What's the main argument for ltc?

Development moves faster and with far less bureaucratic hurdles than with Bitcoin. No contentious forks (thus far), they had SegWit first and chances are they will have working, beyond proof-of-concept Lightning first.

Err... Development is actually just rebasing from bitcoin. I've not seen anything go the other way.

In particular, as coauthor of the lightning spec, I assure you that we're all working on bitcoin. Litecoin happened because it was trivial; interestingly, we've deferred Bitcoin mainnet because it's likely to attract real users, who'll risk real money.

Just to affirm what this guy said, these reasons are pretty much why LTC should be on your radar. The heretic in me actually things that LTC could potentially have more value than BTC these days just because they seem to have their shit together development-wise.

The only reason they haven't had contentious forks and there is less bureaucracy is precisely because nobody cares.

Nobody cares about what? Litecoin or the bullshit?

They just successfully completed an atomic swap with bitcoin so yes LTC is defitely innovating too.

Another valuable coin missing from that list is Decred (DCR, https://decred.org/ ).

The on-chain atomic swap (no Lightning Network involved yet) was executed using Decred technology ( https://github.com/decred/atomicswap ) and was preceded by a DCR-LTC swap ( https://blog.decred.org/2017/09/20/On-Chain-Atomic-Swaps/ ).

(De)credit where credit is due. ;-)

More decentralization because you don't essentially need a special purpose FPGA (or whatever is popular these days) specifically meant for mining coins, as these special purpose machines are typically more centralized in larger farms.

I'm not familiar enough with LTC to say if this will continue for the future, but thats the standard argument in favor.

If this is your main reasoning argument for Litecoin, then I have bad news for you, as you're completely wrong. Litecoin's PoW algorithm uses Scrypt and is second only to Bitcoin's SHA256 as far as ASIC-based mining goes. This has been the case for 3 years now and at this point Scrypt ASICs are hashing between 100-200X more efficient both on a $/H and W/H basis vs GPUs.

(Anyone with a passing familiarity w/ CCs would know this, but still, it feels like statements that are so blatantly counterfactual should probably be noted.)