Introduction

The Terra blockchain is a decentralized network of validators and stakeholders. Validators are responsible for verifying transactions, producing blocks, and participating in consensus, which makes them critical to the security and stability of the network. Delegators are individuals or entities that have deposited their stake (LUNA tokens) in a validator, allowing them to take part in consensus and contribute to the network’s security without the need to run a validator node on their own. Validators are rewarded for their work on the network, and these rewards are shared with their delegators. This incentive system helps to ensure that validators and delegators have an incentive to remain active on the network and be actively involved in consensus-building. It also encourages new participants to join the network, thereby increasing the security of the Terra blockchain.

In nutshell, validators are an essential part of the Terra blockchain and are in charge of maintaining the network's security and stability. These nodes participate actively in the consensus procedure, generating blocks and verifying transactions. Validators are essential to preserving the network's integrity since they take part in consensus and stop malicious activity. A delegator is an individual or entity that has given a validator control over their stake, which consists of their LUNA tokens. Delegators can indirectly participate in consensus and contribute to network security by transferring their stake, which frees them from having to operate a validator node themselves.

Question:

Terra - 12. The Eclipse Score

Invent an easy-to use "Eclipse Score", based on at least 3 metrics, to rank validators who are not effectively representing their delegators via voting.

(For example: if you participate in only 5% of votes but are in the top 5% of validators by LUNA delegated, you should have a terrible Eclipse Score!)

Provide a table and at least 1 visual that displays validators according to your Eclipse Score, e.g. a bubble chart with LUNA Delegated vs. Votes Attended. Additionally, analyze the top 5 best and top 5 worst validators according to your Eclipse score.

Methodology

The techniques used for data wrangling and analysis are covered in this section. The essential features effecting the contribution of the validator's vote were combined from two tables "terra.core.ez_staking" and "terra.core.fact_governance_votes" to ensure the stability and trustworthiness of the network. Both tables are joined by using the “VALIDATOR_ADDRESS” column from terra.core.ez_staking table and “VOTER“ from terra.core.fact_governance_votes table. The following steps are performed to calculate the Ecplise Score:

Results and Discussion

In some cases, very low amounts of LUNA are assigned to the validator compared to other validators and the validator’s participation rate was high. In this case,

Top 5 best validators based on their Eclipse score:

1. Validator with the address of “terravaloper1yy5379nk24xffpq4egw7ue4ad3j7l79ma2enmz“ has the highest Eclipse score of 68.54% and has an excellent balance of normalized voting participation of 68.27%, normalized delegated stake amount and normalized vote reliability.

2. VALIDATOR_ADDRESS: terravaloper1kxyzu575tex8dv6d5uk26t4q8x5lusw082w3t8

   Eclipse score: 16.68%

3. VALIDATOR_ADDRESS: terravaloper1puzp2yjqps43x7nse33svljc550xjz35jxg432

   Eclipse score: 14.81%

4. VALIDATOR_ADDRESS: terravaloper1gtw2uxdkdt3tvq790ckjz8jm8qgwkdw3uptstn

   Eclipse score: 13.32%

5. VALIDATOR_ADDRESS: terravaloper188e99yz54744uhr8xjfxmmplhnuw75xea55zfp

   Eclipse score: 11.27%

Top 5 worst validators based on their Eclipse score:

1. VALIDATOR_ADDRESS: terravaloper1qdspntsxzna8l8sp4ma8x0dyk0q7cgzvfj9vgc

   Eclipse score: 0.053%

2. VALIDATOR_ADDRESS: terravaloper1g0deaesmwy2qg9zupynx6ychspe60lquu3eyze

   Eclipse score: 0.011%

3. VALIDATOR_ADDRESS: terravaloper1rgdps79da300t87twwkcly6ec3vlpgpfgnsel5

   Eclipse score: 0.013%

4. VALIDATOR_ADDRESS: terravaloper15389y60f0rk596tm38qewqycr8ztt8hshsf2mk

   Eclipse score: 0.015%

5. VALIDATOR_ADDRESS: terravaloper1arm3fzjp7qv8e5gyhraeykcpjte6wh562yre97

   Eclipse score: 0.017%

   \

Conclusion

We could draw the conclusion that from the perspective of the Eclipse Score, a validator with high participation and more successful votes in terms of voting in line with the consensus of the network (high reliability and consistency in voting) as well as handling higher amounts of stake is preferred. To prevent a false Eclipse score, all of these parameters should be normalized to the sum for all validators. In general, a validator’s high vote reliability and participations lead to high Eclipse score, if the delegated stakes are similarly substantial as well.

1. Normalized Voting Participation:

To find the normalized voting participation, first of all, the number of the participations of each validator was calculated. Afterwards, this number was divided to the total participations of all validators. Finally this number is multiplied to 100 for percentage representation.

2. Normalized Voting Reliability:

The second parameter effecting the Eclipse Score is reliability of the validator. We could define a voting reliability parameter which represents the consistency of a validator's vote with the consensus of the network. The information for this parameter resides in the terra.core.fact_governance_votes table. Therefore, this table is merged with the terra.core.ez_staking table using an inner join with “VOTER“ and “DELEGATOR_ADDRESS“ columns, respectively, and the two columns of “VOTE_OPTION_TEXT“ and “TX_SECCEEDED“ are added as well. Based on these two columns we could define how reliable a validator’s vote was. If a validator’s vote option was “YES” and the transaction was successful (TRUE), that vote was reliable. Therefore, a new column was created and if this condition was occurring a value of 1 was assigned to that new column called “VOTE_RELIABILITY“. When the validator’s vote was not YES and the transaction was TRUE (successful), this validator had was not aligned with the consensus of the network, therefore, a value of 0 was assigned to the new “VOTE_RELIABILITY“ column, to show that this vote was not a good one. Each validator’s vote underwent the same procedure.

Finally, for each validator, all of the values in “VOTE_RELIABILITY” column were summed up and added in a new column called “TOTAL_VOTE_RELIABILITY“ which represents a specific validator’s reliability in voting.

Finally, in order to account for the weight of large numbers of votes in the calculation, each validator's reliabilty score is normalized to the sum of the votes cast by all validators.

Delegated Stake Amount:

Total amount of delegated stakes by each validator was divided to the sum of all delegated stakes by all validators to have a normalized values.

relation between the validator’s voting participation and voting reliability

The following figure shows that the reliability of the validator's vote is directly related to their participation in the voting process. In other words, the consistency of the vote improves as the validator participates in more votes.

Relation between the validators’s voting participation and validator’s stake

The findings indicate that as the stake amount for the validators is raised, so does their voting involvement.

Relation between the validator’s voting consistency and validator’s stake

Similarly, we observe that increase in amount of delegated stakes to the validator lead to the enhancement of their consistency and reliability in voting.

Eclipse Score

The Eclipse Score is related to all of the parameters described above and we could write the Eclipse Score as following:

Eclipse Score = (Normalized Voting Participation) + (Normalized Voting Reliability) + (Normalized Delegated Stake)

Top metrics to evaluate a validator's vote

1. validator’s participation:

   In nutshell, voting participation is the proportion of votes cast in which a validator took part. The regularity and dependability of a validator's participation in voting events on a blockchain network is measured by the metric known as voting participation. A validator that routinely participates in voting events and contributes to the network's governance has a high voting participation rank. On the other hand, a low Voting Attendance rank suggests that a validator is not actively participating in network governance or is missing voting opportunities. The stability and security of the network is ehanced by the higher participation of the validators. It's important to keep in mind, though, that a high voting participation score does not necessarily mean that a validator had correct decisions or that the votes are in line with the network's unanimous agreement; rather, it demonstrates the validator's reliability and support for the security and stability of the blockchain.

   \

2. Voting Reliability:

   The vote reliability indicator measures how closely a validator adheres to the network's consensus. This measure is essential for evaluating a validator's efficiency because it shows their commitment to the network's security and stability. A validator that consistently aligns their votes with the network consensus is given a high vote reliability rating. A validator who regularly veers from the consensus or abstains from voting events may have a low vote reliability score, which could jeopardize the security of the network. The validity of a validator's voting skills are reflected in the vote reliability measure.

   \

3. Amount of the delegated stake:

   It is, to put it simply, the stake that has been assigned to the validator. The amount of cryptocurrency (in this example, the Terra token, or "LUNA") that has been assigned to a validator is referred to as the "Delegated Stake" in the Terra blockchain. The amount of LUNA that token holders have delegated to a validator as part of Terra's delegated stake model enables them to take part in consensus and contribute to the network's security and stability.

Distribution of the three metrics

The boxplots and density distribution plots of the parameters effecting Eclipse score are shown in following figures. Moreover, heatmap of the reliability, delegated stakes, participation and Eclipse score are shown.