Even if the sample were random, the numbers could be misleading. It's clear from browsing any category by release date (almost random) that many app store apps

Even if the sample were random, the numbers could be misleading. It's clear from browsing any category by release date (almost random) that many app store apps have no effort put into them at all. Trying to use a sample of mostly junk apps to gauge the chances of success for an app that has had serious effort put into it is just going to give needlessly discouraging numbers. The best bet is probably to look at apps with both a similar quality level and a similar amount of marketing behind them.