TL;DR: NWC vs LN address doesn't really matter; just use what is available to you. It probably looks cleaner in your wallet to use NWC since it can send+recv.

Mhh, it's a bit nuanced:

I wouldn't say NWC is the most reliable since you're using infrastructure that you don't control (nostr relays). However, @Alby runs their own relay, so there is some guarantee that this relay is well maintained. Additionally, this is more a theoretical problem. I haven't had any problems with Primal's relay ever for example.

The advantage of NWC is that you don't need to have a public IP and mess with firewall settings. You can simply use any relay to connect both ends.

However, if you already run your own node and have a public IP so we can reach you directly, you could run LNbits or phoenixd since both work with simple HTTP requests. Websockets are a bit more complicated so more could go wrong which is what nostr is using.

But I think in practice, it doesn't really matter if you use HTTP (LNbits, LN address, phoenixd, CLN), websockets (NWC) or gRPC (LND).

LNC is the only connection that does something really unique. Imo it's over-engineered and you can't even run it in multiple tabs at once so we need to always initiate a connection for every zap and then disconnect; making it pretty slow.