I think the deal was there was already a pretty high threshold for risk. I don't know the percentage exactly but the problem was the o-ring thing put it over the threshold which should triggered a a no-go.

For example, you could say "we'll tolerate a 30% chance of loss of life on this launch" but then an engineer comes up and says "an issue we found puts the risk of loss of life at 65%". That crosses the limit and procedure means no launch. What should not happen is "well, we're going anyway" which is what happened with Challenger.