Quote from: d5000 on June 10, 2019, 07:22:29 PM

Quote from: franky1 on June 07, 2019, 06:49:39 PM

average of 7 channels per active node.[...] so if your funding 7 channels (onchain data minimum 1 in 7 out) then the close sessions needed at the end of 2in 2out X7(14in 14out)[...]
so if someone is only transacting once a week. they are not really benefiting if they are using monthly channels(4 a month)

First, the median value would be more relevant for an analysis, because big hubs may be distorting the picture (the last time I checked the network structure it was a mix of the hub-and-spoke and the decentralized model, and there were definitively some big nodes with hundreds of channels).

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
median is 6. yet using your own analysis of separating users vs hubs.
users only have 1 connections and hubs have 7.
users only have 5 connections and hubs have 70.

making 6 meaningless statistical error. its just some single use case sat amungst more meaningful data
median does not actually provide anything meaningful. as 6 has no relevance to the other numbers. the number 6 has not been derived from calculations nor from analysing other stats. its just a number that exists

when you say hubs have hundreds of channels. who do you think that those hubs are connecting to.....
when you say users have few of channels. who do you think that those users are connecting to.....
hint for each channel there are 2 sides. meaning for every 2 nodes with active channel between them thats a channel.
so adding up the nodes and adding upthe channels and averaging it out is fairer than your 'median'

Quote from: d5000 on June 10, 2019, 07:22:29 PM

I don't think that the average person needs more than 2 or 3 channels, often one alone will be sufficient.

Second, why should you open channels monthly if you only transact once a week? Remember that also average persons can receive money via LN, and that's where the potential lies. People would be most likely directly be connected with an exchange so they can "refill" easily, e.g. with their salary.

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.
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.
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
oh and if people are locking to one channel with an exchange. watch the exchange ramp up their fee, because they only have the exchange as the only route.

think about scenarios. not the one time utopian perfect experience.
EG did you know when they done the LN pizza promotion, there was only a 10% success rate. not due to any pizza or delivery or settling fiat issues. but the channel issues of LN

Quote from: d5000 on June 10, 2019, 07:22:29 PM

Quote

the fee war is not some technology limit.

It is actually if you want to keep the network decentralized and secure

actually 32mb was a technical based limit of networking related to packets. but 1mb is a arbitrary limit of human imposition
fee war is a wild west zero control human decided imposition that solves nothing but used to avoid real solutions.

Quote from: d5000 on June 10, 2019, 07:22:29 PM

Quote from: Wind_FURY on June 10, 2019, 09:01:31 AM

It will scale the Bitcoin Network, but I believe no one knows by how much. 100x? 1000x? Or maybe less?

It depends on "usage patterns", but we can at least do a basic estimation.

Let's say we define an "user persona" that wants to transact one time per day. He normally refills his channel via LN one or two times per month, but each 3 months he needs a "refunding" via an on-chain channel opening transaction.

This would give us approximately x50 scaling (up to roughly about 300-500 tx/s or 25-30 million tx/day): we have about 100 LN transactions each 3 months (3 x 30-31 plus LN refills) and need a new channel opening, whose size approximately doubles a "normal" 1-in-2-out transaction.

With channel factories and taking into account the same "usage pattern", we could reach about ten times more, so we get a potential of hundreds of millions of transactions per day.

However, we have to take into account that we need some space on the blockchain for 1) normal non-LN transactions, and 2) channel closing transactions in the case of scams. I would estimate roughly that this would takes to a capacity of a third to half of the number I wrote above.

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.

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

but lets take a guess and say that a 3rd of a block is LN reserved. as your scenario suggests
average block ~3k tx so call it 1k LN users a block = 144k a day
but because those users are not transacting onchain on day 2 a fresh set of LN users can use the 1k = another 144k a day,
same for day 3-90= ~13mill users a quarter before they repeat

where as the 144k a day transactors if doing 1 tx a day onchain. is just 144k which again is ~90x

so based on 1 tx a day with LN sessions of 3 months using only onchain bytes of lean measure (not that many channels)
LN

but if a LN user has 3-5 channels, the onchain bloat means less than the 1000 a block/144k a day.
but if a LN user sessions close sooner than 90 days, the onchain bloat means less than 13mill before ln users refresh

now. to get to the point of transaction numbers.
UTXO stats do not any large percentage of spends happening daily. most bitcoin users are once a week-every couple years
the percentage of daily spenders is low percentile.
so until / unless LN provides access to merchants where people can buy absolutely everything (toilet roll, car fuel, bread)
dont expect many to take advantage of daily activity.
apart from day traders/mixers/spammers

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.

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