The main issue is world randomness.
You have no idea and irregardless of evidence that you have gathered if something will work or not. Based on this, it would be very bad to work on something for a year without knowing or getting frequent feedback (cash sent to your bank account by customers).
So the only way is to treat this like poker. You place several bets over a period of time (max 1 month). This way you launch 12 projects in a year and have a better success chance of hitting something that actually works. You have to place bets until the card is right, then you can go all-in.
