That’s a very good point on the high view-buy people. There’s definitely some more work to be done! I do feel that the shear force of masses of people contributing to this application will prevent a lot of abuse. The people who administer this system would need to like a bit of a tinker; it sounds like the equivalent people at Google relish it

In terms of excluding on the high view-buy ratio, perhaps we’re thinking the wrong way. These could be the most discerning buyers and the best deciders of what is actually good as they may look at everything and only purchase the best. Perhaps not so easy to model as at first thought….

I was thinking this morning over reheated cheesy-crust pizza about exactly the issue you raised. An anti-trolling tweak to the algorithm could be to exclude people who as users have a bad view-buy ratio, that is they’re unlikely to buy anything anyway, and perhaps only include people who have purchased more than a minimum number of applications.

I don’t think we see direct abuse of the Top apps list; we just see what the masses like to purchase, which is what everyone else purchased. Algorithms based on mass behaviour are going to be harder to influence; I expect that a competitor would not be able to convince enough people to view, but not buy, a competitor’s application without revealing themselves to the community as a cheat. Recent events considered that may not be a reasonable deterent (cf. the blatant Classics ripoff and such like).

I suspecut Apple would need to keep the algorithm under wraps (in the same way Google does its search algorithm) and possibly add some protection from multipleviews, whether from the same IP or within a certain timeframe, depending on whether any abuse occurs.