Quote from: franky1 on June 12, 2019, 02:37:28 PM

firstly median is just a single number in the middle of a range
1,1,1,1,1,6,7,7,7,7,7
5,5,5,5,5,6,70,70,70,70,70

Even if it's completely irrelevant, because it's an extreme case you're fabricating here - in both cases the median is nearer of the "typical" usage.

Quote

people pay rent once a week but they get a wage once a month. most live paycheque to paycheque and can only plan/budget one month at a time.

Okay, here we have distinct usage patterns in mind. If you want to use LN for a significant part of your wage/salary, then you very likely would transact more than once per week. For me, LN is more comparable with a prepaid card which you would charge once every few months and then use the credit several times a week.

Quote

having one channel to an exchange means one channel is funded with the wage. but what about the other channels.
oh and if you think everyone will just have 1 channel with an exchange. then watch that exchange get DDoS'd and
watch the exchange have to lock in a months value with every user.. bank2.0 centralisation i hear you scream.

The most "sane" way would be to avoid too big hubs.
1-channel-users (which are very likely to be connected with a hub) would be the typical PayPal/Coinbase user. Not those caring about decentralization. The risks of LN and Coinbase/PayPal are similar in these cases.

My guess is that you would have two big classes of "send-only" LN users: the "PayPal-type" users which would be only connected to an hub, and the "prepaid card-type" users, which would fill their LN channel via on-chain transactions to make online payments, but considerably less than once per month.
Then you have a smaller third class, the "sender/receiver" type, for example, freelancers that are spending money and receiving honoraries for their work. This one would be the most useful for the network, but I completely agree that it will be a relatively small class (maybe 10% of the total).
You have also the "receivers-only", which will be also a smaller class.

Quote

oh and need i forget not everyone wants their whole wage in bitcoin. because bitcoin cant buy you everything. so then throwing a whole salary into LN is not helpful

Correct, but where did I write that?

Quote

funny part of your maths there is that to avoid onchain refills each fortnight would actually require factories to re balance channels without onchain usage.
so you cant say using factories after already describing a offchain factory scenario would further 10x up the scenario.

A channel factory rebalancing would not add no transactions nor kB to the blockchain.

Quote

also even if there was a one channel user. with only a one onchain setup and one onchain settle.
the 1LN payment a day would be a ~88 deburdening of the bitcoin network.(90 payments done in ln instead of onchain. but then discounting the 2 setup/close onchain)
so not sure where you are getting the whole other random numbers you quote. especially 25-30million a day

That number was obviously a maximum number if the behavior of the one-channel user with 1 LN tx/day and 1 onchain tx/3 months was the average and the whole block space was used for LN channel openings and closings. So it's actually even more conservative than you estimation (a third of the block space/13 million LNtx per day).

In theory it could be even higher, if the typical user would use it for "micropayments" several times a day.

Quote

but if a LN user has 3-5 channels, the onchain bloat means less than the 1000 a block/144k a day.

That would only be the case if the user refills all his 3-5 channels each 90 days. You could also refill one of your 3-5 channels each 90 days.

Quote

but if a LN user sessions close sooner than 90 days, the onchain bloat means less than 13mill before ln users refresh

Yeah. That's pretty obvious, but then we would have another average "user persona" Wink

Quote

so if you think out of say ~500k transactions a day it some how will equate to 144k users moving over to LN a day.. forget it.
the amount of users desiring LN for daily use would be less.

Also pretty obvious, but that's not the point I was talking about which was a rough estimation about "maximum scalability".
But I, for example, am currently not using LN, but if fees went up again like they did in 2017, I would do that, even if I transact only roughly 5 times per month. Because LN would allow me to open the channel at a weekend or at night, when fees are almost zero, and then transact in the moment I want, even at 12 o cklock a wednesday when average fees even currently are at 100+ sat/byte.

This possibility to avoid fee volatility, combined with a faster confirmation, is actually very likely a major factor which will drive people to LN. Not massively at first, but gradually.

Quote

oh and check visa stats. even when you can buy most things with visa, even they only have their customers do less than 2 tx a day average
so cool down on thinking everyone will jump into LN.
so cool down on thinking everyone will be doing dozens/hundreds of tx a day

Haven't I wrote here in this thread that I guess that 1 tx/day per household is actually the maximum demand I estimate for Bitcoin? It may have been at another LN thread, but I never wrote about hundreds or dozens per day per person.