Topic: Calculating Satoshi's coins (Read 1139 times)

I think that we need:

1) improvement of the hashing/computing power
2) new ways of calculating

I don't think that there will be a formula to get from the public key through a simple calculation the private key. But I'm sure that the 256 bit range can be calculated with less than the today's 128 bit (pollard). Let's say we bring it down to 80 bit and at the same time we improve the hashing/computing power, so it could be possible to solve this.

Building the group 

• 1 point addition to have Point (x1, y1)
• 1 subtraction to get y2
• 2 multiplications to get x2 and x3
• comparisions to get lowest x and lowest y

Then you will have all x and y coordinates for all 6 points with the effort of less than 2 point additions, what will increase the speed enormously.

Example: pubkey = 02991eb8eb2e45b4bc9c71bc9a022832e712a8dc1b2db62bd7456e49b2d9f7dac8
This point becomes Point (x1, y1), but we don't know if it is Point 1, 2, 3, 4, 5 or 6.

from our offline server:

(Now we can say that the example point was Point 1, but that is not important.)

Remember:
x1 = x4  and  x2 = x5  and  x3 = x6
y1 = y2 = y3  and  y4 = y5 = y6

Lowest x = x1  or  x = x4
x = 0x991eb8eb2e45b4bc9c71bc9a022832e712a8dc1b2db62bd7456e49b2d9f7dac8

Lowest y = y4  or  y = y5  or  y = y6
y = 0x14c3c6d1a538e95f34bf05f71d9e90b8ba6195e33f0d6dd7c97695e21a09ca63

That Point (x, y) would be the reference point to go on with. From that point you jump to another Point (x1, y1) according to your kangaroo / rho.
It doesn't matter if you jumped to Point 1 or 2 or 3 or 4 or 5 or 6, your reference point would be that Point (x, y) in all cases.

That makes kangaroo / rho faster. For example: A 'tame' that jumps to Point 2 will go on with Point 4. A 'wild' that jumps to Point 5 will also go on with Point 4 and we would have a solution.

But this only works if you have the full Bitcoin range (1 ... n) like in this project and not in a range like the puzzle #120 (2^119 ... 2^120 - 1).

Pollard's kangaroo / lambda / rho accelerator for our project



It leads to inner loops, but all solvable.
Profit: with one point addition, one will cover 6 points.

We want to solve Satoshi's coins, because these coins were mined to serve the purpose, to let the Bitcoin network run. Without these coins, we wouldn't have Bitcoin today, they are special as they had to be mined. Satoshi mined them, marked them and didn't transfer them. Satoshi alone can respond to our project e.g. can transfer them to other addresses. If someone can sign messages to our points, we will be sure that it is Satoshi as they were marked in a way, that you can separate them from others.


That does not mean, that Satoshi isn't owning other coins (not sunken ships) than these marked.

Why are you limiting yourself to Satoshi's coins? There's nothing special about those private keys, if you got a way to crack those, you can attack any key which have had their public key exposed.

'Satoshi's points' what we are calculating are coinbases only and mined in 2009 and 2010.

We have to start somewhere ...

For example with a site like https://lbc.cryptoguru.org/about. They are calculating keys for hashed addresses. (We think that is impossible for Satoshi's addresses.)

One other important thing for our project is, that when all 'Satoshi's points' are calculated, owners of the interaction token would sell all their coins as they would be valueless after solving the last point. The token price would crash before solving the last points. For that problem we have following solution:

For the last 4000 points, the solving participants will get their 25 BTC (12.5 BTC for wild and 12.5 BTC for tame) reward as usual.
But the purchase of the token with the remaining 25 BTC will be different. We will place them all as 0.01 BTC bid orders after solving one point, so that all token holders can sell for 0.01 BTC at least. That means: 4000 points with 25 BTC each = 100,000 BTC and the price wouldn't go below 0.01 BTC in the end as that would cover all 10,000,000 tokens.

The participants will provide the system with distinguished points only and after solving a point, we will publish it and the solving participants will broadcast their reward address, so we can transfer them their reward. We expect that the first solutions will take some time and the intervals will be long enough to make it manually and not automated. Also the purchase of the interaction token will happen manually.

At the beginning ot this project we want to improve the system to solve the points. Once the points will be solved, we will find an automated solution to the rewarding. But we expect that the purchase of the interaction token will remain manually.

How exactly is this supposed to interact with a private key solver program again?

The software would have to be modified to buy or sell tokens as progress is made in exhausting search space. And it's pretty tricky to modify these programs because I once did that, and then someone complained on Github that it wasn't finding keys anymore.

We wanted to have a decentralized exchange for this project and created an own token as ADAB spin-off. We chose the Waves platform.

Exchange: https://waves.exchange/

Explorer with information on the token:
https://wavesexplorer.com/tx/9zMruSw8PPVPvRWgmrDi8QzsaGwNfBqdx9L5sMgASDAK

Name of the token: BitcoinADAB
Quantity: 10,000,000
Decimals: 8
Reissuable: no (10,000,000 tokens is max.)
AssetId: 9zMruSw8PPVPvRWgmrDi8QzsaGwNfBqdx9L5sMgASDAK

To find the token in the exchange:
Switch from 'Verified mode' to 'Community mode'
Search with the AssetId: 9zMruSw8PPVPvRWgmrDi8QzsaGwNfBqdx9L5sMgASDAK
Pairs: BitcoinADAB/BTC and BitcoinADAB/WAVES

This is our interaction token for this project.

What we want is, that each participant of a pool calculates one tame or wild. The main server has to coordinate that all, so it is the best solution so far.

We have another idea. I hope that we can publish it here today or tomorrow.

Well the current Kangaroo builds use a herd of 1024 tame and wild points so I don't think it matters a lot how big you make the herd size on the CPU since the controller code runs through the herd with a simple for loop for CPU. On GPU It's a bit constrained though, hardware is usually optimized to run 128 or 64 items in parallel, and in this case it's 128 pairs at a time. And the GPU kernel runs through this 64 times per call to GPU-land so 8192 different privkeys iterated per call, 8 times more than the CPU on paper.

Of course because of floating-point arithmetic advancements in each new GPU card the actual speed boost is much higher.


The ADAB exchange you keep talking about had a messy scam accusation against them, apparently, they didn't conduct their ICO or whatever properly so I suggest you should look for another exchange in that region to cooperate with.

We don't know how your kangaroo is working. We wanted to say, that if you use the projects system for a point in a range (2^119 - 2^120) and not the whole order range, that you will have a 2^256 problem. So it is not usable for it.

Yes, we use 2^0, 2^1, ... Even if 2^0 and 2^1 are neighboring points, we don't know, from which points these are target points. We are open for improvements, if people explain it.

Pools won't buy the token, if the token wouldn't help them. But they will understand, that it will help and support them. And if a pool participates in that project, that means, that they believe in it. If they don't want to use the token, they could reward their participants with BTC in advance.

It doesn't matter how much ADAB we have, we don't have 100%. And if we started our own token, then we would have 100%.

As mentioned in earlier posts, ADAB wanted to build a crypto exchange and for a short time they had a running exchange but then closed it. We think that they had licensing problems. We know that projects in that region (ME) are high quality and that they have the required equipment. With that project, we invite them to join us. They could become the largest pool in that project.

You should probably study the code before talking on it. The current kangaroo, does "jump over" the order of N and does come back around. Run and print to screen the points and you will see or print them to file.

I still do not think if you jump tames 2^0, 2^1,...anything lower than 2^244ish, they will be of value. So if you are using the standard suggested jumps of 2^128, if those tames don't start in a targeted range above where the private key could be, or even when they wrap back around N, they will be too low to be of use/value.

"Hey pools, we need you to buy this old scammish coin to reward your miners because a. we may never find a collision and b. even though you may have the most hashrate, you may not have half of the solving collision and therefore you get nada."

You must have a lot of ADAB tokens in your possession haha!

In our case, tame points will be random points (private key: 1 - FFFFFFFF FFFFFFFF FFFFFFFF FFFFFFFE BAAEDCE6 AF48A03B BFD25E8C D0364140 = full order range) and wild points (Satoshi's points) are in the same range, so we can use the related wild points as tames after solving one point. Wild points and tame points will jump 2^0, 2^1, ... , 2^255 and they will 'jump over' the order of FFFFFFFF FFFFFFFF FFFFFFFF FFFFFFFE BAAEDCE6 AF48A03B BFD25E8C D0364141 and as it is a finite field, they will come back into the field. That system is not usable for the puzzle, as you have there a small range (2^119 - 2^120) and not the whole order range. Nobody could test this for Bitcoin so far, but it has been tested for a smaller order.

Thank you, that you are trying to improve it.


Edit:
In that system, it doesn't matter how many wild points and tame points we have. The server will provide the pools with starting points, so that the number of wild points and tame points (all tame points will be different random points and can be anywhere in the full order range) are approximately equal. The main problem is the key rate of the whole system. We will need 2^128 (can be more or less, Pollard is not deterministic) calculated points to solve the first point. And that is not impossible but nearly impossible, so we want to try it. We think, that when the project runs, improved point addition hardware will appear. With the provided data from the pools, people will be able to calculate how far we are from calculating the first point of Satoshi. But we will not know, which one it will be.

The hashrate of Bitcoin was in the first year ~10 Mhash/sec and 10 years later 100 Ehash/s. If someone told us in 2009 that in 2019 the network will hash 10,000,000,000,000 times faster, what would we say? 10,000,000,000,000 more participants? No, the hardware and hardware use improved, without Satoshi asking them.

So we have to run our project, even with a low key rate. And when people understand the project, they will join us.

With Satoshi's introduction of Bitcoin in 2009, people could join and mine coins and got coins. But in our case, it will take a while until we have the first point and the first reward. That was a problem to solve. And we found the solution in the token. Mining pools can operate with that token (reward their participants with that token) before we solve the first point. We already said it, the price of the token will be an indicator for the success of the project and the probability when a 'Satoshi's coin' will be solved. At the beginning, the token price will be speculative, but with higher key rates of the network it will become more and more predictable.

It would be much better if the starting points were not kept constant, the fact that the starting points are initially chosen at random anyway implies that there's no benefit to keeping them constant.


The tames have a different min/max range from the wild points so some wild points won't be valid tame points.