I'd still submit STS-1 was the riskiest space flight ever attempted. Granted it wasn't leaving LEO, but there was basically no survivable abort option ever, other than a short TAL window (feasibility of such abort will never be known), and nothing beyond high final in the orbiter had ever been legitimately tested outside of a computer. That's a lot of modeling and assumptions to bet your life on. Not to mention that the SSME's had a long history of disintegrating, long before the first flight, or when the SRB's started hinting that they might do the same on STS-51L. That and nobody actually had data to model flight anywhere in the atmosphere above mach 6, let alone 25 IMN, at least in anything other than a capsule/MIRV. Perhaps that data was more applicable, and I think it probably must have been, but still, STS-1 was an uncharacteristically bold flight for NASA to undertake, given it's historical penchant for flying several times unmanned to prove a concept in Mercury/Gemini/Apollo. Add the fact that the STS ultimately rode a failure rate (loss of crew/vehicle) about 10,000 times higher than predicted and sold to the Capitol, and one might start to question the modeling that the first flight was based upon.