I perhaps shouldn't have given a fixed time as it is largely dependent on how you have implemented mechanics etc.. 10% of project time seems relatively appropriate but this will vary wildly depending on the title. I stated a "day", as with 4 people all playing the game repeatedly for that period of time, most core issues/bugs could be found and addressed relatively quickly. A nice rule of thumb is the bigger the project, the bigger proportion of your time you should spend on playtesting.