I think the premise of the question is wrong. In business, there's never a case where you're 95% or 99.9% of the way to "success". Success is something that develops over time.

Validating your ideas is the most important part of the journey, and learning how to do it in the context of your particular product or market is what will lead to success. It's also something you get better at over time.

So it that sense it's self-fulfilling; with endurance you get better at validating your ideas, and become more successful.

But some common validators for startups are:

- growth (in traffic, user signups, revenue) - this is considered by far the most important in YC

- user engagement & feedback

- user testing

- a/b testing or multi-variate testing

Whole books have been written about this stuff - most notably by Eric Ries and Steve Blank in recent times. It's probably worth your while to read them.

That is a hard question. Probably the best you can do is ask people who've seen a lot of similar situations.

I'd imagine that you have some sort of spreadsheet to keep track of your predictions, both positive and negative.

I'm curious if your educated conclusion beats a random walk (similar to mutual fund managers for example)

Like all maxims, "by endurance we conquer" is meant only to remind you of some potential solutions. Knowing when to quit and when to stick with it is the hard part.

This is pretty much isomorphic to the Halting Problem. ;)

This sums it up rather well:

http://dilbert.com/strips/comic/2012-09-19/