Every launch increases the risk that something will fail
Actually, that's not true. When a booster comes back, you get to inspect it and see what components are holding up, and what's not doing so well, and fix it. After over a dozen such launches, you get a really good idea what's going on.[1]
When you're launching the first time, you're hoping the assembly crew did a really good job and nothing blows up.
(and I don't want to fly on the booster that ran into a bridge in Texas this weekend)
[1] as long as you don't accept "normalization of deviance" like they did with the Shuttle tiles, not realizing how fragile they were, and not knowing how badly they aged because they never tested aged tiles until after the accident.