Why Charlie Lee is wrong, and how Verge Multi-Algo is different from DigiBytes

Also known as: The story of why DigiByte is the safest, most decentralized blockchain, and I’ll explain why!

You see, not all multi-algo coins are created equal.

Not all “security” is as secure as others

This started thanks to a tweet from Charlie Lee yesterday:

https://twitter.com/SatoshiLite/status/999027979891044352

5 mining algorithms, when done correctly (my emphasis there), will be exponentially more secure than a single algorithm ever could be.

How does a 51% attack work?

Well the idea is that you have over 51% of the hash power, you can attack the network and effectively ‘rewrite history’.

You can also dominate finding blocks statistically because you have the most hash power.

This is bad, because a blockchain is supposed to be immutable. However if somebody is able to ‘rewrite history’ and after a block has been confirmed they go back and rebuild a longer blockchain showing it was never spent, then that undermines the authenticity and the point of a blockchain.

Whattomine.com profitability for a single L3+ ASIC miner

There’s a LOT of hash power currently on Litecoin. 320TH/s is insane. That’s approx 650,000 L3+ Antminer ASICs (Worth around USD$520 each).

This means to 51% attack Litecoin, you would need to find yourself a around 700,000 L3+ miners, worth a casual 365 million USD.

Compare this to EMC2, where they have < 200GH/s, meaning if you had 400 L3+ miners, you could attack their network and begin to rewrite history! A quarter of a million dollars in Scrypt ASICs and you’re in control of the coin.

I know miners with more than that here in little ol’ New Zealand! Some have that in GPU’s, some have that in Scrypt / SHA256 ASICs. It’s totally doable!

That means that some people in New Zealand with their ASIC farms could basically wipe a coin off the face of the planet if they want to, because nobody wants to use a blockchain they can’t trust, or one that doesn’t send their transactions when they want.

To simplify an explanation, wallets will ‘flock to’ and sync with the longest chain available. If you’re able to come in and find blocks faster than the rest of the network (Because you control most hash power), you can attack it.

These small-time ASIC-farms could kill off something like EMC2 both with a 51% rewrite, or, by joining the network, mining for 30 minutes, finding a few blocks, and then leaving. With its Dark Gravity Well (DGW) difficulty retargeting it means it won’t take too long for them to “even out”, but the difficulty average from the last 2016 blocks means their network will take a while and transactions would get delayed.

In fact, this has just recently happened with Bitcoin Gold: https://www.ccn.com/bitcoin-gold-hit-by-double-spend-attack-exchanges-lose-millions/

But I’m going off-topic here, so:

An explanation of how a 51% attack could work

The mining pool distribution of Litecoin is as follows:

Hashrate of pools for Litecoin, 23rd of May 2018

If you were Antpool, LitecoinPool.org, F2Pool, LTC.top, ViaBTC etc… You could basically attack GameCredits, Gulden, Einsteinium etc (all pictured above) because those pools have WAY more hash power, and all those coins are single-algorithm: Scrypt.

Verge has just 2.67TH/s right now:

Verge (Scrypt) on whattomine.com

DigiByte has 4.06TH/s:

DigiByte (Scrypt) on whattomine.com

However, neither of these are vulnerable to the traditional “51%” attack described above.

Why are they not as vulnerable?

Good question!

These are multi-algo (Multiple hashing algorithm) blockchains.

This means that to perform a 51% attack on the network, it’s not just a bunch of ASICs you need. We couldn’t have Antpool or F2pool switch the blockchain they’re mining for, aim at DigiByte or Verge, and successfully perform a 51% attack.

They could do that for GameCredits, Gulden, Einsteinium, but not DigiByte or Verge, thanks to multi-algo.

If they had 51% of the hash power of a single mining algorithm, such as Scrypt, they would still be unable to attack the network in the same way that Monacoin and Bitcoin Gold have recently been attacked.

To successfully attack DigiByte, an attacker would need 93% of one algorithm, and 51% of the remaining 4x algorithms.

So they’ll need a LOT of Bitcoin ASICs, Litecoin ASICs, Qubit & Groestl miners, as well as a TON of GPU’s for Skein mining.

The problem Verge are having is not that somebody has 51% of all their algorithms, it’s due to a poor implementation of the Dark Gravity Well (DGW) difficulty adjustment algorithm. DigiByte on the other hand uses DigiShield (Now named MultiShield after we added support for DigiShield on multi-algo blockchains such as we are now).

Basically, DigiShield prevents one algo from mining more than 5 blocks in a row, even in the event they were able to stack their hashing power in such a way that they tried to beat the difficulty adjustment. This was the issue that Verge suffered back 7–8 weeks ago: https://bitcointalk.org/index.php?topic=3256693.0

However this issue is again slightly different. It’s not an issue with being multi-algo which traditionally would secure a coin more than having a single algo, but rather it’s an issue with the way that Verge have implemented their difficulty retargeting.

Not quite following?

It’s OK, don’t worry if you’re getting mixed-up with hashing / mining algorithms vs difficulty-adjustment algorithms. Even Charlie Lee doesn’t understand:

https://twitter.com/SatoshiLite/status/999496970384297986

You see it’s not that having 5 mining algorithms made it worse (As Charlie originally tweeted), it’s the difficulty adjustment (DGW) that’s the issue. The 5 mining algorithms really isn’t the problem, because this wouldn’t affect DigiByte. DigiShield (MultiShield) solves this entirely, and Verge still hasn’t implemented MultiShield after their original hack at the start of April 2018.

They should, it would make their problems go away in a flash.

Are Verge actually under a “51% attack” then?

Well not quite, because technically the attackers don’t have over 51% of the hash power for the whole coin. They only have 51% of two of the mining algorithms, and they’re taking advantage of other flaws in Verge, being the difficulty adjustment algorithm (DGW).

The Digi(Multi)Shield code is open-source and the suggestion has been made numerous times to the Verge developers that they implement it. It would save them so much heartache, because there’s really no need for these sorts of hacks to take place. We as a greater community of crypto developers (though I’m not a developer, but still) all know better by now.

DigiByte helped DOGE to implement DigiShield back in 2014 which saved the coin.

Bitcoin Gold uses DigiShield to protect against sudden influx / exits of mining power (Though, Bitcoin Gold needs to be MultiAlgo to prevent from the 51% attack they’ve just suffered).

Litecoin mining pools could in theory destroy several smaller blockchains in a matter of hours if they were so inclined to feel malicious. The whole point of the Blockchain is that you trust nobody, that you expect there will be malicious users etc.

TL;DR — Is DigiByte vulnerable?

DigiByte is safe against a sudden influx of hashrate from a single (Or even several) algorithms, thanks to it being multi-algo, means they can’t perform a 51% attack like Bitcoin Gold suffered.

DigiByte is also safe against these other attacks that affected Verge, thanks to Digi(Multi)Shield.

Litecoin is also vulnerable to different kinds of hashrate fluctuation attacks that Jared mentioned in his talk at the Texas Blockchain Conference (Skip to 8 mins 35 seconds in), although it’s about BTC it affects LTC the same:

https://www.youtube.com/watch?v=MiBFKJVF69A&t=515s

In closing

I’d also like to clarify that this doesn’t have to be a dog-eat-dog world out there, this doesn’t have to be a “winner takes all” scenario, and I’m sure the other DigiByte developers / Foundation members would also agree with me there.

We can help each other, through these projects all being open-source, and although I don’t forsee the DigiByte developers going out of their way to bake the code into the Verge blockchain as they did with DOGE back in 2014 (Mostly because it’s a big time-suck that could be better spent elsewhere for very little recognition), the code is still readily available and it would solve Verges issues immediately.

Also, don’t be surprised when you see Charlie trying to shit on DigiByte. Not sure why he’s so mis-informed about it, he’s usually such a switched-on fella. After all, the DigiByte core wallet was based on Litecoin originally (Though our 100% independent DigiByte Blockchain is now closer to Bitcoin than Litecoin thanks to the lack of development by LTC).

And don’t worry, Charlie Lee came around in the end too:

Now although we occasionally take jabs at each other, I would hope the community never forgets that we came from the Litecoin code-base.

So it would seem that all’s well that ends well… Unless your blockchains being attacked, in which case, that sucks, and you need MultiAlgo + DigiShield!