The Xbox 360 has been on sale for eight months and I still cannot name five games on it I would tell somebody to buy the console for. That is the problem Microsoft has to solve between now and November, and after E3 I think it probably will.

The launch lineup was weak and there is no honest way around that. Perfect Dark Zero carried an enormous amount of expectation and buckled under it. Kameo was pretty and slight. Call of Duty 2 was the one genuinely good reason to own the machine in December, and shooters that also exist on PC are a thin argument for a five hundred dollar box. Microsoft got the console out a year early and shipped it without the software to justify the hurry, which is a defensible trade only if the software eventually turns up.

Gears of War is the reason to think it does. What was shown in May is the first thing on this hardware that looks like a generational leap rather than a resolution bump, and the cover mechanic is the kind of idea that reorganises a genre. It also matters well beyond itself, because it is Unreal Engine 3 running properly on a console, and a great many studios are currently deciding which engine to license for the next five years.

Announcing Halo 3 was the other move, and it was pure defensive positioning. There is nothing to play, there is barely anything to look at, and it does not need to be either of those things. It needs to exist in the mind of somebody standing in a shop in November wondering whether to buy a PlayStation 3 instead. Microsoft has one franchise that can single-handedly hold a customer through a competitor’s launch window and it has just deployed it as a placeholder.

Xbox Live Arcade is the part I think is genuinely underrated, and it gets discussed as a novelty when it is actually a structural advantage. Small games, low prices, bought in ninety seconds without leaving the sofa, from developers who could never fund a retail release. Geometry Wars has probably been played by more 360 owners than most of the launch retail catalogue. More importantly it establishes the habit of buying software on the console, which is the habit that matters for the next decade.

Sony has nothing comparable and is about to launch a console while building an online service from scratch. Building that service is not a technical problem, it is an operational one, and operational problems take time that a launch window does not provide.

What I would still be nervous about, in Microsoft’s position, is the hardware itself. The failure rate stories have gone from anecdote to pattern over the past few months, and the warranty on offer is not generous enough for a machine with that reputation. A software lineup does not help if the console it runs on has to be posted back. I do not think that story has peaked.

The honest summary of the first eight months is that Microsoft bought a year and spent most of it badly, and then used one trade show to buy back the narrative. Whether that was enough depends entirely on Gears shipping in the state it was demonstrated in, which is not a small assumption. If it does, the head start finally means something. If it slips to spring, Microsoft goes into the most contested Christmas in console history with a thin library, a reliability problem and two brand new competitors on the shelf beside it.