You’re confusing specific low odds of success for “didn't think it was possible.”

The Apollo missions got unbelievably lucky in that none catastrophically failed despite multiple close calls. However, if you’re willing to try multiple times the odds any mission being successful is much better than every mission being successful.

IE: Six missions landed on the moon. If they each had independent 50% odds then six heads is a long way from impossible ~1.6%, but at least 1 head is quite likely ~98.4%. I doubt we would have tried for a 6th mission after 5 failures in a row, but the point is definitions of success matter a great deal here.

Similarly failures improve odds of success in the future because you learn from mistakes and success means the system is functional eliminating some risks.