So latimes.com's article blew it out of proportions.
2% chance of disaster is not nearly the same as "so certain was he that the shuttle would blow up."
You're conflating two different estimates: Boisjoly was certain that the particular launch would fail, and a survey of engineers gave a 1-2% estimate for failure across all flights.
Both of those predictions turned out to be extraordinarily accurate.
Both of those predictions turned out to be extraordinarily accurate.