Thinking aloud here...
There are three types of Wii Sports purchases. Type A is Wii Sports + Wii purchased simultaneously. Type B is Wii Sports purchased AFTER a Wii (someone buys a Wii and some other game like Zelda, finishes it, then comes back to the store and buys Wii Sports). Type C is normally the rarest kind... purchasing a game BEFORE you buy the system. Type C only occurs when a console is supply-constrained.
The attach rate for Type A (simultaneous purchases) probably remains about the same week-for-week, since Wii Sports is the killer app for the system and the system's been supply constrained since launch. There is a pool of people who want a Wii, a fairly constant percentage of them want it for Wii Sports, and a representative portion of that group scores each week. So Type A purchases rise and fall as console sales rise and fall.
Type B (late purchases) are harder to predict but they probably also follow the sales of the console somewhat... just lagging behind a few weeks. In other words, Type B purchases are driven by people who pick Wii Sports as a second of third game rather than first, and the purchase is made after the most preferred games are finished.
Which leaves Type C (early purchases). The number of these purchases are INVERSELY related to console sales ONLY IF THE CONSOLE IS CONSTRAINED. The fewer consoles that are available, the more frustrated purchasers, and the more people who will say "Since I have the cash I'll just go ahead and buy Wii Sports today and get the system when I can find it in a store." They are an artifact of supply constraint, they completely vanish when the system is sitting on the shelf. And I think they are the best way to explain how attach rate for the killer app jumps when the console sales fall.