Introduction The Decentralized Finance (DeFi) Revolution is taking over traditional finance (TradFi) and centralized exchanges (CEX), with billions in...
CEO Kory Hoang was interviewed by Aaron Fifield from Chat With Traders in December 2017. Kory shared his story as a part-time system trader as well as some valuable experience for those who want to begin their trading career.
RISK DISCLAIMER: None of this is financial advice. Investing has many risks, including risk of loss beyond principle investment. Past performance is not indicative of future results. None of this is recommendation to buy or sell any security, commodity, cryptocurrency or any other financial instruments. Please DYOR.
(Source: Chat With Traders)
— — — — —
Transcript of video below, there may be some grammatical errors.
AARON FIFIELD: Chat With Traders, episode 152 has backing from online broker and technology provider TradeStation. Now folks, there’s a good reason why TradeStation clean up when it comes to Awards, because they know what traders need and they deliver from decent rates to powerful technology and analytical tools to helpful support. TradeStation are 1 broker you should not overlook. Learn more at TradeStation.com/traders. Markets, speculation, and risk. This is the chaplet traders podcast hosted by Aaron Fifield. Hey crew, what is going on? Your host Aaron Fifield here coming up on this episode. What you’re about to hear is a conversation I had with Kory Hoang. Kory is not a veteran trader—he’s not someone who has been doing this 10-20 years. He’s someone who has been doing this for only a few years, yet he’s begun to make decent gains on his trading capital. Kory is also not a full-time trader. Well let me be clear. At the time we recorded this, a few weeks back, he wasn’t, but I’ve since been told, he’s handed in his notice and has taken the leap. So the point being for the average listener there are quite possibly a few similarities between your situation and Kory’s situation, so how does he trade? Kory is a retail systematic trader. He’s running numerous algorithmic strategies, which he’s developed (all of which are fairly simple). These run on various ETFs ETNs and even some cryptocurrencies. Join our chat, we cover quite a bit but mostly his journey and how he’s progressed to this point. There was a moment around midway where Kory did lose me a little bit, if you happen to get lost – just stick with it, because it will all click and I then take a moment to summarise his point also. As Kory talks about a few strategy specifics, I’d like to just remind you that you are entirely responsible for your own trading decisions, okay? Enjoy the episode team. This is Kory Hoang. Where’s a good place to start? I mean it might be helpful if we just hear a little bit about kind of what you studied at University? Did you go to university?
KORY HOANG: Yeah I went to the University of Washington. I took a business degree and I major in finance and marketing.
AARON FIFIELD: Okay and what did you do once you… once you got your degree?
KORY HOANG: So while I was still in school to get my degree, I actually took an internship at Merrill Lynch wealth management in Everett kind of by Seattle. And that’s where I really got my foot in the door in the financial industry and the investment industry in particular. And that’s where I learned a lot about… you know, stock market, financial market trading and all that and… you know, became really fascinated to me ever since that day.
AARON FIFIELD: Okay so what were you doing at Merrill Lynch while you were there?
KORY HOANG: When I was an intern for three months during the summer in Derek Wells management division, I was tasked mainly with basically shadowing other advisers there and… you know, one of the guys there, I remember him to this day, his name was Kelly. I remember I walk into that office one time, and he was in this big glass office facing… you know, this was on the 10th or 12th floor of a building, and he was looking… overlooking the ocean from the port of Everett. And you know, I look in his office and there was… you know, screens and charts and all these cool indicators and everything I knew from that day. And I want to just be like that guy.
AARON FIFIELD: So that obviously sparked your interest in trading, that experience. Was your internship there… did you find that to be any help once you started to get into trading yourself?
KORY HOANG: It did help me get kick-started in terms of… you know, finding out what’s a research… you know, what kind of stocks to look at… you know, what, for example, stock sectors, market sectors are, and things like that versification. Then you know, a lot of… you know, basic beginner stuff for investing in general, so that did help me get started in the right direction. And from I just basically self research and… you know, looking up information online, reading books and figuring out the rest from there on my own.
AARON FIFIELD: Okay okay, so just to put this in perspective, when did you actually start to get interested in trading or actually start trading, so a few years back right?
KORY HOANG: Yeah so my internship Merrill Lynch was actually three years ago, and that was the same time that I actually started trading. You know, I was still in college that time. I had a little spare money, just a couple grand. I started to trade… you know, technology stocks, high growth stocks. I remember this one time, I bet money on KNDI Kandi technologies. It was a Chinese clean energy vehicle company. But a long story short, I made some good money on it initially, made a couple grand off of it. But then… you know, that company, it just didn’t do really well, and then the stock kept tanky and I just kept trading it and didn’t go well at all, and so start losing money and had to take a break after a while.
AARON FIFIELD: So what… what we basing your decisions to buy and sell on back then? Like was there any methodology to what you’re doing or did you just have like a… a good feeling the stock was going to give you a nice return?
KORY HOANG: Yeah to be honest back now is really just gambling. Yeah to be honest, it’s just basically seeing what’s… what’s moving today… you know, what’s the top market movers and what are the major themes in the market, today. One of the things I was researching back then back in school was clean energy and electric vehicles and… you know, so I thought I venture into that space and bet on a couple of stocks from that sector, so there wasn’t really any methodology really. I mean I did look at charts a lot, but I never really found them to be extremely effective.
AARON FIFIELD: Okay so why did you feel as was gambling? I know you said they let you feel like you were probably lacking a bit of methodology, but was that the only reason or was there we bet him too big? I know you said you made some profits and one it was a Kandi or something like that and then it went right back against you just kept holding. I mean what was some of the reasons why it felt like gambling in the beginning?
KORY HOANG: Well I felt like gambling because I did not follow a system. I mean I would think that I have a system, but I would not follow it rigidly, and… you know, it wasn’t… it wasn’t a scientific process. A scientific process mean that you come up with an observation and then you create a hypothesis and then you go out and you do your experiment, you collect your evidence and from then you analyze the result and make a decision… you know, is your hypothesis supported by evidence or not. That wasn’t exactly what I was doing. What I was doing is akin to what a lot of traders that I see… you know, when they just start out and trading or investing. They just have some sort of ideas, some sort of intuition or maybe they heard someone say, oh this is a good stock or maybe they read somewhere that… you know, maybe they should buy a stocks when the RSI goes to oversold. And they would just have some sort of idea of how to trade and they would… you know, think it’ll make sense and they would go with it without really testing out that idea… you know, may perhaps through a back test or for testing with paper trading or something like that. They just go right at it. It’s just like gambling.
AARON FIFIELD: Okay and I just want to ask you before we go too much further, what were your expectations? Like when you started out trading like did you feel as though this could be an easy ride and an easy way to make money?
KORY HOANG: You know, initially I thought I could really day trade and make just 1% a day… you know, and I’ll be good for the rest of my life. And it was just really… you know, naive beginners thought process that I’m sure a lot of people went through when they first started trading. But yeah expectations were high and sure enough it was soon crushed by the reality of trading in the market.
AARON FIFIELD: Mmm… now what caused you to start getting interested in algorithmic trading? I feel as though you may be touched on it before when you started to talk about having more of a scientific approach to the types of strategies you trade, but… you know, you started out as a discretionary trader, you started out just trading stocks and sectors that you knew a little bit about, what caused you to start gravitating towards algorithmic trading?
KORY HOANG: It was in 2015 and there was a TED talk with Jim Simon’s from Renaissance Technologies. And I was very intrigued by that TED talk because here we have… you know, someone who’s a mathematical genius. He’s… you know, had a great career in academics and not only that, he used his knowledge from there and went into the financial industry, started a hedge fund and managed to crank out… you know, the best returns that anyone ever seen in the market so. And he did it all by using algorithms and following rules-based processes so to me that was what I needed to do. Because I saw that as… you know, a scientific process that he was doing and to me, he was a scientist and he was tackling trading and investing in a very scientific approach. And I figure… you know, if I started imitating that process myself, I would hopefully be able to achieve the same sort results.
AARON FIFIELD: Yeah I have seen that talk actually. It’s a really good one and I’ll dig up a link to that and put it in the show notes if anyone wants to also watch that. It was that the first time you’d heard of Jim Simons?
KORY HOANG: Yes that was the first time I heard Jim Simon’s (AARON FIFIELD: Ok), wasn’t the last time. He came out in the news… you know, many times that how… you know about how his Medallion fund is like the best performing hedge fund of all time. It’s just insane the kind of returns that he cranks out.
AARON FIFIELD: Yeah, there’s a really interesting article on Bloomberg about the Medallion fund. I presume you’ve probably read that. It came out probably as a few months ago earlier in the year, so you’ve watched this talk, you’ve realized that you kind of like the approach that Jim Simons has taken… you know, very algorithmic rules based as you described. What did you do from there like what resources were helpful for you to begin learning about algorithmic trading?
KORY HOANG: I would say the most useful tool that I ever came across was Quantopian. For… you know, some people who don’t know what Quantopian is, it is an algorithmic trading platform. There’s open source actually. It’s not a trading platform anymore. They disabled live trading recently, but you can still do a lot research on this platform. It is based on Python and you can code your own algorithmic trading strategy there and they have a forum where a lot of Quants and a lot of people who pursue that subject post a bunch of… you know, algorithm and things that they come up with and we will share idea and collaborate together on strategies. So it was a great place to learn and I started out learning just like everyone else who doesn’t know how to code. You just kind of like copy and paste and put thing together and just kind of like making a Frankenstein algorithm and… you know, run a backtest on it and see if it would work. And you know, you do that day after day, week after week, some months after months, you start to get better and better at it. And at the same time Quantopian offer a course. They had lectures on how to build Python based trading algorithm. And you know, it it’s literally like a full university course on quant trading. That’s my impression on it. I know to some people it might not be very advanced but to someone who did not really have that kind of exposure, it was a lot of help and it kick started my journey into algorithmic trading a lot.
AARON FIFIELD: Yeah their lecture series is very top-notch. I mean I would consider certainly some of the topics there to be quite advanced. Yeah and I love that feature about Quantopian how you can like just clone other people’s algorithms and then play around with them, tweak them, try change a few things, and… you know, see if you can improve on it. And yeah… and they also have a really good forum as well. I mean I don’t normally I’m not really much of a fan of forums generally but Quantopian have a really quality forum like you can… you can find some interesting insights there, so did you know how to code in Python prior to this because as we know Quantopian’s platform only takes Python as the input language? I mean had you had any experience coding in this language before?
KORY HOANG: Absolutely not. Before I discover Quantopian, I did not code at all. I mean the closest that I do to coding was using Excel spreadsheet. And back in college, most of my courses were mainly business and finance courses. None of it was for programming, so if I could go back to school, I was…. probably one thing that I would change to take some programming classes.
AARON FIFIELD: Absolutely I mean that’s one thing I wish I would have picked up a lot sooner how to code.
KORY HOANG: Hey we are. But yeah, so with absolutely no coding experience, I just kind of dive head in and… you know, it took weeks and weeks before I finally figure out how to get basic features running like running a full backtest and factoring in commission & slippage and how to change stock symbols and all that stuff. It took a while, the learning curve was… you know a little bit steep at first but… you know, this was something that was really passionate and that I really enjoy doing. So to me, it wasn’t really… you know, something that was strenuous. It was something that… you know, I was looking forward to doing. I remember staying till… you know… 2 or 3 a.m, I’m just coding algorithms and learning how to use the platform.
AARON FIFIELD: Yeah it can be kind of addictive in some ways, right?
KORY HOANG: Yeah certainly, what, because you know once you understand the capability that this tool can offer you and all it takes it for you to learn it and get better at it… you know, you just kind of get into that mindset and just push yourself to do it.
AARON FIFIELD: Yeah, yeah totally man. Now were you using Quantopian to actively learn Python as well or were you like doing some other online courses and classes, etc from other places on the net?
KORY HOANG: I did take the course from Udemy, but I never really finished up with it so my experience with Python so far is… you know, I’m not like the average typical programmer who… who knows Python and cant code like an app or… you know, some sort of interface for you, so I’m not that kind of guy. I strictly know how to use Python on the Quatopian platform only and my break really came when I discovered TradeStation because TradeStation uses a much simpler coding language called easy language and TradeStation has like charting platforms and interfaces that… you know I was much more accustomed to. Quantopian was very much tech space and code base and very little charts. And I mean you can code them yourself but for me to do that at that point what it was a huge learning curve. So when I found TradeStation and I’ve discovered… you know, the ease of coding in easy language, I started playing around with that platform a lot more and I started migrating my trading over to TradeStation.
AARON FIFIELD: Okay and I presume it points throughout this whole process. There’s been… there’s been points where you’ve kind of got stuck or hit a wall and got maybe frustrated because you can’t work something out. What are you doing those sorts of scenarios?
KORY HOANG: Well, what I usually do when I’m stuck at something is I find someone else who can help me figure it out, someone who actually know how to code. At first it was a bit difficult because usually the people who knows finance don’t know how to code, and usually the people who know how to code don’t know finance, so to find someone who has some sort of quality of both was very difficult. I remember trying to work with programmer who have no financial knowledge and you know it was ok and I try to teach them… you know, what I did know and… you know… my… what I wanted my strategy to be and… you know, try to guide him through the process, but it was very difficult… you know, is… it’s basically another learning curve for them that they had to tackle on, and for the amount of money that was paying them, they obviously were not interested in pursuing that route. So… but you know, eventually when… when you have a forum like Quantopian, you can find collaborators fairly easily and… you know, when I start using the forum there to reach out to people, to coders to help me with my strategy, that’s when I really found that’s when one of my first breakthrough really came. When I found someone to help me implement one of my strategy that I had in my head, it was just a concept and he helped me lay it out in Python and I was able to observe like how the Python code was structure, and I basically reverse engineer and learned it back from there.
AARON FIFIELD: Okay yeah, I mean that’s… that’s a smart way to go about it in the forum. You know, a lot of people really helpful there so… you know, if you have questions, it is a good place to find some answers. How long did it actually take you to go from doing back tests and different forms of analysis to actually starting to run some real money through your strategies?
KORY HOANG: So I began coding strategy and… you know, trying to back testing since, I think, start 2016 and from then, it took me all you know about six seven months to finally get comfortable enough, throw some money at the algorithm that I have code it and… you know, it wouldn’t be a lot of money. I’ll just be like five or $10,000 and I wouldn’t even trade the whole amount… you know, just make the algorithm take very small portion of that money. But you know, it was enough to test out my idea to do some walk-forward testing. And from then on… you know, it helped me build more confidence because… you know, I can see that my money’s being put to work by this machine that I built myself and knowing how this machine works and seeing it in action after… you know weeks and weeks and weeks build up your confidence to the point where you realize that… you know, this is… this is real. You know, like I always read about that’s, always heard about this and… you know, now I’m actually experiencing it so…
AARON FIFIELD: Right, do you remember your first automated trade?
KORY HOANG: You know, to be honest, I can’t remember the first one but I remember my first biggest win from automated trading, it was in the… it was in the summer of 2016. I forgot what happened but that week, my algorithm bought XIV and inverse volatility ETN. I bought it… you know, at a really good time and I saw it… you know, about week later and make some good money on that. And from that on where… you know, I was hooked.
AARON FIFIELD: Yeah yeah. Now it’s a good feeling man, so when did you feel confident to actually go live with this first strategy because… you know, I presume there was… it… of course it wasn’t the first strategy you tried? You tried to like in… what’s it, without trading real money right? So like it backtest in different ideas and tried different things but this particular strategy which you decided to go live with, you obviously felt confident in that, like what gave you the confidence to go live with that particular strategy?
KORY HOANG: I see. When I… when I first created a profitable algorithm I knew it was profitable because it was a very simple algorithm. It’s literally a moving average crossover but you know the asset and the timeframe that I was using for it just work out perfectly. And you know… so just just because I understand how it works and the simplicity behind it, it gave me a lot of confidence that… you know, start going forward.
AARON FIFIELD: Okay and I know some people, I think probably freaked out about the idea that a computer is trading their money for them, right? Was that an issue for you like… did you struggle with that aspect of it, just let taking your hands off and letting the computer take over?
KORY HOANG: To be honest, I think I had it easier down a lot of people because when I started building my strategy and creating my algorithm, that’s was when I started researching into… you know, market anomalies and taking… you know, really quantitative approach and research into… you know, how to figure out what assets and how to trade them. And you know, I think it was luck perhaps, but the framework I develop it ended up working out really well. And I’m still using it to this very day and… you know, all the algorithms that I built and create are from this framework that I use and it is rather simple if I really break it down to you due to its simplicity. It is very robust and… you know, it’s been more than one-and-a-half years of walk-forward testing and live trading now, so it’s giving me a lot of confidence to keep moving forward.
Ad: This episode of Chat With Traders is brought to you by TradeStation. When you become a TradeStation client, you’re putting yourself in a position to better analyze and trade markets, especially as a retail trader. As traders, we have enough challenges as it is. So roll with an online broker who understands what traders need. TradeStation equipped their clients with awesome commission rates and professional grade analysis tools which you won’t find being offered by other brokers. Just as an example, TradeStation allow you to create your own custom indicators and your own custom strategies, then back test these ideas on large amounts of historical data and even automate them. If that’s what you’re into, take this hot tip discover what it’s like to trade with TradeStation by visiting TradeStation.com/traders and open an account. Also supporting this podcast once again is HealthIQ. HealthIQ believes that the best way to improve the health of the world is to celebrate the health conscious through social and financial rewards. By using science and data, HealthIQ secure lower rates on life insurance for healthy folks like runners, cyclists, gym junkies, vegans etc. Here’s a fun fact which should grab your attention: 56% of HealthIQ customers saved between four to thirty three percent on their life insurance. These savings are exclusive to HealthIQ. Just like how you saved money for being a good driver, HealthIQ are here to get you lower rates on life insurance for living a health conscious lifestyle. To see if you qualify, get your free quote today at HealthIQ.com/Chat. That’s HealthIQ.com/Chat to get your free quote.
AARON FIFIELD: How many systems or strategies what are you gonna call them are you running nowadays like at the moment?
KORY HOANG: So when I do these things, I have two different types of algorithms that I run. The first one is my binocular, so this is an algorithm that I develop based on my observation and research into market anomalies. I call it my price anomaly detection algorithm. So what this algorithm would do is it would scan all available securities in the market and based on a proprietary metric that I developed, it would classify which asset at which time interval exhibit price anomaly, meaning this has mean reversion properties or if it has momentum or trend persistence properties. And of course this system would also tell me… this algorithm would also help me identify assets that random walk and according to my framework, I don’t trade random walk assets. My algorithm indicates to me that about 96 95 percent of the available assets and time interval was out there are random walk. And it is not profitable for… you know, you, for most people to come in and trade those assets, because they don’t have any edge, because they’re random. You can’t trade something that… you know, random. But about 5% up stuff out there, they do exhibit certain trends and anomalies that show up time over time. And it’s those assets that you want to focus your trading on and this is when I start using my second set of algorithm. These… so the first one was the binocular, the price anomaly detection algorithm, the second set of algorithm, I call them my rifles, so this is when after I have identified my prey like a hunter, I would take my rifle, some very simple rifles at time then I would go and I would hunt down my prey to put to… put it in a… you know, in a simple to understand way.
AARON FIFIELD: Okay okay, now I just want to go into a few of those things you said there okay, so I guess the first one being framework… you know, you’ve used this word a few times. I just want to make sure that we’re clear on what you’re actually referencing when you talk about a framework, so what are you referencing?
KORY HOANG: Okay all my research so far ever since I started investing, trading 3 and a half years ago indicates to me that the market is mostly efficient. If you study… you know, French and Fama Modern Portfolio theory, they say that prices, asset prices are efficient because the market discount all available information into prices almost instantly so… you know, the best thing you should do out there is really just buy and hold on index fund and just let it ride. So my internship Merrill Lynch kind of had me subscribe to that idea for awhile until one day, I realized that… you know, perhaps that’s not the whole truth out there. Perhaps there are certain pockets in the market where inefficiency still exists. And so you know, it was at that point I realized that efficient assets are random, because when they are efficient, there are no extractable alpha or there’s no edge in trading them. Therefore that’s why they are efficient, and they are random. You can’t predict these things. However, certain assets out there are inefficient meaning there are repeatable patterns that keep showing up and people are not exploiting them at that pattern or what I call anomaly is large enough for you to explore yourself. They just definitely do it so that’s basically what my framework is about… you know, I used the price anomaly detection algorithm to scan all available asset… you know my binocular to find my preys. The ones that… that are random and you know, that can’t be hunted down, I disregard them and I don’t trade them, but the ones that do show a high level of anomaly that you know is exploitable, then I would come in and trade those assets with my more optimized algorithm.
AARON FIFIELD: Okay, so in some ways, this binocular algorithm, I guess you could call it, that doesn’t actually take any trades, it’s purely just a scanning tool essentially?
KORY HOANG: Essentially yes, so what this algorithm does is it uses an oscillator, a technical oscillator to measure the expectancy of momentum or mean reversion for any particular asset at a certain time interval. For example, I can use this algorithm to scan for… scan all ETF on a daily interval and at the end of the scan, it would produce me a market map. This market map it looks like an x and y plot. The x axis would be for momentum expectancy and the y axis would be for mean reversion expectancy, so in the middle up the plot where the two axes intersect is zero. So in the bottom left quadrant, it’s the random walk zone, that’s where 95% of assets tend to end up, because they’re random and unpredictable. However in the top left quadrant and the bottom right quadrant, those are the quadrant where you would find momentum and mean reversion anomalies where assets would exhibit certain trends that are persistent throughout time and assets that fall into those zones, I would assess them and find a winner off for me, the highest risk reward ratio and then I would go in trade just those asset, only maybe four or five of them. And these tend… these usually tends to be ETF, so you know, it’s been pretty good for me following the system that I develop.
AARON FIFIELD: Now the price anomalies that you’re talking about, can you give us just maybe an example of one anomaly which you might use?
KORY HOANG: Certainly so when I’m talking about anomaly, I’m actually just talking about two. And it’s momentum and mean reversion. Matter of fact that how I got started thinking in this direction is when I went to Quantcon hosted by Quantopian in 2016, I think was April in New York, so I flew out to New York and I met up with some people there. And you know, we went and we saw a couple of lectures and talks by some of the best… you know, people in the industry, some of the brightest quant out there. And there was a… there was a person who came all the way from India. His name was Manish Galan of I think SG analytics. Yeah and he came all the way from India just to give a talk at Quantcon. And so I… you know, I figure I’ll go to his talk and see what he has to say and it was at that talk that he opened my eye to price anomalies. It wasn’t really… you know, fully a hundred percent clear to me at that point but… you know, afterward after went home and taken to consideration what he talked about during his lecture there, I finally understood what he meant. And from then on, I developed the rest of my framework from there. So when I was in that talk, he showed me what a topic that he touched, I was… you know, figuring out how to trade certain assets and certain style like you can’t use the same trading style for every single asset. And so he put up on the projector a picture of an RSI oscillator. And so he… I look at the RSI… you know, everyone’s familiar with the RSI indicator, and so I look at the RSI and he said if I was to buy when the RSI is oversold and sell when the RSI is overbought, what technique that is used and you know, for a while everyone was like stumped. And then he said that is a mean reversion technique and I thought about it. I was like… you know, what that makes sense, you know, if you buy when it’s oversold and you wait for it to bounce back and then sell it when it’s overbought, you’re waiting for it to mean revert. And then, he also said what do you think a momentum technique would be, I… first I couldn’t figure it out, but then it was very simple. He said a momentum technique would be to buy when it’s overbought and sell when it’s oversold, the exact mathematical opposite of mean reversion. And so from my experience from that lecture there, I went home and thought about it and I realized that… you know, what he is right and I started doing some more research and trying to come up my own evidence and support.. you know, what he had to say and in the course of looking for these evidence, I stumble upon these anomalies that is momentum and mean reversion in the market. Certain assets you can use momentum technique on them and it would work really well, but what that means is if you use mean reversion techniques on these assets, you’re gonna do very poorly. And so also for certain assets neither mean reversion nor momentum technique would work well on them and… you know, those asset tends to be random walk, because… you know, they don’t show a persistent trend neither a momentum a mean reversion trend for you to follow. So that is basically what my entire framework is based on identifying which assets shows momentum properties, in which asset shows mean reversion property. Identify the one that exhibit the highest level of those anomaly and trade those asset only, meanwhile avoid trading all the random walk assets.
AARON FIFIELD: Okay now I don’t want to confuse things here, but the example you gave about the RSI between what makes what might be a mean reverting characteristic and what might be a momentum property or characteristic, you said if the RSI is pointing to overbought and you sell that, that’s a mean reversion type of trade, if you… you can look at it the opposite way, and it’s a momentum type of trade, so how does that work? Because doesn’t… wouldn’t it give you a signal to also go long as well as go short the same stock?
KORY HOANG: so this is where this is the part where you have to identify which asset exhibit momentum anomalies and which asset exhibit mean reversion anomaly. For example, I’ve done this myself so I know SPY on a daily interval exhibit mean reversion property. Right now it does. Back in the 1960, it did not exhibit mean reversion properties. In the 1960, have you bought the S&P the day after it was… it has gone up, after it has gone positive, and you hold it and sell it the data it finally starts going down, have you followed that basic momentum system, you would have made… I can’t remember the number but you would definitely have outperformed the market many… many times. But sometimes around the 1980s and 1990s that system stopped working, however the reversal of it started to work and that was mean reversion. So whenever the market started the S&P start going down, the data start going down, have you buy it and then you sell it, the first data start going up that system when I started making you money. So because I did this research myself and I saw it in the S&P 500, I figured this anomaly must also be… it must also exist another asset and so I went out, I scan all every single one of them and sure enough, I was able to identify certain ones that have very strong momentum anomaly, couple of them are inverse volatility products like XIV SVXY, jr gold miners etf like GDXJ exhibit very high momentum anomaly on the intraday interval. Also junk bonds this was kind of surprising but on the two our interval, junk bond HYG the etf exhibits momentum anomaly, so if you use a simple RSI system where you would buy it when it is overbought and sell it when it is oversold, you would make decent money over time trading those assets.
AARON FIFIELD: Now you are scanning this… a scanning these different symbols every day or is this just what you’ve done in your initial research and then you’ve traded main reversion on that those particular names?
KORY HOANG: Yeah right now, I’m working on a system where it would be constantly monitoring it. But for now I’m just taking a snapshot every quarter so at the start of each quarter I’ll run the scan and I’ll run it on many time interval for example run the scan on all ETFs on a 30 minute interval, in an hour and then 2 hour and then a daily interval and then perhaps weekly interval, monthly, quarterly at the most. And so after I run to scan at each time interval, I would get a bunch of result. And so I would get that XY plot for every single time interval and on these XY plots on each time interval, it will indicate to me which asset is currently exhibiting momentum or mean reversion anomaly. Once I figure out… you know, out of all those, which one gives me the best odds then I go after it.
AARON FIFIELD: Okay, I think this is all starting to make sense and come together. So just to summarize what we’ve been talking about here, every quarter you run this scan through your binocular algorithm, right, and it determines which symbols, which stocks, or ETFs, or whatever it is you’re trading, has strong characteristics of mean reversion or strong characteristics of momentum, and then you take kind of the correct their strongest out of the strong and then that’s what you trade your other strategies on. So if it’s… it shows strong properties of momentum, you’re gonna trade those names with momentum strategies, is that correct?
KORY HOANG: Yes, a lot of traders, they tend to think that… you know, today the secret lies in a strategy, you know, in the perfect indicators and the perfect rules… you know, the perfect money management system. I mean all that play a part in becoming… you know, in creating a successful strategy. But from my experience that’s not where the secret lies it’s not in the system that you trade, it… it lies in the asset. If the asset is random, there is no way you can ever try to make it work for you. I mean personally, I have not found a way yet. I shouldn’t say never, but… you know, personally I haven’t found a way to trade random assets, but when you found when you can find certain assets that exhibit certain patterns that keep repeating over time… you know, you need to be… you know, oh, you got to be a really chicken, a really big chicken so now go out and take your chance at that. Because they keep showing up over and over and over again.
AARON FIFIELD: Okay, yeah now this… this makes sense, I’m following with you, I was a little bit lost for a moment there, but ya know, it’s… it’s very clear to me now. I might just say if anyone has any questions around this, would you be open to answering those in the comment section of this episode on the website?
KORY HOANG: Yeah, so I wouldn’t necessarily share all my secrets but… you know, I from my experience with Quantopian, other people have been very willing to help me and help me get started to where I am today. So you know, I’m just trying to return the favor and doing the same. Obviously I have a day job and I’m trading my algorithm trades on the side so… you know, I don’t have a lot of time to cater to everyone, but I try to answer it as many emails I can if that makes any sense.
AARON FIFIELD: Of course, yeah I know. We obviously would be respectful of your time, but I’m just thinking if this part wasn’t clear to someone… you know, I don’t expect you and I don’t think anyone expects you to give away the secret sauce. Yeah just if they need a little bit of clarification around your framework there but anyway, let’s keep moving so let me ask you this: How do you come up with strategy ideas? Okay so once you’ve classified different assets into either being having mean reverting characteristics or momentum characteristics, how do you actually come up with the ideas for how to trade those assets?
KORY HOANG: Now this is a little bit interesting because I draw from my own personal life experience to… you know, come up these strategies. So just a little bit background with myself. I was born and raised in Vietnam. I actually wasn’t from the United State. I came here in 2004 with my family when I was 12. So for a good part of my life, I was… you know, growing up under communism, under a system that was… you know, a hybrid between communism and capitalism to put it that way. And when I was in Vietnam I had the chance to learn a lot about the Vietnam War that was actually one of the major topics that you had to learn about in school. And I had to learn about the tactics that were used during the war, I don’t know why they were teaching kids those kind of things back then. But you know, they did it anyway so you learn a lot about guerrilla tactics. And what you know, the Vietcong, the Vietnamese communists used to against the Americans during the war and some of that theme, it kind of stuck with me for the rest of my life throughout everything not kind of do guerrilla tactics that makes any sense. So for example, one of the things that a guerrilla would do during that war is they never openly confront the enemy who is superior to them in open battle, right? So that’s kind of like me the end of a time trading I… when I create strategy, I don’t create a strategy to try every single thing out there, because I know that I don’t have the capability or the skills to do that. I want to focus my trading on you know things that I know that I’m really good at. Just like guerrilla tactics, you don’t go out, you fight everywhere, you focus your… you know, on the most vulnerable target and on the most… the one that do the most damage and put the best bang for the buck. And you know… keep my, I keep my strategy is very simple just like… you know, when during that war… you know, we had two sides. One was like technical… technologically hundred times superior and the other side was… you know, fighting in jungles and using bamboo trap. But you know, we all know how that war ended up. And simplicity at the end of day, one that war, a idea of other reason so that’s why I keep my strategy very simple. I try not to over complicate it, a lot of my strategies are built which is a couple of variables, just a couple of inputs, maybe two three less than forty, fifty line of codes at best. It’s working out really well surprisingly… you know, something so simple and… you know, not very complicated could deliver such robust result over time.
AARON FIFIELD: Now most of these strategy art is revolve around indicators?
KORY HOANG: Most of what I use is technical indicators and I don’t use it like a lot of what people call technical analysts. Technical analysts tend to look at a chart and draw some sort of trend line or see some sort of patterns emerging or maybe they’ll look at a technical oscillator and like the RSI and they would come up with some sort of a subjective idea about it. Now I don’t want to say that… that doesn’t work or it won’t ever work but in my experience, when you follow that route, there’s a lot of subjectivity and a lot less science, and it… it becomes more an art than a science. So when I use technical indicators, I… you know, don’t… I also don’t draw trend lines, I don’t… you know, see random patterns like you know bullish flag or whatever in my chart. I use cycle indicators as basically triggers to initiate my traits for and I use them in very quantitative manners for… so for example if… if I use the RSI, I would only buy when the RSI hit a certain level and only sell when I hit a certain level and that would be defined and I would follow that without fail every single time.
AARON FIFIELD: Okay, and what’s the maximum number of parameters you’d feel comfortable including in a strategy? Let’s say a number of parameters or conditions which must be true for you to enter into a position like… I know, you’d like to keep things very simple, so… just so we really get a grasp on that.
KORY HOANG: Usually when I create strategies, because I don’t put too many inputs in there and you know overfit the strategy or data mine sometimes, it tends to have very few inputs maybe I would say less than 10 inputs, most of the time probably won’t even get above five or six things… you know those input with some… it would be something like the length of the RSI or the length of the moving average and we’re not… I want to short sell or not or what is the person allocation per trade and what the stop loss and profit targets are, done, you know, not nothing too fancy, nothing to complicate. It… it depends on how strong the anomaly is so for example, on… and you can go out, you can verify this yourself on a daily interval. If you look at SPY and if you were to use an RSI, a two period RSI and if you were to buy when that RSI is below 30 and sell when it’s above 70. That system produces very good result over the past 10-15 years. I don’t recall exactly what the numbers was but the Sharpe ratio was very high. It was above 1 compared to just buying and holding SPY itself. A matter of fact that system during 2008 managed to avoid a lot of the drawdown that SPY experienced.
AARON FIFIELD: Okay and when you say buy when these conditions are true, how do you exit that position like is what’s the what’s the guard line there?
KORY HOANG: On certain system, I would have a profit target but most of my system the exit condition would be based on my indicator so for example if the RSI is oversold, I would buy and if it’s… when it’s overbought then I would sell. And I would not use any sort of stop loss or profit target because that system I just described there as a mean reversion system. You tend to want to let mean reversion system have some room to jiggle before it mean revert back up, so using stop loss can sometime cut you out a trade too early and a lot of time with… you put in a profit target you don’t capture the entire move of the trades. And you would exit early so in my experience using profit target and stop-loss with only be reserved for asset and time intervals that offer lower intensity and momentum or mean reversion.
AARON FIFIELD: Okay and I realized I just… you’d already explained how to exit that particular strategy before I asked you exit when the RSI crosses 70 so ok, so where are we here? Now as you have been live trading a few strategies now for… how long has it been? It’s probably been about 18 months since you first started live trading with algos?
KORY HOANG: I see I think I’ve turned my first algorithm on July of 2016… so twelve years… I mean 12 months (AARON FIFIELD: and a bit) Yeah about 15 months, yeah.
AARON FIFIELD: Yeah have there been any times when you’ve… you’ve stepped in and overridden your strategies?
KORY HOANG: A couple of times, and I’ve learned that I should stop doing that, because… you know, the time that you know I miss out the most was after the election of 2016. After that election and when the market was still going up initially, I thought it was kind of weird because I had my expectation was… you know, the market would not react well to a Trump presidency for various reasons, so I decided to go against what I typically do which is to trust my algorithm to do to train for me, because you know… I already already identified the anomaly, I already designed the system, already tested it and it’s working. I should stick to it but for some reason after that election, I had a little of doubt so I decided to turn my system up all for a couple of months, and I did not turn them back on into March of 2017, and that was a really big mistake because had I had it on during that time, I would have made… you know, close thirty forty percent during that period. So, lesson learned.
AARON FIFIELD: Lesson learned, so what sort of returns are you aiming for… like you know, what would you classify is a good year for you?
KORY HOANG: So according to the back test I’m running on my strategies, I’m running seven different algorithms right now and when you put them all together in a portfolio, each one of them is taking 15% of my account per trade, so my account was not leveraged, I don’t use leverage at all. I do short sell but not on a lot of asset, matter of fact only short sell on Junior Gold Miner ETF and these system when you put them all in conjunction about working side-by-side each other, I’m expecting the backtest tell me I can expect around 60 to 70% in annual return. Now I know that sounds really well, but again I am trading inverse volatility securities and… you know, these securities have been shown to deliver incredible returns on certain years but also during other year they could produce very incredible disastrous result which is part the reason why I use algorithm to trade them and don’t trust myself to do the manual trading.
AARON FIFIELD: yeah I mean I don’t think that sounds absurd. You know what I mean, like you’re not trading 10 million dollars so… you know – I mean for 60 70 % yeah… I… you know, I think I don’t see an issue with it. Now, how do you monitor your strategies in real time? Like is there anything you do want a day-to-day basis to just check that everything’s running smoothly as it should be?
KORY HOANG: Yeah so… you know, I work a day job and I don’t have the luxury of monitoring my strategies all the time. But I host it at my home computer and I’ve remote… remotely connected to my home computer every now and then to check up on my algorithm and see how they’re doing, see if all the trades are executed or the API still connected. One of the big lesson I learned this year during the summer was I took a vacation. I took about two weeks off to go to Southeast Asia, to go back to Vietnam and travel. And during that time I still have my algorithm on and they were running just fine. And so it was a couple days before my trip was over that the API went offline for some reason. And so for a while the algorithmic trading platform now I was using which is called multi charts was not connected to my broker. And because I was oversea and did not have regular… you know, internet connection, I was not aware of that. And so by the time I found out… you know, I was already down I think five… about 5% or so in max drawdown which you know wasn’t a lot. But you know, it was a stupid mistake that could have been prevented, had I… you know, decided to check on it more often, so definitely learn a lesson there, too. You know, even if your algorithm are automated and your strategies are doing a training by themself, you still should check in every once in a while to perhaps daily if you can to make sure that everything is running smoothly because… you know, things can always go wrong. You can’t bank on the fact that your strategy will keep on working perfectly 100% at a time.
AARON FIFIELD: Another lesson learned. Now how’s it been for you to hold down a full-time job and do this, I guess part-time? You know, like you do this after hours, is that presented any challenges for you? Or do you find it’s… you know, it’s a good lifestyle?
KORY HOANG: It’s been pretty good for me. I mean my algorithms in total since I started in March, they’re up close to 17 percent this year so far. I know that’s nowhere near the 60 or 70 percent market I’m aiming for, but they spent a couple of times this year when I turned them off or did not… you know, let them run 100 percent of time so… but the risk adjusted return is very good. My Sharpe ratio right now is about 2.0 so… and having that running just on a side while I’m working, I find it to be incredibly easy. Maybe it’s just because I’m not handling a lot of money right now. It’s just my own personal money and perhaps that’s why it seems easy not to worry about… you know, your money being traded by a bunch of robots in a market. But you know, I go to my day job, I take care of my duties, perhaps I cleanse over to my algorithm at home to be a remote connect every 2 or 3 hours or so, make sure that everything is running fine and get right back to work. And at the end of day, I go home and tell… you know, how the algorithms performed today. And you know, if I made money and I lost money, most… most of the time I go home and it’s all green. Some days I go home is red. But I’ve seen that so many times now that I’m not fazed by it anymore. No, I’m just like oh ok cool and then… you know whatever else that I want to do and… you know, keep the algorithm running for the next day and repeat.
AARON FIFIELD: Yeah, and are you working towards going full-time into training or are you just happy to keep doing things as now for the time being?
KORY HOANG: I do, in several different paths. One of my plan is to eventually start a hedge fund and implement this… these strategies that I’ve developed for the hedge fund. But in order to do that, I need to start building track record first. So that’s one of the first thing that I’m focusing on doing it right now.
AARON FIFIELD: And tell us a little bit about… you know, there’s something very interesting about you and that’s the fact that you’re donating a portion of your profits to charity so do you just want to tell us a little bit about what you’re doing there?
KORY HOANG: Yeah so I have a startup. It’s called Quant Prophet. This is one of my startup projects. The website is www.quantprophet.com And what we do is we are a financial tech startup here in Seattle. And we are developing a platform where we will be hosting algorithmic and quantitative investment strategies that we collect online from various sources like Quantopian and blogs and… you know, podcast, even… and books. And we would sort of just collect all these strategies from online. And we would go through them and make sure… you know, we picked the quality ones. We remodel them, test and test them ourselves. And then once we have… you know, a good library, we put it on our platform and we make it available for people to subscribe to. So you know, they see, for example, a gold trading strategy on my platform that they like, yeah, they can look it up see all the statistics of the trade signal that has taken in the path… in the past. And if they want to, they can subscribe to that strategy. And they would receive trading notification, trading signal notification from it so as part of this startup, we are also running a charity program call, we call it trading for charity. So what we’re doing right now is we seeded the program with $30,000 in March. And we hook it up to my trading algorithm system that I’ve been using. And I began the year… we… in March, we said that by the end a year, we will take half of our profit generated by the algorithm and we would donate it to step up for Laos, which is a nonprofit organization here in Washington State. One of my friends, he worked with you know… a university of Washington professor. I can’t recall his name right now but he has a nonprofit organization here in Washington state called step up for Laos. And he, for $75, he creates a prosthetic limb, and he sends it over to Laos where it is given to children victim UXO unexploded ordinance which is a… you know, a remnants of the Southeast Asian conflict 30, 40 years ago. There are actually a lot of people, a lot of children victim in Lao who step on land mines and… you know, get their limbs and legs just torn off. And it’s a really sad situation over there. So you know, I’m personally from that region of the world. And you know, I kind of saw that… that scene myself before in life. And I decided if I’m gonna start business or start a project, I might as well as some corporate social responsibility aspect to it and try to make a difference. And you know at the end of the day, if you’re a trader, you can’t never have too good… too much good karma. So you know, hopefully, by do… by doing this hopefully I will inspire other people to… you know, take the same approach, take some of their blessings and winnings from the market and give it to the less fortunate, because a little bit of your trading profit over here means a whole world to someone… you know, half a world way who… who doesn’t even make more than a dollar a day.
AARON FIFIELD: Yeah, now respect that man, good on you, that’s… that’s really impressive. Well that’s… let’s close this out, Kory, is there anything else you’d like to add? I mean what if someone wants to find out a bit more about you? Are you on Twitter? And you mentioned a website before, where’s the best place to go?
KORY HOANG: Yeah so if anyone wants to contact me directly. They can email me my email is email@example.com, so that’s my personal email. You can also reach me at firstname.lastname@example.org.
AARON FIFIELD: Okay and I’ll also mention you are fairly active in the Chat With traders Facebook group, I mean that’s the… that’s how I came across you. I mean I wouldn’t know about you otherwise so I’ll just mention you, hanging out in there as well, if… you know, wants to join the chat with traders Facebook group, it’s https://chatwithtraders.com/facebook, that will redirect you directly to the group on Facebook. You’ve just got a request to join there and of course, I’ll accept you in. Kory, I appreciate you taking the time to do this. Thanks very much.
KORY HOANG: You’re welcome, Aaron.
Ending: You’ve reached the end of this episode of Chat with traders, but rest assured there are more episodes loaded with real market insight and zero hype on the way soon, so to stay updated with each great new release, subscribe to the podcast in iTunes, and we’d love it if you leave a rating and review. We’ll catch you next time on Chat With Traders.
Stably is a fast growing FinTech company from Seattle founded in 2018 by a team of former bankers and Amazon software engineers. The team has raised close to $3M in funding to date from angel investors as well as leading venture capital firms like 500 Startups, BEENEXT and Pay It Forward. Stably is also the blockchain developer for a wide range of cryptographic tokens and stablecoins backed by real assets, such as USD, EUR, CAD, precious metals and real estate. Our mission is to make financial transactions faster, cheaper and more transparent through a borderless digital money platform powered by open-banking and blockchain technologies.
For more information, please visit: www.stably.io
— — — — —