For the most part, questions and challenges related to bounty programs are released in one-go, with mostly the same deadline across the board. This is regardless of the questions’ difficulty level and complexity. While this style of release is still able to generate good submissions and community engagement, we see some problems that come with it.

Firstly, some bounty questions around the protocol or blockchain involved are left unanswered given the time constraints and the difficulty level. Another is that analysts are not able to first familiarize themselves with the topic of concern and strategically dive deeper into the bounties which leads to submissions that can be much improved. Lastly, due again to not having enough time, analysts are not able to complete several bounty questions which decreases their potential earnings from bounty programs.

Our hypothesis is that

1. If bounty questions are released based on difficulty level with easier ones going out first, more questions around the protocol or blockchain can be answered. This is because analysts can have a chance to first familiarize themselves with the protocol and tackle complex questions later on.

2. If the volume of questions are controlled with a staggered release, analysts will have enough time to think through their solutions and generate higher quality submissions.

3. If analysts are given enough time, they will be able to tackle more bounty questions and maximize their potential earnings.

The solution we have come up with to test out our hypotheses is to do tiered challenges for bounties. This is wherein bounty questions are released in a series of weeks at an increasing difficulty level. The first tier of questions is for analysts to familiarize themselves with the basics of the particular topic, where that might be about a particular protocol or blockchain. The next tier is for diving deeper into more nuanced aspects of the topic. Finally, the last tier is for solving harder, more complex questions not only regarding the protocol or related blockchain but possibly the surrounding ecosystem as well.

We are currently implementing this as an experiment with our Convex Bounty Program, although due to time constraints, we were not able to explicitly classify the questions according to difficulty level.

We started with the release of the first tier last February 7, 2022, giving participants until the Friday of that week (February 11, 2022) to submit their analyses. Since then, we have finished releasing the subsequent tiers on a weekly basis at a rate of five questions per week. Currently, our team of graders are in the middle of assessing submissions for each tier and we will soon announce the complete results.

The results of this experiment are not yet complete but so far:

• We received 48 submissions which is 4x times more than our last bounty program.

• Comparing averages, this program generated three submissions per bounty question compared to our last wherein the average was one submission per bounty.

• 14 out of the 15 bounty questions had a submission which is another improvement over our last program that only had 70% of the questions answered.

• Two questions were able to get seven submissions each.

Stay tuned for the complete results of his experiment after we conclude the Convex Bounty program by joining our Discord and following us on Twitter. We would love to hear your feedback on our bounty programs as well.

For the most part, questions and challenges related to bounty programs are released in one-go, with mostly the same deadline across the board. This is regardless of the questions’ difficulty level and complexity. While this style of release is still able to generate good submissions and community engagement, we see some problems that come with it.

Firstly, some bounty questions around the protocol or blockchain involved are left unanswered given the time constraints and the difficulty level. Another is that analysts are not able to first familiarize themselves with the topic of concern and strategically dive deeper into the bounties which leads to submissions that can be much improved. Lastly, due again to not having enough time, analysts are not able to complete several bounty questions which decreases their potential earnings from bounty programs.

Our hypothesis is that

1. If bounty questions are released based on difficulty level with easier ones going out first, more questions around the protocol or blockchain can be answered. This is because analysts can have a chance to first familiarize themselves with the protocol and tackle complex questions later on.

2. If the volume of questions are controlled with a staggered release, analysts will have enough time to think through their solutions and generate higher quality submissions.

3. If analysts are given enough time, they will be able to tackle more bounty questions and maximize their potential earnings.

The solution we have come up with to test out our hypotheses is to do tiered challenges for bounties. This is wherein bounty questions are released in a series of weeks at an increasing difficulty level. The first tier of questions is for analysts to familiarize themselves with the basics of the particular topic, where that might be about a particular protocol or blockchain. The next tier is for diving deeper into more nuanced aspects of the topic. Finally, the last tier is for solving harder, more complex questions not only regarding the protocol or related blockchain but possibly the surrounding ecosystem as well.

We are currently implementing this as an experiment with our Convex Bounty Program, although due to time constraints, we were not able to explicitly classify the questions according to difficulty level.

We started with the release of the first tier last February 7, 2022, giving participants until the Friday of that week (February 11, 2022) to submit their analyses. Since then, we have finished releasing the subsequent tiers on a weekly basis at a rate of five questions per week. Currently, our team of graders are in the middle of assessing submissions for each tier and we will soon announce the complete results.

The results of this experiment are not yet complete but so far:

• We received 48 submissions which is 4x times more than our last bounty program.

• Comparing averages, this program generated three submissions per bounty question compared to our last wherein the average was one submission per bounty.

• 14 out of the 15 bounty questions had a submission which is another improvement over our last program that only had 70% of the questions answered.

• Two questions were able to get seven submissions each.

Stay tuned for the complete results of his experiment after we conclude the Convex Bounty program by joining our Discord and following us on Twitter. We would love to hear your feedback on our bounty programs as well.