Augur Development Update: November 2020

a year ago

7 min read

As 2020 comes to a wrap, so does an eventful month of November for the Augur protocol. After a slow start, Augur v2 set new records for open interest and volume.

In spite of the bump in usage, Augur's work is never done. Product improvements are inbound which will make life easier for traders, liquidity providers, and consumers of market odds, alike.

Table of Contents:

  • Augur CE
  • Catnip & Election Markets
  • The Graph x Augur
  • Augur AMM
  • REP v2 Migration
  • Protocol Deep Dive: The Case for Invalid
  • Links

Augur CE

At the beginning of the month, Augur unveiled its first major product update since v2 launched. Augur Chad Edition (CE) allows users to bet in alternate currencies while inheriting similar security guarantees.

CE is a response to changing market conditions which make wagering in specific currencies more attractive:

Many long term crypto holders do not want to lower their exposure to ETH or BTC in todays market conditions. However, there are still many traditional bettors and traders who still prefer to use a stable currency and do not want additional market exposure when they're trading.

Read more about Augur Chad Edition in our introductory post.

Catnip & Election Markets

November was the Augur protocol's biggest month to date. The U.S. presidential election is a worldwide spectacle, and Augur was well-positioned to earn to a share of the trading activity.

Augur Election Stats (All Markets)

Recently launched Catnip.Exchange, a market platform built on Augur and leveraging Balancer for their AMM (automated market maker), was the leading presidential market in terms of volume and open interest.

Catnip is initially launching with single markets, and has followed up their 2020 election market with one asking if Democrats will sweep the House, Senate, and Presidency.

Catnip UI

Catnip broke its own daily volume record each day for nearly a week during the frenzy, reaching over $1m traded the day of and after the election.

Catnip Daily Volume

The election markets catapulted the current open interest in Augur v2, defined as the amount of money at stake, past highs set by Augur v1.

Augur Open Interest via The Block

As the 2020 election season winds down with only a few more races and questions left to be settled, Augur will lean on continued product innovation to unlock the next round of traction.

Learn a little bit more about the innovations that are planned below.

The Graph x Augur Protocol

Better late than never. Like many applications, Augur and its users are now benefitting from faster-than-ever data syncing courtesy of The Graph, an indexing and API layer for web3.

Augur's subgraph will serve as the foundation for more interesting and faster data displays in future Augur releases. This should vastly improve user experience.

Augur AMM

The rise in popularity of constant product market makers throughout the Ethereum ecosystem makes a lot of sense. The UX is fantastically simple, and liquidity providersthe equivalent of market makersjust have to provide capital to start earning fees (instead of crafting a proper book).

In witnessing the success of Catnip and other AMM-based prediction markets during the election, it became more apparent than ever that the simplicity on both sides of the liquidity equation would be a major boon to usage.

Augur is working on a custom implementation that hopes to achieve the best capital efficiency and lowest slippage amongst prediction market AMMs.

Augur's AMM will be packaged in a brand new user interface and is slated to launch in the near future.

REP v2 Migration

The REP v2 Migration continues and now the majority of REP resides in the v2 system!

As always, you can check to glean migration details including the amount of REP in each universe, DEXs and CEXs supporting REP v2 pairs, contract addresses, and more.

Protocol Deep Dive: The Case for Invalid

Truth, particularly unwitnessed objective truth, is an unsolvable problem.

An event in the world is said to have occurred but only a well known serial liar witnesses it. Did the event actually occur? Would you bet on it?

This problem rears itself with little provocation in the realm of decentralized prediction markets.

In a centralized world this class of problem is relatively easily handled by administrative control over the wording and resolution of a given market. Consider the very realistic case of a website meant to be the source of truth going down or suffering from a bug that makes it produce obviously incorrect data. In a centralized world clarifications, edits, and new resolution rules can be made after market creation meaning that ultimately a trader is betting on the perceived honesty of the central party controlling the market. This generally works out fine since central parties care about their reputation, but carries the inherent problems of being centralized.

In a decentralized prediction market edits after creation must never occur since no special trust can be given to market creators to not be malicious. For example a market creator, if it were in their power, could simply switch the outcome descriptions moments before market end time, resulting in shares no longer representing what traders actually expected. This means that the letter of the market, and the nature of how it resolves, must be apparent from the moment of creation for traders to have confidence in it.

Invalid Markets In Augur V1

In Augur V1 the proposition of a market resolving “Invalid” was initially thought to be something that would be relatively infrequent. The system requires market creators post a bond which they only have returned if the market resolves to a non-Invalid outcome. This bond increases in the system with the frequency of Invalid markets. That bond and the general gas cost involved with market creation seemed like a reasonable deterrent to malicious Invalid markets.

As it turns out malicious Invalid markets were rather profitable, both in the trading and resolution phases of a market. The Invalid payout in the first version of Augur simply divided value among outcomes equally. This meant even if traders suspected the potential for Invalid it was non obvious from the prices in the orderbook and it was fairly trivial to sneakily construct a book where if the market resolved Invalid the creator of that book would profit. The prospect of Invalid markets is something that most reporters and fans of prediction markets in general have reasonable disdain for. This hesitance to make markets Invalid was also a tool to profit by baiting well meaning participants into making a market resolve “reasonably” when in reality it could only ever objectively resolve as Invalid.

The existence itself of malicious markets isn’t solvable without sacrificing decentralization. They’ll always exist, so the real problems that had to be solved were:

  1. Hiding Invalid markets from traders
  2. Making it unambiguous what markets are Invalid

In the first version of the protocol there was little information available to objectively determine if a market is invalid.

In the first version of the protocol there is relatively little that can be looked at to objectively and programmatically determine a market is Invalid. Without moderation this is the only way however to hide a likely Invalid market from would be traders. The best solution that could be thought up and which was ultimately implemented in the V1 client was to hide markets where the current best bid/offer would profit as a result of a market resolving as invalid. This is not ideal since a market could be legitimate fairly reasonably with this characteristic.

Since the Augur protocol is decentralized and the resolution of markets is ultimately in the hands of REP holders the second task was up to the community. As a result a fairly consistent ethos of strictness was born and documented well at the site The Majority Report.

Invalid Markets In Augur V2

In Augur V2 the architectural change was made to make “Invalid” a first class outcome. This means that it can be bet on explicitly and that when a market resolves “Invalid” all payouts will go toward compensating holders of “Invalid” shares.

This change has two primary benefits:

  • The payout for when a market resolves Invalid is much clearer
  • The orderbook for Invalid, particularly the bids, provides a clear signal of a market potentially being Invalid.

Looking at the existing Bids for Invalid a client library can calculate if a bid currently exists which would be profitable to take assuming the market resolves as Valid. If such a Bid exists it is reasonable to suggest that the market may be Invalid, and thus hide it from a trader’s normal view.

Obviously a malicious party may attempt to hide markets by creating such Bids. Since a market will only be hidden if taking the order is profitable however this is effectively paying a decentralized team of Augur moderators to scan for such behavior and fix it. In time market creators could even employ automation to take such bids to earn some additional fees and keep troll Invalid bids off the book.

This more direct form of indicating Invalidity means that reporters can be much stricter with rules as well. In the past some leeway was awarded to markets that didn’t follow the letter of the market but instead resolved based on the spirit of the market intent. This set a potentially dangerous precedent that will now not be needed as the capability to signal Invalidity exists before reporting begins.