For those who missed the clarification, here is the thread by Charlie Lee on Merged mining and 51% Attack in Charlie Lee’s Own Words:

Since, so many people have asked me if this is an issue, I figure I should respond to this and clear up any misunderstanding of how merged mining works and how it affects 51% attack. Let me start off by saying that this network security concern is totally unfounded.

First, let me explain how security works in a POW cryptocurrency.  Miners have an upfront cost (cost of hardware like ASICs/GPUs) and an incremental cost (cost of electricity) when mining a coin. To make money, miners have to mine enough coins to cover both costs over time.

On the flip side, we have attackers. In order to 51% attack a coin, they need to have 51% or close to 51% of the network hashrate and they need to mine the coin for some time to build up a 51% attack chain to override the honest chain and reverse some transactions.

How much time the attacker needs to mine depends on how long it takes for their victim to trust a transaction to be confirmed. So if the attacker is going to reverse some transactions to Coinbase for example, they would need to wait for 12 confirmations, which is 30 minutes.

This incremental costs can be calculated by calculating the electricity costs to run 51% of the mining hashrate for 30 minutes. Charlie Lee has also made a reference to an app called “crypto51.app” as a useful site to check that cost. We see that right now, it costs about $15,000 to attack Litecoin for 30 minutes.

But it’s not actually this cheap to attack Litecoin. There’s also the upfront cost that I first mentioned. This upfront cost can be calculated by figuring out how much money it costs to buy all the ASICs/GPUs needed to match the hashrate that is currently on Litecoin network.

Depending on which miner (A6+ or L3++) you can find, total costs is between $133M to $151M. This is of course assuming you are even able to find and buy that many ASIC miners. I’m talking 149,573 A6+ or 515,201 L3++! And you have to set up a mining farm with cheap electricity.

For this attack to be profitable, the attacker will have to make enough money to cover the costs of the attack: $150M upfront + $15k incremental. This is because if he successfully pulls it off, he will destroy the value of his Scrypt miners and won’t be able to recoup costs.

He will also need tens or hundreds of millions of dollars worth of LTC to perform this attack because he has to deposit these LTC to exchanges and double spend them. After a successful attack, LTC price might crash, which makes the attack even less worthwhile.

This attack is theoretically possible but not practically possible. Satoshi designed this system such it makes more sense for miners to be honest and mine the coin normally, rather than attack the coin and hurt their investment and mining income.

There’s also another way to get the hashrate needed to attack. That is to rent the hashrate. This is what the NiceHash-able percent refers to. Litecoin is 6% NiceHash-able. This means you can rent a total of 6% of the Scrypt hashrate. That’s nowhere enough to do a 51% attack.

Whereas, Bitcoin Cash (33% NH-able) and BitcoinSV (40% NH-able) are much easier to 51% attack. If you can rent the hashrate, then there’s no upfront costs and only the incremental costs are needed. That’s why it’s important for a POW coin to be mining algorithm dominant.

Litecoin is Scrypt-dominant and that is one of the reasons why Dogecoin switched to merged mining (AuxPoW) with Litecoin in 2014. Litecoin and Dogecoin miners are combining forces to mine both coins at the same time. So it would cost almost just as much to attack Dogecoin.

So how does the recent increase in Dogecoin marketcap and liquidity affect this 51% attack scenario? The simple answer is that it doesn’t at all. It still costs $100M+ to acquire the hashrate needed to do the 51% attack. That’s still prohibitively expensive.

Merged mining doesn’t make it “easier (and considerably cheaper)” to attack both network at the same time. The upfront costs and incremental costs is not any less.

The incremental costs are actually more! Pools like LitecoinPool are paying out PPS 155%. This means miners are getting 55% more in mining rewards than before because of the extra DOGE they get with merged mining. So incremental costs are $23,250 instead of $15,000.

The only difference with merged mining is that if you have the resources to 51% attack Litecoin, you can also 51% attack Dogecoin at the same time. But what good does that do? It’s not more attractive to attack Litecoin. There’s no higher ROI to attack Litecoin and Dogecoin.

If you can pull off this attack, you can make money by depositing LTC to an exchange and double spending it. Adding the ability to also double spend DOGE at the same time adds nothing. Re-read the original hypothetical attack and remove the DOGE from it. It’s the same attack.