That realistic assessment should start by assuming that in the absence of further data you are a typical member of your class. If you have ideas about ways in which you are different to your class, try to validate them with actual data before making judgements based on them, because another feature of typical members of the human race is that they frequently invent spurious reasons to believe they are atypical.
I'd say most people think they have a huge market though. How do you separate out the ones who are right (and thus the special person for whom the statistic didn't apply) from the failure cases (who believed in essentially the same preconditions but were incorrect/implemented wrong/didn't move quick enough/had some extenuating circumstance/didn't get funding/had the wrong partner/hired the wrong dev/didn't pivot/failed).
That's why experience is such a key factor—if you've already done it 3 times you have a lot higher chance of making the right decisions along the way.
I'm not saying it's entirely luck, but it's not deterministic either.
Clearly, there is a grand misunderstanding of statistics here.
In order to make yourself an exception to the statistics, you're going to have to scientifically explain how you are going to become better. Everyone believes they're better. How are you, truly, going to go above and beyond? Do you have better connections than the average? Do you have the experience needed? Do you have a better idea which is downright guaranteed to succeed?
If you can answer Yes to all of those questions, congratulations, you are the average Startup Individual, and the statistics do apply to you. You're going to have to work really hard, have all the right things fall into place, and do everything right, and even then you still have an extremely large chance of failure. It is the nature of the situation.
Starting a startup isn't like playing the lottery. You have maybe 20% more control over your outcome. But 80% is still in the hand you're dealt, both as a person and as a company. Make the best of it.