ULA's track record with the Atlas V: 100%
Yes, let's take one vehicle in its fifth generation (not counting subrevisions), and ignore its track record with all of its earlier versions that led up to this point and all of their failures, and all of Lockheed and Boeings' other launch vehicles over time, with all of their failures. Lets also ignore that they're going to have to switch engines soon, to an engine with zero track record.
Payloads typically launch on schedule or within a few weeks.
.... Some payloads have been waiting literally years due to delays.
Let's totally ignore that Atlas V launches once per two months, while SpaceX launches once per month, and that almost all of the wait time was due to investigation backlog. When it comes to hitting launch windows, SpaceX has a higher average success rate than average than Atlas V
And lets entirely fail to mention the point that ULA charges nearly double what SpaceX does per kilogram. Or that SpaceX is doing everything while rapidly evolving its rocket, to the point that they've basically even switched propellants partway through (denisification radically changes their properties). And while at the same time running an aggressive recovery and refurbishment programme and developing a heavy lift vehicle, with a small fraction as much capital.
As if liquid boosters can't fail catastrophically? Check out SpaceX's last failure. Liquids are hardly immune to catastrophic failure.
And actually more to the point, you've got it backwards. The SRB failure on Challenger was slow, more like a blowtorch. The explosion was when it compromised the external tank (which, obviously, stored liquids).
Solid propellants aren't like explosives. More to the point, you have to keep them under pressure to get the sort of burn rate that is desired for a rocket.
Could you remind me how many people SpaceX has killed? Boeing and Lockheed have certainly killed people in the past.
If you're referring to the AMOS 6 ground failure, ignoring that part of the whole point of flying a stack unmanned as much as you can before you fly it manned is to shake out any problems, is that a manned mission would have almost certainly survived that. Unless the launch escape system failed, despite the drama, that was an eminently survivable. How do we know this? Because AMOS-6's hypergolic propellant tanks didn't ignite until the satellite hit the ground. AMOS-6 had the fairing as some extra protection, but on the other hand, the satellite itself isn't nearly as durable as a crew dragon.
The launch escape system ignites within milliseconds of a failure being detected and almost immediately reaches full thrust, accelerating away at 10gs. Here's a graphic of Dragon's abort test superimposed over the AMOS-6 failure. Things like this are the very reason that launch escape systems exist. NASA's last manned space vehicle lacked such a system entirely. And while their design for the Shuttle ultimately wasn't chosen, you know what? Lockheed's proposal didn't have one either. And it had a strong impact on influencing the final Shuttle design outcome.
SLS Block 1 is less than 10% higher payload to LEO than Falcon Heavy. Not a particularly meaningful difference. Don't confuse Block 1 with Block 2 (which will probably never fly; the current schedule doesn't call for it until 2029 - and that's not accounting for the current delays).
New Glenn doesn't count as a hedge?
SpaceX and Blue Origin would not use solids, not because there's something wrong with solids per se, but because they're not "fuel and go", which makes them expensive to reuse - and SpaceX and Blue Origin are all about reuse.
A lack of experience with hydrolox surely factors into the picture for SpaceX and Blue Origin; they'd get significantly higher payload fractions by using a hydrolox upper stage. But they're willing to accept lower payloads in order to simplify their manufacture and ground infrastructure, and in particular because the need their propellants to be storable, and storing LH for long periods is a PITA. Storing methalox is quite difficult, but nothing compared to hydrolox.
Solids really aren't that bad when reusability isn't a concern. They're very high thrust, which is exactly what you want out of a booster, and they're structurally very simple. Their low impulse and high structural mass are not particularly important aspects for boosters. Reuse of solids however gains you very little, because there's so much work in refurbishing them.
That's not the reason you don't use it for a first stage. The disadvantages of hydrolox (which are numerous) are offset by its incredible specific impulse. But for a first stage, specific impulse doesn't matter that much, while thrust matters a lot. Thrust is in large part proportional to fuel density, as a turbopump sweeps out a fixed volume per rotation, so the denser the fuel, the more mass (and generally all else being equal, energy) it pumps per rotation.
Another aspect is that first stages are big, meaning that cost is more important than specific impulse. By contrast, when dealing with an upper stage, a small increase in mass has a huge increase in first stage size, and since first stages are so large and expensive, that's a big cost. So you generally want a higher ISP upper stage. With the caveat that "storability" requirements for engines that need to restart can shift the balance; because hydrogen is so deeply cryogenic it's difficult to store for protracted lengths of time. Also, the longer you plan to have a stage in usage without maintenance, the more you tend to favour simple propellants over high performing ones, particularly when you're dealing with small, light engines. So for example if you have an interplanetary probe you'll tend to favour a self-pressurizing hypergolic system so that you only have to rely on a couple valves working, even though self-pressurizing propellant tanks are heavier and hypergolics tend to be lower specific impulse. Engines that are smaller still are often monoprops for an even greater degree of simplicity.
"You need tender loving care once a week - so that I can slap you into shape." - Ellyn Mustard