent of Gauntlet, forever having to bolster our health by paying more money, whilst the ranger shoots the food.