I also think that the UASF is a good idea. Hashrate follows coin price. If
the UASF has the higher coin price, the other chain will be annihilated. If
the UASF has a lower coin price, the user activated chain can still exist
(though their coins can be trivially stolen on the majority chain).

The success of the UASF depends entirely on the price. And actually, the
price is easy to manipulate. If you, as an economically active full node,
refuse to acknowledge the old chain and demand that incoming coins arrive
over the UASF chain. In doing so, you drive down the utility of the old
chain and drive up the utility of the new chain. This ultimately impacts
the price.

I think it would be pretty easy to get high confidence of the success of a
UASF. Basically you need all the major economic hubs to agree to upgrade
and then exclusively accept UASF coins. I don't have a comprehensive list,
but if we could sign on 75% of the major exchanges and payment processors,
and get 75% of the wallets to upgrade, then the UASF would be very likely
to successfully obliterate the old rules, as miners would be unable to sell
their coins or pay their bills by stubbornly sticking to the old chain.
It's less risky than a hard fork by far, because there is zero risk of coin
split if the UASF has majority hashrate, which will follow majority
economic value.

A serious proposal I think would get all the code ready and merged, but
without setting a flag day. Then we would get signatures from the major
institutions promising to use the software and saying that they are ready
for a flag day. After that, you release a patch with a flag day 12 months
in the future. People can upgrade immediately, and have a full year to
transition.

That gives tons of time for people to upgrade, and tons of confidence that
the UASF will end up as the majority chain.

If we cannot get enough major exchanges, payment processors, and other
economic hubs to upgrade,  the flag day should remain upset, as the risk of
coin split will be non-zero.

I would suggest that a carefully executed UASF is much riskier than a soft
fork, but far, far less risky than a hard fork.