It took around 160 minutes for the BTC.BTC pool to converge to the market price of BTC.

Findings

• No drop in price on November 18th 18:00
  • RUNE already lost a lot of value during the downtime especially just before trading got enabled
  • This might very well have lowered the effect of RUNE being sold as users were "buying the dip"
• After November 18th we can see that bots were swapping from RUNE to other assets (SELL_VOLUME) more often than they swapped to RUNE (BUY_VOLUME)
  • It appears like bots were selling their RUNE slowly instead of just after obtaining it

Thorchain, which is built on the Cosmos blockchain, is a decentralized liquidity protocol that specializes in cross-chain connectivity, allowing traders to swap tokens between different networks. It connects a ever increasing amount of chains, currently it provides services for

• BTC
• ETH
• BNB
• TERRA
• BCH
• DOGE
• LTC

Terra Collapse & ThorChain

Terra which was one of the TOP 10 cryptocurrencies experienced a total collapse in mid May 2022. Prices of most On-Chain assets went to 0, the chain experienced heavy congestion and was halted several times. This triggered a automatic suspension of trading with the Terra chain which to date has not been resumed.

The goal of it being the removal of all Terra asset liquidity pools.

Naturally this means that trading Terra assets will never be resumed.

Other Proposed Resolutions

Instead of the steps that actually will happen (see above) there were other proposed steps to resolving this situation which did not get a majority from the node operators.

1. Move remaining liquidity into treasury controlled address

   • Refund LPs
   • Refund nodes that got slashed due to congestion

2. Resume trading

Next Steps

On June 3rd an annoucement about the next steps has been made:

• Single-sided LPs are refunded (asymmetric withdrawals with usual slip calculations)
• Then dual-sided (symmetric withdrawals)
• Impermanent Loss Protection will be applied

Figure: On Chain LUNA price (In liquidity pools) vs WLUNA price on ETH DEXes

• May 11th: TERRA trading halt
• May 29th: LUNA donation into LPs (No swapping happening, just addition of LUNA)

Especially the 2nd plan got very little approval as everyone assumes that arb bots would scoop up the remaining RUNE in the liquidity pools & dumping it onto the market. This would not only hurt the liquidity providers directly but also indirectly due to a large quantity of RUNE would be sold.

In this dashboard we will look into a previous scenario where trading was halted and later resumed:

ThorChain Consensus Failure November 12th 2021

Yesterday, THORNode got stuck in a consensus failure, meaning the active nodes were not agreeing on the state chain in a given block. The ThorChain block production came to a full stop for several days as nodes had to fully resync the chain. On November 18th trading was enabled. With prices on ThorChain being stuck for a while they deviated quite a bit from the actual value at the time of re-enabling.

Most currencies dropped ~10% during the downtime allowing users to buy the asset on exchange and swap it to RUNE for a profit.

It took around 25 minutes for the ETH.ETH pool to converge to the market price of ETH.

Bot Activity after trading re-enabling VS usual bot activity

In order to know if bots were more active than usual when trading got resumed we will first have to look at how active they are usually.

Which addresses are considered to be Bots?

Detecting bots is a massive topic on its own, as it is not the core topic of this dashboard we will stick to a simple definition for bots. An add "is a bot" if it

• Completed 10 or more swaps within one minute OR

• Completed 1008 or more swaps within a week (One every 10 minutes)

If you want to see more about bots within ThorChain check out this awesome dashboard made by @Zook#2707!

Findings

In the month leading up to the consensus failure bots accounted for

• ~80% of all swaps
• ~60% of all swap volume

Bot activity after trading got resumed

Findings

We can clearly see that bots dominated the swapping for the first few hours after trading got enabled. Add to that the previous finding that prices took a couple hours at most to return to their market price we can conclude that most profit was indeed made by bots.

Bot wallet behavior

We already found out that bots scooped up most of the RUNE that was available. But we should also look into what happened to that RUNE as well.

For that we will look at 2 metrics:

• RUNE / USD price
  • If a lot of RUNE gets sold we would expect a drop just after trading gets enabled
• BOT swap volume from vs to RUNE
  • If bots keep their RUNE we would expect swaps to RUNE to exceed swaps from RUNE

All following graphs will use the same 56 addresses as bot addresses. Isn't it crazy that 56 addresses can be responsible for such a large amount of on-chain activity? See below!

Bot activity leading up to the consensus failure

Summary

• Trading of Terra (v1) assets has been halted on ThorChain as of May 11th 2022.
• A decision has been made to not resume trading and instead refund liquidity providers and remove the Terra liquidity pools permanently.
• The alternative solution of resuming trading would have likely ended with:
  • Bots scooping up 80%+ of the remaining RUNE in the Terra liquidity pools.
  • Most of that RUNE being sold on the market possibly damaging the price of RUNE and hence the entire THOR ecosystem.