Gate Institutional AMA Series 17 - Mastering Crypto Trades with Hummingbot: Your Platform Guide

2023-12-08, 05:02

[//]:content-type-MARKDOWN-DONOT-DELETE
![](https://gimg2.gateimg.com/image/1702010966539550538humingbot.jpeg)
Gate Institutional Department hosted an Ask-Me-Anything session with Michael Feng, Co-Founder of Hummingbot, on Gate Group YouTube channel: https://www.youtube.com/watch?v=a_mWK3Y8h7o&list=PLE_x6o4N6eXY28rgJA0RRQLadTq-qDbrS&index=5

Follow Hummingbot on social media
Website: https://hummingbot.org/
Twitter: https://twitter.com/_hummingbot
LinkedIn: https://www.LinkedIn.com/company/hummingbot-foundation/ 
YouTube: https://www.youtube.com/c/hummingbot
Discord: https://discord.com/invite/hummingbot

Guest
![](https://gimg2.gateimg.com/image/1702010966737601770humingbot1.jpeg)
**Michael Feng, Co-Founder of Hummingbot**

Welcome to the 17th episode of our Gate Institutional Ask-Me-Anything series, where we sit down with special guest Michael Feng, Co-Founder of Hummingbot. In this insightful AMA episode, we dive deep into the world of Hummingbot - an open-source platform designed to help you create, test, and deploy fleets of automated crypto trading bots. Discover the fascinating journey of Hummingbot, learn what sets it apart from other crypto trading bot platforms, and explore its commitment to security and user-defined architecture. Michael Feng also shares insights on Hummingbot's business model, its approach to handling market volatility, and its user-friendly setup, making it accessible even for beginners.

## Introduction
**Gokay (Host):** Hello, everyone! We're back, and this is the 17th episode of our Gate Institutional Ask-Me-Anything series. I'm Gokay from Gate.io Institutional Services. Today, we're hosting an AMA series featuring Hummingbot - an open-source platform that helps you create, test, and deploy fleets of automated crypto trading bots. I'm thrilled to introduce our special guest, Michael Feng, Co-Founder of Hummingbot. Thank you for joining our stream. Nice to meet you, Michael! How are you feeling today?
**Michael:** Thank you, Gokay, it's pretty good. It's pretty early in the morning for me since I'm in California, but thank you for having me.
**Gokay:** That's great! We're eager to hear your insights and experiences, Michael. I'm particularly excited to dive into the details of Hummingbot and learn more about it. I'm confident that you have a wealth of information to share with our audience. 
So, let's get started with our questions.

**Q1: So, tell us a little about yourself. Who is Michael Feng and what's your background?**
**Michael: **Okay, well, I'm glad you asked me because I think that it's very institutional, and traders and people appreciate my background. I was born in China, but I had the blessings to go to the USA early on because my parents went there for graduate school. I was 6 years old when I moved to the USA from Shanghai. I grew up in Florida, went to college in the US, and then started my career on Wall Street where I structured derivatives called collateralized obligations or CDOs. Have you ever seen the movie "The Big Short"? It's all about how these financial instruments are created. So, I did that for a total of nine years in New York and in Hong Kong, and I guess my claim to fame is that I happened to structure or create the model for the first CDO backed by subprime mortgages.
So unfortunately, I did have a hand in causing the 2008 crisis, and that was actually why I left finance. I felt honestly quite depressed because I wanted to get into finance to create something new, to really experiment with the levers of technology, and to enable people to do something we couldn't do before to better allocate risk and reward.
What I ended up doing was causing a lot of people to lose their homes and jobs. So I moved to Selton Valley. I got a master's from Stanford University in industrial engineering because I wanted to leave finance behind and just do something that was more on the tech side, more on the engineering side. Oh, and then I caught the startup bug, a creative startup, and then this one was where we basically tried to extract data from PDF files using computer vision. We ended up selling that company to Nitro, which still makes PDF software today. I gained a lot of experience working at Nitro and learning how to actually build a product, manage a team, and work in the tech industry.
So in 2017, when I left Nitro, I started to create something new, a startup. That's where I really found crypto, and for me, crypto was like this fusion of tech and finance. That's when things came together, and we created a company that eventually developed Hummingbot. But even then, it wasn't a straight path. In Hummingbot, the first product we created was actually a hedge fund. With one hedge fund, we actually made good money, but one day, we received a letter from the S.E.C telling us to shut it down. They wanted us to shut down because they thought we were engaged in something called general solicitation. Anyway, it was kind of a big mess, but it ended up leading us to shut down the fund and open-source our stack, which became about.
So that's how the full story came to be. For me, I think my life has been a journey from working in a very zero-sum industry like Wall Street to one where I feel like I get to work in a very positive-sum industry, which I define as open source. Basically, creating open-source projects and finding ways to make them sustainable through crypto and through the power of crypto technology.
**Gokay:** I think it's a great story. If we ever have the chance, I want to create just one episode about your background. I believe it could even be like a movie, with a big impact, and I think it would also be motivational for a lot of people.

**Q2: How did Hummingbot come to life and how is it positioned in the market?**
**Michael: **So, I mentioned a little bit about how when we first started, it was like this hedge fund and turned into an open-source project. But I would say for the first year we did this (in 2019), it was more about trying to find a way to sustain this open-source project. Because, with open-source projects, you give away code on the internet, how do you actually make it viable? In 2019, we had this idea that we could serve it by coming up with a market-making solution. It was a tool for market makers. We had this idea that if you could aggregate market makers together and then reward them on a time-based basis, you might actually be able to create a decentralized way of doing market making.
So, in quarantine, we wrote a white paper called "Liquidity Mining," a marketplace-based approach to compensate market makers. It was more like a theoretical paper, but we actually had in mind an order book-based platform that would compensate users for providing liquidity on order book exchanges. That's why we wrote the paper. We wanted to build this platform, which is now called Hummingbot Miner. Gate is one of the big partners for the platform. However, the term "liquidity mining" kind of took on a life of its own and became adopted in the DeFi world for basically the same thing: incentivizing market makers through the use of rewards they're paid every block. So, that is happening, but today, Hummingbot is more than just a tool for use by market makers.
About a year and half ago, we split off the miner platform and our parent company, Craft, was doing from Hummingbot. The reason is that Hummingbot is just open-source code, and you can do a lot with it. So, we realized that it could be used for more than just market-making strategies; you could use it for direct trading, balancing your portfolio, or any type of automated process you may want to do. In addition, today we've also expanded the user interface in Hummingbot. Now, there's a dashboard application that helps you orchestrate and deploy multiple bots. There's also a library where you can import it into your Python code and use it as a library. Overall, we thought Hummingbot could be like the Linux of crypto trading. That's why we separated the Open Source Foundation, which we now help run, from the parent company, CoinAlpha, which is into market making and runs the decentralized miner platform.

**Q3: What sets Hummingbot apart from other crypto trading bot platforms?
Michael:** I think there's actually a large landscape of crypto bots out there. So, there are open-source ones like ours. Platforms such as 3 commas, Bitsgap, as well as all the exchanges' platforms. Then, you have a new wave of Telegram bots that people are using. But to me, the core need that bot platforms provide is really about automation. How do I automate what I'm doing? How do I mirror someone's bot? But how do I not do things on my own but leverage a bot to copy, mirror, or trade automatically? So, I think there are various ways to accomplish that, and I also believe it's more about I really appreciate everyone who tries that instead of manual trading because it usually means they're like-minded individuals. However, I do encourage everyone who's running bots to try the open-source ones, whether it's us or CCXT or Quan connect, because I think open source ensures that your keys, which define your ownership of your tokens, are held securely by you running them locally on your own infrastructure instead of giving your keys to someone else and having them run bots for you. So, to me, that's the biggest difference between open-source bots and closed-source ones, which I actually put Telegram bots in that category because you don't actually own your keys, I believe. So, that's why I would say that's the biggest difference. But even within open-source bots, where Hummingbot really, I think, has tried to set ourselves apart is through community. We genuinely try to cultivate a community and leverage a bounty system that essentially maintains a code base through community developers. We also essentially run a school that teaches people how to use our platform, and we aim to grow the community. And that's what I think is the essence of this part.

**Q4: How do you ensure the security of users' funds and personal information when using the Hummingbot trading bot?
Michael: **Then the No. 1 thing is non-custodial. So, you run it on your local infrastructure. But beyond that, with Hummingbot, the first thing you do is enter a password. That password uses any key you store in Hummingbot. Additionally, because we have integrations with lots of different exchanges, this is why we have a governance system to restrict how many exchanges are on the code base. With every new exchange we integrate, there might be some library we're importing, and there might be a security vector that the library contains. That's why we have a monthly release cycle where we test every new addition in the code base. We use governance to ensure we're supporting the exchanges our community really needs and avoiding adding some random malicious exchange, which could be a threat. So, for us, we're not a protocol; we're client software. The risk profile is different. We're not as concerned about hackers who can hack into smart contracts, but we are concerned about malicious code getting in that could circulate everywhere or compromise the security of user data. Probably the most significant factor is connecting to an exchange that poses a security threat and then having users deposit funds to that exchange, potentially resulting in losses. This happened to us with FTX last year. So, for us, a lot is about identifying the best exchanges to integrate into while also keeping the bad ones out.

**Q5: How do you handle volatility and unpredictable market events? Does your trading bots have mechanisms to prevent losses during flash crashes or pumps and dumps?
Michael:** I think this is really, like I would say, the power of the bot because it allows you to automate certain hedging. For example, we have a strategy called cross-exchange market making, which allows you to provide liquidity on one exchange using information from another exchange's order book. This way, we get filled on the maker exchange, and the bot automatically hedges on the taker exchange, hopefully extracting arbitrage profit from doing so. Other features we offer include a kill switch, allowing you to turn your bot off if you don't reach a certain threshold of profit or loss. We also have a Telegram integration where you can receive notifications and control your bot through a Telegram chat. There are various ways to manage your bot, but overall, I think it's more about leveraging technology, like a bot, to help you defend against volatility and uncertainty even when you're fast asleep.

**Q6: How much technical knowledge is required to set up and manage the Hummingbot trading bot? Can a beginner get started easily?
Michael:** So, I would say, it's definitely not for everyone. I wish it were, but I think there is a class of people, regardless of how technical you are. It's really about whether you want to take the time and effort to automate your trading because it is a bit of an investment. But we do see people who don't have any technical knowledge at all, who are able to become fairly successful. That's because now it's not just a command line tool; we have a dashboard. As long as you can get it installed, it's basically an application running that you can use from the browser. However, it does require you to use Docker to set it up. So, as long as you can understand how Docker works and set it up through Docker, I think even someone who's not a developer can still use it.

**Q7: Can users customize their trading strategies with Hummingbot?
Michael:** It's definitely, we have a few different frameworks for how users can define their strategies. Initially, we had strategy templates, for instance, for arbitrage, where you input the exchange, the trading pair, and the desired profitability. We just ran this predefined strategy and found that it was not very flexible. We had various strategy templates, but some of them added many different configuration parameters. So, we also introduced the concept of _script_s, which are simple single files that define a straightforward process. Now, we have a library of different _script_ examples that people can copy and customize.
Currently, we have a more powerful framework that we call V2 strategies. These are designed to be back-testable using our dashboard back to the engine. This is also what we teach in our bootcamp because we believe that to create back-testable strategies, there needs to be a cohesive connection between the dashboard, the engine, and the application. I don't want to delve too much into the explanation, but overall, we aim to create an easier way for even non-technical people to back-test, create, and deploy strategies that can compete at a professional level with market makers. Hopefully, our tool, the Hummingbot dashboard, and our school will elevate people to that level.
**Gokay:** So, as I understand, everything is ready; you just need to put in a little work. Even if there are challenges, and everything seems daunting, if you, let's say, maintain a consistent effort and dedicate your time, I believe you can achieve great results.
**Michael:** Alright, and exactly that's one more thing. I've seen many people in our community do this, where we literally have people who have no technical background, no training experience, coming by, and they made over a 100% return in the past year. Part of this is because of understanding which markets you want to play in, which markets to avoid, and just finding the corners of the market where you can earn some reward by running a bot and exploiting that for as long as possible. Okay, so I think many people try to do this successfully.

**Q8: Can you share any specific examples of how Hummingbot has successfully maximized trading income? 
Michael:** So, we just started doing this, but we're going to have monthly trading competitions in our community. We essentially have a small prize, and it's primarily for educational purposes. Everyone starts with $100 on one exchange, runs bots for a two-day period, and submits the results. Based on the volume they generate and the money they make, we provide them with some rewards. We also recognize and reward those who write up their strategy and share it with others participating. We recently completed this, published a blog post about it. And from my experience, I ran a market-making bot. I was pretty sure I would lose money, and I did. However, some other participants ran their own bots, and some of them did exceptionally well.
I believe it's a good example of what trading is like. Many people think it's either about making a significant profit or losing a lot of money. Most of the time, it's a random distribution of results. Observing this in action has significant educational potential for both those implementing the bots and those on the sidelines.

**Q9: How has Hummingbot's experience been so far working with Gate.io?
Michael:** Honestly, it's been great because we've had a deep partnership with Gate.io ever since we started. Gate was one of the supported exchanges on the Hummingbot platform, where people can run Hummingbot and earn rewards by market-making for various projects. Also, because we have a fee-sharing agreement with Gate, anyone that runs Hummingbot on Gate helps support the Hummingbot Foundation, as a portion of their fees will come back to us. This comes at no additional cost to the users. So, for us, we're always encouraged to do more to expand the usage of Hummingbot on Gate. Right now, we actually maintain a dashboard that shows how many bots are running on Gate.io.
Just on the spot, we've seen 21 million in volume over the past month, averaging about a million per day. At any given time, about 80 to 100 bots are running. So, there's already a lot of usage, but I think hopefully through initiatives like this and through education, we can encourage people to do even more.

**Q10: What role do you see crypto trading bot platforms like Hummingbot playing in the future of trading?
Michael:** To me, I believe our role is to democratize high-frequency trading (HFT) for everyone. This is something that, when I was working on Wall Street, felt like a situation where major institutions like Morgan and Citibank had a significant advantage over average everyday traders due to their access to computers and sophisticated systems. However, the technologies, algorithms, and systems that were once proprietary have become more open today, moving towards the power of open source.
So, I see Hummingbot not as a competitor to other open source projects but as a collaborative effort. We work alongside them to expand and educate people on how to engage in trading effectively.

**Q11: And finally, what is the most important point you want our audience to take away from today's AMA session? 
Michael:** I know I've mentioned it already, but I think people often perceive algo bots or bot trading as a kind of black box. It's like they believe that finding the right black box might be the key to winning the lottery and making a ton of money. Conversely, if they stumble upon the wrong black box, it's as if they could lose all their money. However, I would say it's a bit the opposite, isn't it? It's almost like you're building and running a program; the outcome is defined by what the program does. Instead of viewing it as a mysterious black box, I encourage people to genuinely learn to code, understand how the program works, so that they can control its actions. By doing so, you can essentially program it to do what you want.

## Ending
**Gokay:** Michael, thanks a lot for spending your time with us. Your insights have been truly informative, and I'm sure our viewers have a better idea about You and Hummingbot now. I really enjoyed our conversation and learned a lot.
Feel free to reach out to Hummingbot, visit their website, and delve into further details.
Before we end our session, Michael, would you like to add anything else? 
**Michael:** Nothing else I would say to us; I think the main thing that supports us is running, as I mentioned, bootcamp, our school for aspiring market makers and algo traders. It features monthly bot battles, training competitions, as well as new courses and content that we're adding each month. Our goal is really to use that to support the foundation, support Hummingbot, but also to expand the knowledge of trading
**Gokay:** Thanks to everyone who watches our AMA series. Follow us on social media, for more information write to our key account managers.

Twitter: https://twitter.com/gate_io
Gate.io Institutional Twitter: https://twitter.com/Gateio_Inst
Telegram: https://t.me/gate_zh
API Telegram: https://t.me/gateioapi
LinkedIn: https://www.LinkedIn.com/company/gateio/
Webpage: https://www.gate.io/institution

<div class="blog-details-info">
<div>Author:**Hummingbot**, Gate.io Researcher
<div class="info-tips">\*This article represents only the views of the researcher and does not constitute any investment suggestions.
<div class="info-tips">\*No content herein shall constitute investment advice and does not constitute any offer or solicitation to offer or recommendation of any investment product or project.
<div class="info-tips">\*The inclusion of any third-party products, services, entities, or content does not constitute an endorsement, recommendation, or affiliation by Gate.io. Please conduct thorough research and seek professional advice before making any investment decisions.
<div class="info-tips">\*Gate.io reserves all rights to this article. Reposting of the article will be permitted provided Gate.io is referenced. In all cases, legal action will be taken due to copyright infringement.

</div>