Agreed. Some hard numbers, like from the Startup Genome would help, and undermine some of OP's claims - the high failure rate certainly implies substantial risk since you can only roll the dice once or twice, and the constant pivoting of startups suggests minimal value to planning.