The attack vector against tokenized systems using voting systems to find truth

Tokenized voting systems have been proposed for several projects in the blockchain space as a means to discover truth. In a single market with any number of options, the option with the majority of votes is deemed the truth, and participants that vote in this majority receive a reward. Else, they’re penalized. The truthful option is the schelling point which actors in the system can flock to create a majority. Economics theory deems this a stable approach in incentivizing individuals to contribute truth.

An example of such a system would be: A prediction marketplace with multiple markets running simultaneously. Closing the prediction marketplace requires token owners voting, and the outcome with more stake is immediately recorded as truth in the blockchain.

The deviation from economics theory is when multiple markets are present, creating multiple schelling points. In a naive tokenized system, all these schelling points provide equal reward and thus they are all equally “right”. The distribution of attention will then skew. An intuitive example — majority of individuals would be willing to verify that Germany lost to South Korea 0–2 in the 2018 World Cup but significantly less people would be willing to verify that, for example, Rubin Kazan U18 wins Nizhny Novgorod U18 7–0. The markets requiring more time and effort than others will see less attention and stake.

In these situations, the naïve voting implementation without any settlement period is vulnerable to a last-minute attack whereby the attacker puts forth stake matching the amount of stake by honest voters plus ε. The lateness of the attack prevents the community from manually directing attention to these attack sites. Note: Markets with lower volumes will be targeted in the attack. Less tokens coupled with the immediacy of rewards makes it likely that the attacker can cash out his reward before the protocol is frozen/forked.

Some possible solutions and thoughts:

1. Add a settlement period for markets after they close before awarding tokens. Thoughts: Under normal operation with rational actors, majority of stake is locked up (optimizing rewards requires minimizing time holding unlocked tokens in a work-token model). The 51% attack is massively cheapened because it’s now “the 51%-unfrozen-stake attack”.

2. Incentivize attention towards smaller markets. Thoughts: Individuals incentivized to hunt for smaller markets might not hold sufficient stake. It will improve the equality of the distribution, but will it be enough?

Overall, I just don’t have a good feeling about voting systems across multiple markets. That said, the thoughts aren’t sufficiently rigorous to merit any protocol changes for projects with this model

If there’s anything I missed out would love to hear it out!