Lamplighter is, by many metrics, a pretty efficient ice:
- 2 to rez, and generally 2 to break, although later on Echelon breaks it for 1 if the runner is not tagged. Particularly efficient against Carmen and Revolver which are very popular killers.
- Fairly bad on a face-check, either giving a tag and ending the run, or forcing the runner to pay 3 (and not ending the run).
The trashing text seems like a big drawback, but can be worked around by only using it to protect assets. For example in AU Co.: The Gold Standard in Clones this could be used to protect Bladderwort, Cohort Guidance Program, Phật Gioan Baotixita, and then that text will never come into effect. Similarly it could be used in Poétrï Luxury Brands: All the Rage decks that use Scatter Field to protect remotes.
Compared to Whitespace, it feels similar in that it ends the run if the runner is too poor, but not otherwise, and a rich enough runner doesn't need a breaker to get past it. However, the cutoff for whitespace is 10 credits to run through it, while for lamplighter it's only 3. Another difference is that the runner can use Overclock and Cezve credits to pay the tax, due to the difference in mechanics.
Obviously, it's more attractive in corps that have other ways of tagging the runner, particularly, if they don't clear tags after Oppo Research then they have to break both subs of Lamplighter.
The biggest downside that I see to this ice is, similar to Palisade, it really would rather not be on a central server, although for different reasons. However, this can lead to much worse Turn 1's for the corp, if the only ice in your opening hand is Lamplighter. So, perhaps it's most interesting for corps that really are installing so many threatening assets that it isn't essential to ice HQ on turn 1.
I think it's totally fine to use this on central servers turn 1. Especially to defend HQ before even drawing your first agenda. Yes, it will probably trash itself at some later point, but until then, it provides a cheap and efficient tax. Certainly not the ideal use case, but not that terrible either.
— Krams