There is one problem with what you say about its spread at levels that saturate healthcare systems and cause extreme amounts of sickness and death: They may not be the case so much because the virus is particularly deadly or even much deadlier than the flu -which may indeed be the case: that it's only part of an order of magnitude deadlier than H1N1, we don't know yet until we get more concrete infection rate surveys- but there are reasoned arguments supporting this idea.
Instead it saturates in such ugly ways because it's so contagious (as the comment below this mentions). Thus, highly contagious + entirely virgin population of susceptible people (of all types: asymptomatic, lightly symptomatic, moderately, severely, mortally) will inevitably result in very high infection rates and thus an initial saturation that's also high even if hospital-worthy cases are a small percentage of total known and unknown cases.
This saturation may or may not be capable of getting worse than what we've seen so far in NYC, Northern Italy, Spain and Iran to a lesser extent, but though we don't yet know this for sure, if it is the case, the virus being allowed to spread far and fast will likely cause these bad but possibly tolerable saturations to happen quickly only once, and then simply cease as enough people become immune or at least resistant, and healthcare systems steadily adapt.. So far, it seems that Italy and Spain might actually be close to as bad as it gets even with widespread infection of the population, since credible estimates are presupposing at least 2 million cases in total in Italy at this point, for example. Even in New York there are signs of slowdown already. Again, if that is in fact the worst case scenario in any given region, its cost, as tragic as it is, might be worth weighing against the possibly more catastrophic cost of prolonged quarantine and its resulting economic and social consequences.
We could possibly be delaying an inevitable event with only briefly heavy localized consequences and moderate general clinical consequences by artificially creating a much worse long term problem that kills too, albeit more indirectly and among a wider demographic through missed clinical care appointments of other kinds, depressive suicides, economic ruin, substance abuse and other general population harm factors we can only start guessing at so far.