>The genius of SpaceX's approach is that they built a landing system where it's OK to be unreliable for a while. For booster landings it's probably okay to be

>The genius of SpaceX's approach is that they built a landing system where it's OK to be unreliable for a while. For booster landings it's probably okay to be unreliable forever as long as the failure rate is estimated properly and the value of recovered boosters exceeds the cost of developing and maintaining recovery capability. Since there are no people or cargo involved it's a straight exercise in accounting.