Double Spend Attacks on Exchanges


#22

It’s important to protect the ecosystem. This is real threat. We don’t want anyone lose money.


#23

Thanks for the explanation!
A little addition, the estimated attacking hashrate may be as high as 175MH:
https://www.coinotron.com/coinotron/app?action=ChartNoLogon&span=0&type=C&name=BTG

A side effect is the mining difficulty right after the attack is very high so the rest of the miners have hard time finding a block after the attacker leaves.

Miners, it’s important to keep mining so the cost of the attack remains as high as possible.

update 4 hours later:
and right after the difficulty fell, they are back at mining full speed. So basically they are milking the BTG network out of the mining rewards and leave the rest of the miners to mine the hardest blocks. It’s a vicious circle.
The worse part is, this may no longer be considered an attack (unless they still orphan some blocks, but it looks like they stopped) and it may just be a consequence of being a smaller coin amongst coins sharing one algorithm. PoW change may indeed solve this but only if BTG is using an Equihash variation that noone else uses.


#24

The Coinotron pool likely represents miners chasing the “highest profit” coin at any moment, not a malicious attempt.

Yes, it is definitely harmful to steady miners and to the ecosystem.

This is true.

Also true. The PoW change means we will leave the “big pond” of Equihash power for the “small pond” of EquihashDifferent power - a pond where we are the biggest fish, so the risk will be lower (until others choose to use the same EquihashDifferent power that we do, which will bring more water into our pond.)

Also of note, our upcoming fork includes improvements to the DAA - Difficulty Adjustment Algorithm. If our Difficulty adjusts more rapidly to the incoming surge of hashpower, the participants in the surge will earn fewer “excess” coins, limiting their profit. This will also make malicious attacks more expensive. This will help protect the steady miners and smooth out the flow of blocks.


#25

As of May 24th, it is still correct that the last known attack was on May 19th.


#26

h4x3rotab and I were discussing this and believe the current difficulty algorithm verses the new one will not affect the attack. A faulty algorithm could assist the attack, but the current and next difficulty do not have the type of problem I have in mind. But the new DA will be of some benefit if he was or would be able to pick an optimum time to begin the attack, during an oscillation that should not be there in the new DA.

If he’s able to lower the difficulty during selfish mining by something like timestamp manipulation, it might help the precision of deciding when to take it public and beat the public chain, but it does not directly help him. He has to do more chain work which is difficulty x blocks, not number of blocks. If he cuts difficulty to 1/2, he has to mine 2x more blocks to beat the public chain. He’s got 2x more blocks, but he’s not getting those blocks cheaply as h4x3rotab explained. It’s the double spends he gains on.

The public chain has to get 50 blocks before the first spends are confirmed, so the private chain getting more blocks does not help that aspect.


#27

I see; our algorithm is not a significant factor.

So it’s not about a flaw being exploited. It’s about having:

a) lots of power to make 51% attacks and
b) lots of money so their double-spends can trick exchanges into giving them lots of other coins

Got it.


#28

I have been following BTG since its creation, invested in, running full node, mining and I am a strong believer in decentralization but this is the first time for me to try and participate :slight_smile:

What is the feasibility of using a hybrid PoW algo combining or PoW&PoS for example 2 ASIC resistant algos together and specify one for odd blocks and the other for even blocks, this way the attacker mining private chain to perform 51% attack will need more resources in two different algorithms to produce longer chain than the rest of the network. dose anyone here thinks this might help secure BTG?