Polkadot

This is more of the ^{_{(“power rangers”-ilk, c.f. also)}} Ethereum-esque, recycled, repackaged dog shit flawed design concepts, served up on a silver platter of technobabble to fool the gullible neophyte speculators.

Greater fools buy technologically vacuous pink polkadork polkadots ICOs.

I have been unable to find any other critique of Polkadot. I’m flabbergast that I’m ostensibly currently the only person in the cryptocosm capable and willing to do such a critique. I suppose others just got tired of repeating the same fundamental flaws they’ve been repeating for every failed attempt to scale Ethereum with proof-of-stake over the past 3+ years.

“Dr.” Gavin Wood¹ ostensibly hints from the start of his polkadot background colored “pink paper” that he’s pulling some technobabble nonsense out of his arse when he (presumably intentionally as it’s repeated throughout the paper) misspells ‘canonicity’ as ‘canonicality’. It only gets worse from there…

The Polkadot design is fundamentally flawed in at least the same ways as previously mentioned for Ethereum’s Plasma proposal. For example, in §6.6.1. External Data Availability they repeat the flaw I explained in my post about Plasma as follows:

Thus, such unresponsiveness can’t be deterministically penalized because it’s (in terms of what can be proven cryptographically³) indistinguishable from an attack where a majority of the validating nodes pretend the minority is unresponsive and cause the minority to forfeit their bonds. The threat of which forces all validating node into a colluding oligarchy in order to secure the promise of protection. But a promise is not trustless (aka trustproof) thus in actuality such a design lacks a (non-malevolent oligarchy) Nash equilibrium.

The cited section is chock-full of game theoretic assumptions on incentivized behavior which isn’t formalized nor holistically analyzed for out-of-band adversarial flaws. Amateurish nonsense.

It only gets worse from there…

Seriously these decorated² “PhD”, “Thiel Fellowship fellow” imbeciles shouldn’t be trusted to do anything related to economic game theory. For example, the only recourse they have proposed for the failure of fraud proofs or collusion outside safety margins is centralized consortium overlords which hardfork as necessary:

Once a block is out of the challenge period, it is considered unassailable. If it does turn out to be bad, then it must be fixed in the same way as a protocol upgrade. As such it is likely that validators and other large stakeholder would act as Fishermen to protect their investment.

Dr. Gavin Wood had argued after the Ethereum DAO attack that democracy (which by the way is oligarchy rule) should have the power to trump canonicity. Gavin even tried to obfuscate that there would have been no Nash equilibrium to converge on a hardfork without Vitalik’s and Ethereum foundation’s leadership in support of the hardfork (which was actually the case that there was chaos until Vitalik put his foot down and demanded the rollback hardfork).

Their quoted design assumption above that validators will be incentivized to always be honest is at-best not formally justified (and in actuality it’s incorrect as I broad-stroke explained for the flaws of Plasma). There’s no amount for a bond that can disincentivize an opportunity cost to maximize earnings (such as via collusion). The only temporal Nash equilibrium of proof-of-stake systems is (oligarchy collusion for) extraction of the maximum rents from the system that can be extracted until eventually the systems implode which isn’t sustainably a Nash equilibrium because relies on trust (that some powerful coalition in the oligarchy won’t defect to pursue a different strategy). Evidence that not even one proof-of-stake system has remained in the Top 10 at coinmarketcap.com. They rise up into the Top 10 then eventually fade away as the (power-law distribution centralized oligarchy control over the) majority of the stake maximizes extraction. The common neophyte retort, “Which proof-of-stake systems have entirely failed due to 51% attack?”, entirely ignores the fact that the oligarchy is extracting maximum rent insidiously and/or surreptitiously.

Besides their paper has no holistic liveness and safety analysis (i.e. analysis taking into account all the moving parts in the holistic economic game theory, and not just presuming/pretending that formalized Byzantine agreement aka BFT consensus algorithm protocol for the master relay chain in isolation is indicative of anything). When they ever get around to formalizing that, they will discover as I explained for Cosmos, that there’s either no canonicity or there’s denial-of-service liveness attacks. The range of vulnerabilities in such a complex interaction of bonds, fraud proofs, challenges, disparate sub-chains, etc.. is far beyond the flaws I enumerated for Plasma. It behooves them to attempt to make a formal analysis but of course they won’t because it’s virtually impossible and they’d surely find insoluble flaws such as the ones I enumerated.

The same philosophy of a centralized consortium overlorded system is evident in their adoption of the hardfork checkpointing as a way to avert long-range nothing-at-stake attacks.

Also the (in)security provided by Polkadot only applies where the confirmation latency for any shard (aka parachain) will be as slow as the block period for master relay chain. Thus Polkadot’s (in)security is inapplicable for example for instant transactions in the context of millions-of-TPS transaction volume scaling. Although a Polkadot parachain could offer some technology for scaling instant transactions, it would have to provide it’s own security and canonicity. As I had explained for Cosmos, separate incentivized security and canonicity, means there’s no global security nor canonicity. These are just fundamental invariants that Polkadot hasn’t even formalized yet so they don’t realize how fundamental flawed their entire concept is. Thus this “pooling security” is essentially nonsense (and ironically in that video Robert Habermeier claims that those who design blockchains don’t or shouldn’t have to focus on security because they could instead delegate to “pooled security”, which is absolute nonsense that indicates they don’t understand formally how security, safety, and liveness are intertwined in the holistic economic game theory in ways they haven’t formalized).

All of these sort of overly complex Rube Goldberg machines obfuscating crap designs have been shown over and over again in the community discussion (e.g. Peter Todd et al) to be insecure. They’re overly complex because they have no coherent robust solution to fundamental flaws. Ethereum and its offshoots such as Polkadot continue to bark up a rotten tree and pursue non-formalized nonsense designs. And apparently continue (c.f. also) to do quite well selling technobabble ICOs to gullible neophyte speculators.

At least they admit their paper is essentially useless fluff:

Bond-confiscation and conversely reward provision has not been deeply explored.

The underlying issue is that none of them have any clue how to solve these fundamental flaws. I do. But I’ve quit for reason of declining health and torture threats to those who aren’t anonymous.

Note I posted a comment on the Medium blog about the BFT consensus algorithm being proposed for Polkadot’s master relay chain:

The paper does not present a holistic analysis of the game theory of the range of vulnerabilities in Polkadork’s complex interaction of bonds, fraud proofs, challenges, disparate sub-chains, etc.. It behooves them to attempt to make a formal analysis but of course they won't because it's virtually impossible and they'd surely find insoluble flaws such as the ones I enumerated (on my blog).

A fundamental issue is that proof-of-stake is vulnerable to the power-law distribution of wealth, and because there's no external resource consumed to incentivize a Nash Equilibrium of non-malevolent behavior. Thus all PoShit systems are wealth extraction systems, contrary to the BS about the stake being incentivized to be honest. For example, do you ever see any such whitepaper mention shorting the market or other out-of-band attack vectors. Nope.

¹ _{Don’t confuse him with the other “biobankrupt wedidditreddit” dork “power ranger” Gavin Andresen who incorrectly thought “Faketoshi” (aka Craig Wright) was Satoshi and was formerly in control of the commit access for the open source Bitco[i]n Core f*rk-off development.}

² _{Imagine the decorations on the Christmas tree on their “power ranger” lapels.}