Obviously there is no magic number to dictate play time. All that's important is whether or not the game has fully explored its game mechanics and world. Replayability is a different creature, and relies on deep design goals. Not all games need replayability; if a game lacks mechanics geared toward competition or cooperative play, multiplayer mode should be cut. Game critics are particularly harsh on play time, and hold most games responsible for hours and hours of content. In reality, each game is different and should only be held accountable for enough hours to run the course of its mechanics. If you're playing a game to fill time, you should probably just read a book.
John Rose, Codemasters Ltd.
This is a difficult question as different players will answer it different ways. I, as a thirty-ish gamer with plenty of disposable income, will rarely ever complain that a game is too short. I enjoy a nice 12-20 hour gaming experience. I rarely will ever play a game more than 20 hours before moving onto the next game. It takes an outstanding game to keep me playing for longer (e.g. Oblivion, Okami, competitive games) So I am a gamer that quickly consumes numerous games and actually prefers a "short game." Twelve hours is a good measure for me. I think they survival horror genre has made this an accepted time for me. Many of the Silent Hill, RE series, etc. are around this time. Anything less that 12 hours I find to be too short. That being said, Gears of War was too short. It was an outstanding game though and after I beat it, I immediately starting playing the game through at a harder difficulty.
Brad Merritt, Cartoon Network
One of the more interesting approaches to addressing the play-length vs. quality issue is to realize that a big portion of the cost involved in making a high-quality game is the technology and effort to get started. Creating a fully realized, boxed play-experience with a new engine and a high amount of polish is a tremendous effort which almost always results in either lower quality then what could be possible or shorter playtime. To capitalize on the investment we've recently seen companies adopting the episode formats prevalent in Half-Life episodes and Telltale Games' Sam & Max. This makes it possible to keep using the same production methods and technology in multiple releases, at high quality, with short play-times and without "cheating" the consumer.
I would imagine that the financial plan for something like Sam & Max doesn't necessarily count on recouping the investment for the engine and initial production on the first episode, but rather spread the budget out over a number of them. The consumer gets a well-crafted, price-worthy experience from the get-go and is confident that all-in-all the playtime will add up to something akin to the open-ended games such as Oblivion. As a bonus this model forces the developer to not rely on hype but rather on quality to sell their game, as the consumer can make an informed choice as to whether he/she wants to buy the next episode, based on how good the previous one was.
Fredrik Thylander, Digital Illusions
Personally, I do not understand how people with real jobs and lives would prefer 30-40 hour single player games. Given the degree of innovation and inspiration of the titles released nowadays, 10-15 hours of gameplay should be enough to spend with almost any game. As a consumer, I quite dislike the 60-dollar minimum price point. I would strongly prefer to try out the game mechanics for a smaller amount of money and pay more if I like the experience. This model works well in the casual space and in a more bothersome way with the aftermarket of used console games. I hope that with the new-gen consoles and downloadable content, consumers will start demanding to pay for a game as they play it.
Mart Lume, Simon Fraser University