First Melon Manager Competition in Retrospect
Following the completion of the competition, we thought it would be fruitful to put together a quick write-up detailing how we felt the competition went and what could be improved for future competitions. Before we get dive into this, we at Melonport would like to extend a massive thank you to all the people who participated in the First Melon Manager Competition! We were so impressed with your feedback and general desire to learn about what we’d built and how it works. It’s been incredibly useful to us and we fully intend to work that feedback into our development process moving forward — more on that later though.
The competition aimed to enable users to test the Melon protocol in a safe environment with play money/assets on the Kovan Ethereum testnet. The prizes for the top 3 participants were Melonport t-shirts this time round, though we will increase the prize pool significantly as we move forward.
The event ran from noon (12:00) CET Monday 24th July 2017 until noon (12:00) CET Friday 28th July 2017. We had a total of 104 participating accounts in the competition, though as we had no intention of keeping any identifying information on the participants, there’s no way to know if some users created multiple accounts or not. We take the fantastic level of interaction between the competitions participants and our developer team in the Melon project Gitter channel as a better metric to show success of the event. It really was quite phenomenal!
The only identifying information we collected from users was email addresses (which have also since been deleted) so that we may contact the winners of the event for an address to send their prizes to. This is a theme we’d like to continue moving forward — we will never ask for more information than is required from Melon users, and anyone who wishes to engage with us shall be able to do so in an entirely voluntary manner. No recycling of email addresses into our newsletter list!
That said, if you’d like to receive our monthly newsletter you’re welcome to sign up here!
For us at Melonport, the most important part of this competition was the incredible amount of feedback we received from all the participants.
Some of the main points were:
- It’s not immediately clear how to use the order-book by clicking on the orders a user might want to take (buy/sell). This user experience can be refined moving forward.
- One of the issues a user must deal with (as a consequence of the software being built on top of a blockchain) is the confirmation times between each block. This can sometimes leave them scratching their heads as to what exactly is happening on screen. Also, for some interactions with a fund, multiple transactions are required (investing, taking multiple orders at a time, creating new ERC20 tokens to be held by fund, etc). We can do more here to alert the user as to what they should expect and what is happening. At the same time, we don’t want to saturate them with too much information!
- The servers that we used for the competition started to see very heavy load towards the end, especially on the last day leading to some disruption in user experience. We didn’t want to fully migrate at the time as that would have meant even more downtime. In the future, we will be sure to run the competition on far more powerful infrastructure to solve this issue. We also have some refactoring planned that should significantly increase the efficiency of the Melon Portal, lessening its burden on the servers that it’s hosted upon.
- We built a negative spread into the liquidity provider which makes the markets on our testnet decentralised exchange. We did this to create an incentive for people to consider building bots/algorithms to take advantage of the arbitrage opportunities. Initially, this spread was too wide and occurred too frequently, it was fixed by the end of the competition. We’ll be careful how we implement arbitrage opportunities in the next competition, though we absolutely do want to maintain the incentive to build bots and algorithms. Of course, the next event will also have security compromise/system abuse rewards built in. If you can take advantage of the system we’ve built, at this stage that’s fine because once we know about it we can fix the underlying issue. One of the main reasons we are running the competition series is to battle test the protocol.
- Sometimes it took a long time to get more liquidity for an asset pair; our liquidity provider was not providing liquidity fast enough when all orders were executed by users. This can be solved by changing the liquidity provider to be more reactive.
- Users have let us know their desire to be able to see (or at least download) their trade history. We’ll make that a priority for the next release.
- Users expressed that they’d like to have a way to retrieve the Testnet Ether Token in their MetaMask account when transactions occasionally fail when interacting with the “invest” function in a fund. We’ll endeavour to create a solution for this ASAP.
- Users would like to be able to invest in other users funds. We hope to have this feature ready soon — there’s no doubt that “fund of funds” will be an important part of the future Melon Fund ecosystem!
The portal webpage saw over 5000 individual sessions, so we’re pleased to see that many of the funds were being very actively managed for the duration of the competition. Of the 104 participating funds, 20 did not invest in any assets beyond Ether, and 10 invested nothing at all. This could have something to do with the accessibility of the order book (where users can take their invested Kovan Testnet Ether and diversify into other assets).
We did some quick calculations on Melon’s impact on the Kovan testnet. Melon Funds themselves along with the associated user-based management activity were .incredibly light, accounting for just 0.18% of all the Kovan testnet transactions in existence. The liquidity provider and data feeds on the other hand account for 79.54% of all the txs in the week of the competition. This meant we were using around about 1–3% of the Kovan testnet’s total capacity.
“What!? That’s a bit high!” we hear you say. It’s worth noting that of the 79.54%, 75% of the txs came from just the liquidity provider itself. In normal circumstances (Melon on the main net), there would be no need for the liquidity provider to exist as Melon would plug into various real decentralised exchanges which have actual users, not a manic and verbose bot! Also, most decentralised exchanges now keep their order books off chain for efficiency with settlement on chain. This leads to a much smaller capacity footprint.
For future competitions we’d like to do a more sophisticated statistical breakdown of what occurred and what exactly people were buying/selling/holding.
In conclusion then, we feel the competition was a brilliant exercise for us at Melon and we are so pleased with all the fantastic feedback we got from our growing community. It was a great step forward in the journey to version 1.0. Now, we need to focus on building this feedback back into our dev schedule, iterate upon it, and create an even better user experience for the next competition (details coming soon!) which will have prizes of real value (not that our limited edition Melon t-shirts aren’t exceptionally of course!).
Before joining Melonport as Head of Business Development, George previously worked with the Ethereum Foundation starting as Business & Partnership Director in Mid 2014. He quickly established himself as a key communicator, moving to the role of Head of External Relations where he successfully advocated the Ethereum platform to the world and coordinated the Ethereum Foundation’s yearly developer conferences in London and Shanghai. During his 2 years at the Ethereum Foundation, George grew a strong appreciation and understanding of the Ethereum community and the ecosystem it inhabits.
Learn more about Melonport: https://melonport.com/