What can explain the performance of Initial Coin Offerings?

Master's thesis in Applied FinanceI study features the most successful initial coin offerings (ICOs) have in common and analyze factors which can predict post-ICO token performance. I find that 79% of startups did have a working prototype, minimum viable product or open-source code available before the start of ICO campaign; 84% of ICOs had an active Reddit discussion channel at the moment of token sale and showed on average 93% increase in subscribers over the two weeks before ICO end date; 80% of startups did have core team members who worked together on the project more than one year before ICO event, with the average team age of 2.3 years; 81% of core teams did have previous work experience in startups or venture capital industry and 76% work fulltime on the startup project; 91% of startups did have a fixed number of tokens issued during ICO, i.e. hard cap token sale structure. I further document significant ICO underpricing of 200% average token return at the end of the first trading day with median value of 78%. Moreover, I provide evidence that 3-month subsequent token returns can be strongly predicted by ICO underpricing (1st day token returns) along with fixed token issuance (hard cap) during the ICO