Topic: PPS pools what is proper reserve number for a pps pool? (Read 397 times)

yep  which is why bitmain would need to be the greatest defender of BTC if the attack occurs.

Not liking that very much at all.

Sometimes  an argument over point a being right or wrong can lead to point b or point c.

If this attack occurs and bitmain chooses to support BCH and not BTC  we could see a big shift in this game.

In fact it seems pretty easy to do maybe 90 days of work at it and cost would not be that high.

I don't want to start another thread as this thread was closed more then a month ago. I opened it as a direct request from Dave.

With this I will close the thread and thank Dave for further expanding points B and C

I would also like to thank  Kano for expanding on point C.

point B the attack could be done
point C the  defenders of  would be mostly asic builders/farmers
point D bitmain could switch sides and move to BCH.

I will lock this again and if someone wants to start a thread about these ideas which sprung from Kano and I arguing about Meni's Formula and the  proper application of it when it came to this pool.

Please feel free to do it as it is more of a speculation thread.

While that all being off topic ... ... ... you do realise don't you that most of mining on the large pools will never change away to another pool?
They are mostly either clients locked into mining on those pools, or are the pool owners mining on those pools.

I will leave this open for a bit as I think Dave F.  has discovered a weakness with this idea.

I want to give him some credit for this ideas are welcome.

I do have one idea if this attack was done it would appear bitmain would be best suited to defend btc from this kind of attack.

then the other top asic builders would be next choice to fight this kind of attack.

Taking this to the limit.

6 blocks an hour  * 24 hours in a day  * 12.5 coins per block = 1800 BTC generated per day.

1800 * 5500 (USD value of BTC at the moment) = 9,900,000 USD Lets round that up to 10,000,000.

Let’s give everyone a 10% bonus so 11 Million USD a day in our PPS pool in outgoing BTC.

Let’s do this for the 3rd quarter so 90 days. That is $990 million USD. Just shy of 1 Billion USD out of pocket if this killer pool mines 0 blocks and has to buy all the BTC that they are giving out.

Since it’s a PPS pool that they control, they can point the hash to ANY SHA coin. Kiss every SHA256 coin that is not BTC goodbye. Then they point it back to BTC for 2 or 3 difficulty increases. And then in the end they point the hash rate nowhere. And the block chain comes to a crashing halt. Are you seriously going to mine at CKpool or Slush or P2Pool when it’s going to take forever to find a block because difficulty spiked because with an extra 10% marginal hardware gets turned back on? And just about everyone else is mining here? Or are you going to mine here with the rest and get your extra 10% all day every day?

And then they shut off and go away. How long will it take us to recover?

Who would do this? I don’t know. Do you think if I went to Chase / Bank Of America / HSBC / CitiBank / Western Union / Wells Fargo and said “Hey, each of you cough up $175 million and I’ll set Bitcoin / crypto back 5+ years.” They would ask how or just write me a check? It's also petty cash for just about any government. Even after the divorce settlement Bezos probably still has it in cash in his safe at home.

-Dave

DaveF asked me to open this thread. In a pm


he will be posting his refined idea soon.

I gave this a few merits and with this idea by DaveF  I will close the thread as it is a better reason than any that I gave

This is the best reason not to mine on a pps pool that is over paying.  Better then Phil or Kano

size of pool does matter according to the developer of the reserve formula.

meni says bigger pool = bigger variance risk
..

simple enough not to argue with kano about it anymore and once again i did not mean to say you were attempting to mislead people with application of the formula deliberately .

facts are  very simple here.
 the formula is a base number subject to multiple linear factors.

a 1 ph pool  in the case of a 42,000 ph network  will lose money slower then a 10 ph pool

so at  current difficulty
1   ph ---------    needs to pay out at    0.04148 btc a day
10 ph ---------    needs to pay out at    0.41480 btc a day
100 ph--------    needs to pay out at    4.14800 btc a day
1000 ph-------   needs to pay out at   41.48000 btc a day
10000 ph------   needs to pay out at 414.48000 btc a day

the formula still reads  a reserve of 4000 coins for all of the above

but as meni said in his article the larger the pool the more variance risk the pool operator takes on.

ie he can lose his reserve of coins faster if he has 20 percent of the network.
then is he has 0.01 percent of the network.

as day says  1 hour payout 1 day payout 1 week payout matter

as does full transparency of the pools hash rate.
and of course actual reserve number and who holds it.

Sometimes I don't like the internet due to the difficulty one may have in explaining what one means.
a flat in the USA ALMOST NEVER MEANS AN APARTMENT.
it means your tire has lost its air.

Also english can be ambiguous in other ways.

I would never mine in a pps pool that only takes the x fees as its fees today.

But in 2017  a pps pool doing that was pretty safe from going bankrupt because tx fees were about 1.4 btc  so 1.4/13.9 = about a 10% fee.

Lots of miners out there but take a 1ph guy that has paid up s9i's set at 80 watts a th.

if he has 6 cent power and mines at a tx fee pool only  he earns a about 49.78 usd a day in btc.

this pool charging only tx fee and paying the 12.5 btc every day looks very good to that miner.

so he want to know what chance do I have of the pool tapping out.  he plugs in the formula and gets 4000 coins and says to himself damn no pool op will hold that much and moves on.

A miner that under stands game theory says  if I mine here and get paid 49.78 a day

and if I was mining at viabtc getting only 48.2866 a day  I gain $1.49 a day

if he goes broke since  he pays daily  I lose what I would have got at viabtc 48.2866

so I need to hope that this pool sticks around for 48.2866/1.49 =  32.4 days  and if I do  I win.

the reserve need for a terrible losing streak causing the pool the quit in under 33 days is what?

depends on pool size 1ph or 10ph  size pool he may stick around even with a small reserve.

10,000 ph pool that big reserve is needed.

AS A miner that now is doing multiple joint ventures in solar arrays I can't take excess risk.

I have buysolar myself on one venture
I have buysolar myself a warehouse owner  his son on another venture.
we are soon to do another solar array joint venture.

So pps pools must be what I do.  with most of the hash power .

I won't use a tx fee only pool as I believe it will fail with the current rate a tx fee pool really is 1%

lots of reasons for me to not mine on one and mine on a pool like viabtc.com via pps

as for 5n I mine some private hash on ckpool
I also mine on mmpool using a system design by meni rosenfeld.  but that hash is trivial and is exclusively mine.

As much as I want all of the hash I manage on a pool using 5n I can't and I must decide which pps pool works for my multiple partners.

if you look at viabtc  they are 3% + tx fee  or about 4%  they pay daily so I simply do not fear bankruptcy and just look towards steady payments.

In hard times like now I put some of the profit to regular pools

here I am on mmpool

http://mmpool.org/statistics

Estimated Payout

Rank   Id   DGM Estimate   PPS   Shares   Rate (MHash/s)
1   06ae84f6   3.55057111   0.00000000   1,171,509,380,517   0
2   2d0ba29d   0.51050521   0.00000000   130,638,721,300   117,281
3   a5d78fff   5.73805197   0.00000000   130,161,014,576   0
4   b0d22e5b   0.37382912   0.00000000   121,152,709,055   0
5   c932f3ed   0.37382913   0.00000000   90,817,204,703   0
6   71602a00   0.37382912   0.00000000   45,557,299,863   0
7   c03a6b93   0.09342429   0.00000000   32,332,000,000   0
8   67076efc   0.15285052   0.00000000   26,758,862,431   0
9   a60330ba   0.05375181   0.00000000   19,990,766,359   0
10   49ecce78   0.09328962   0.00000000   17,375,959,894   0
11   eb4a7fcb   0.05036050   0.00000000   15,152,643,537   0
12   d7ec4cdf   0.07419160   0.00000000   14,289,498,257   0
13   dad84f7a   0.07854794   0.00000000   13,424,228,895   0
14   1fa7a7f0   0.06336479   0.00000000   11,609,886,265   0
15   d17ab73b   0.03541869   0.00000000   10,567,830,845   0
16   8f21a6e7   0.05854443   0.00000000   10,125,655,014   0
17   29501a4c   0.01845540   0.00000000   5,369,606,803   31,254,156
18   b14f3c1d   0.01281848   0.00000000   4,414,315,691   0

Some other things to think about.

Although it would not change the reserve requirements, frequency of payouts do matter a bit too. Once an hour vs once a day vs once a week etc.
i.e how much the pool could wind up owing you if they go broke. If I don't trust you and I only get paid once a week, I'm gone....

Cash / fiat reserves. If I want to start a pool to dominate the mining world and I have a big enough bank account I can do just about anything I want. 105% pay out, no big deal if I don't mind taking a bath buying BTC on the open market. Think venture capitalists throwing money into it.

Using the theory of a 10,000 ph paying 415.00 BTC a day at the price of $4k a BTC that is $1,660,000 (using round numbers here for BTC now at $3800 and a 105% payout)

So if the pool *never* finds a block in 30 days and I get $50 million USD of venture capital I can still run for a month. How many pools would I kill in the process who saw their hash rate drop to 0.

Just my thoughts.

-Dave

That's not correct.

The point is about limits on the luck - again nothing to do with the size.
Nothing to do with losing blocks.

It can be a 200PH pool or a 2EH pool.

That is simply a recipe for disaster.
If you want to go down the path of saying that it's OK to have a too small balance thus it's OK for a month coz it's small - well so what?
You are just putting people in the position of losing out in a month.

You need to look at the history of how this has gone before.
As I said in the previous thread, you've not been around for long.
Learn from the mistakes of the past (though most people are too stupid to do that)

You have some bazaar idea that everyone mining on the pool is going to be able to work out the day that it's gone bust and be able to magically guess in advance when they are going to have a run of bad luck.
It wont happen like that.
What will happen is the same as every other time before, some people will move on - by luck or by good guessing, others will hang around and keep mining then finally it will all close down and those who stayed around will lose out.
Yes just like the altcoin scams - so yeah I guess that's what you are hoping for.

Just give up and stop giving people bad advice
(and also pretending you said something else ... you do get that what you said is still there ... no one but a fool believes you when you pretend you said something else)

Lets try to explain the reserve.

4000 to 1000 for a 1% fee  depending how close you are to the ½ ing

and the bigger the pool is  the more variance risk it takes on.

So pretend someone has a pool that the only fee is the tx fees and is giving all of the 12.5 out. this is roughly 1% fee

Suppose they put the 4000 reserve it an escrow handled by coinbase.

What would happen.  Well any miner with an economic understanding of this would move to the pool.  why is that?  zero variance risk  and the 4000 coin reserve is proven.

The pool would grow like mad as the miners simply have to track the size of the pool the luck of the pool to see what happens to the reserve.

If the pool is having good luck no worries mine on.
If the pool is having bad luck the reserve shrinks very quickly.
If all is transparent  miners can time the jump from the sinking ship.
With almost no risk.

Basically it is why I got into the argument with kano about 4000 being not accurate for a 1% pool.

Wish I phrased deceptive to "applied in error".

The truth is no pool paying pps at 1% fee would survive with any proven reserve.

reason is the pool would grow as it would be the best pool to mine at.

it would grow to 30 or 40 % of the network.

and  it would be attacked by anyone with the ability to with hold blocks.

There is a thing I still don't get.
If any of pools declare to have a 4000+ BTC reserve fund (even 1000+ is enough) with a PPS payout - won't this pool get instantly attacked by miners (or even other pools) to get all of these funds?

you are correct. and i did read what I said and it does look like I accused kano of deception.

I did not mean it that way .

 simply meant the applying the formula is more difficult then it appears to be.

I just want to say that it is sad to see you two guys fighting, because you are both great guys and very important contributors in btc and here at btctalk.

yes  it was a poor choice of words on my part.

I now prefer to say  the 4317 number was calculated by the formula given by meni r. correctly

But needs to be correctly applied based on the caveats I listed  in the above posts.

If they are not addressed the number is not helpful for a miner or for a pool op.

Phil  don't you think  you simply  made Kano feel attacked when you used the word deception?

He just went back at you and called you an idiot.

there will be a lot of cut and paste  from paper below


(0.1%) that is my number not meni's

the operator should keep a reserve of at least  

R=B ln (0.1)/2f


I agree with this in principle  as no one want to be waiting for their payment for mining and the pool op has gone broke.

the formula  has

R for reserve
B for block

(0.1) for bankruptcy risk in the case above.

2 as a constant and f as a fee

in this example I will use

12.5 for b
6.9079  as the solve for  ln 0.1

2 for the constant

and 1% as the fee.

so 12.5 x 6.9079  / (2x0.01)

86.34875/0.02 = 4317.4375  is the solve

this is not wrong or right  it is the solve for the formula when you use 12.5 coins a block

and use a 1 % fee  

and  use a 0.1% chance of bankruptcy.

I can clearly show that more is needed then this to determine if my pool goes broke or does not go  broke.

Both miners and pps pool ops could use  this guidance.

As a miner  proof of reserves is more important  then  the exact of amount of reserve needed.

So lets say  you want  to know what number meni's formula give for a 99.9 % chance of payment

in the case of this a pool doing a 1% fee via  grabbing tx fees which average out you get 4317....

that is a starting point.  to further find a correct reserve number you need  to understand

meni's formula uses  B as block size  and the formula does not reflect that B is subject to ½ every 4 years.

So the day after the  latest ½ ing  the B is stable for 4 years
The day before the next ½ ing  the B is stable for only a day.

so the formula when solved  does not reflect that it must be used as part of your decision process.

next is  what is the fee  is it a flat %  or is it a flat % plus tx fees

this is very important to use in your decision making process.

if a pool does tx fees only and if they are about 1%  of the full block + tx fees

you use .01 in the formula
if  you are 4 years away from the ½ ing  the 1%  won't change  for 4 years

if you are 1 day from the ½ ing  the 1% won't be true in a day.

so 12.5 + .125 = 12.625  the .125 tx fees are close to 1%

next ½ ing

6.25 + .125 = 6.375  the .125 tx fees are close to 2%

so that 4317 number  is subject to a change to 1079.25 at next ½ ing

thus the correct number for a pool operator to have if they  do a 1% fee is a range of  from 4317  to 1079  during this current block size  of 12.5

there are more factors for the miner to think  if they should use a pool

next one if pool size.

1ph pps pool
10ph pps pool
100ph pps pool
1000ph pps pool
10000ph pps pool

all pay  different amount of coin  each day

a 1ph pays 0.04148 btc a day

a 10 ph pays 0.4148 btc a day

a 100ph pays 4.148 btc a day

a 1000ph pays 41.48 btc a day

a 10,000 ph pays 414.80 btc a day

meni does address size matters I will find the quote in a minute


So miners must take their time and start with meni's formula then tweak it to get the proper reserve number.

you can see above that a 10 ph pool holding back just the tx fees as their fees is only paying  0.4148 btc
a day  so  if they  do  4317/0.4148 they could pay for 10,407 days  and never have to hit a block for all that time and still not be broke.  In fact much longer  since they would be multiple ½ ings in that time frame as kanos says 1461 days = 4 years  there would be 7 ½ ings    so rewards given out  would  last  more then forever  and that is with out hitting a block ever.

so  to assess what the pools reserve needs to be   to do 99.9 % safe  is not  a simple matter

even if you pick  that large number 4317 to be safe  you could be wrong  because the pool is a big pps pool  say 10,000 ph this pool need to pay 414.8 btc a day   that comes to the 4317 being gone in under 11 days  if the pool never hit a block in that time frame.

So meni's number is the starting point  to determine a correct reserve.
time left before the next ½ in
size of the pool
and of course  what fees  are since if tx fees  they have been at high points some time as much as 2 coins a block of 12.5

once you have  factored this all in as a miner  and let us say you are using a reserve of 250 coins or 140 coins due to pool size  being small.

you need to see real proof that reserve is there and that pool op is will to lose that number of coins.

all of the above applies to established pps pools such as viabtc  or new pools.

and the lower the fee a pool takes the more likely it is to go bankrupt.

to all feel free to have some popcorn.

this is my final post in the other thread


I will get back to this later today.

Self modded as I don't want people to use  terms like you are an idiot and don't have a right to an opinion.

as anyone reading the pools section  they know Kano and I are having a war over his use of meni rosenfeld's formula for proper  reserve for a pps pool.

Kano you and anyone else can feel free to post but not free to call people names.

I really want to educate people about using statistical formula correctly

not solving them correctly.