Objective: To find out what attribute of a product is most valuble to the consumer, and what value for that attribute is most preferred.

E.g. : Of the attributes of a watch, Strap Color, Dial Color, Strap Material, Functionality etc , let's say Strap Material is most important to a certain user. And for this, out of Leather, Metal, Plastic and Rubber, he prefers metal most.

The game:
A roulette listing the attributes is shown to the user who is given two separate currency denominations (fake) and asked to place his money as he prefers.

After he's chosen 2 attribtues, two roulettes come up one after the othhr, listing the attribute options. The user again places bets on each roulette on whatever option he prefers. Essentially here teh data collection is compelte - we should know what attributes are most important to the user and what values for those attributes are preferred.
To finish off the game, we simply declare him a winner if the roulette stops on the option he bet.

Will this work? Out of two denominations in the first roulette, do you think people will place the higher buck on the attribute most preferred? Will they then place the money on the attribute option they want most?

Remember that each attribute has an equal chance of winning onthe roulette and there is no other bias also...