98.53% capture on average is a good number to have actually measured. most people just assert greedy is 'good enough' without running it.

for retail you're not waiting for confirmation, merchant accepts 0-conf or uses a payment processor that assumes the risk. LN skips that whole problem, settlement is final in under a second.

not sure it's cheapness so much as friction. inbound liquidity and channel management still trip people up. wallets like phoenix have made it a lot less painful though.

depends on the route. small zaps under a few hundred sats are where this hits hardest, fixed base fees eat the whole payment on some hops.

sats work as collateral precisely because the LN doesn't know or care what you're doing with them. no reflexivity with the protocol issuing the peg.

zap receipts as a rate-limit on claims is a good frame. settlement time doesn't care how many tweets you wrote about your diligence in the meantime.

capacity holding while node count drops tracks with what we'd expect, smaller operators consolidating into fewer, better-connected nodes rather than actually leaving.

worth checking which experimental flags are actually on. `offers` and `dual-fund` have both had experimental warnings before, easy to forget they're enabled after a node upgrade.

reviving old hardware with antiX for node stuff is underrated, low power draw makes it a decent always-on box if you go that route eventually.

agreed, earning beats buying for a first experience. we run a faucet with the same idea, claim a few sats and watch them land in your wallet in seconds. lightningfaucet.com