David Aferiat is the founder & managing partner of Trade Ideas, which publishes actionable market intelligence to help customers capture alpha in the capital markets. In this episode of Power Your Advice, Doug Heikkinen and David discuss how Trade Ideas makes technology normally available to hedge funds available to financial advisors and investors.
- How Trade Ideas makes artificial intelligence (AI) accessible to professional traders
- How AI provides the guardrails to help advisors make decisions
- Why David thinks advisors may be reluctant to adopt this type of technology
- Where David sees AI & Trade Ideas headed in the next five years
Resources: Trade Ideas
Douglas Heikkinen, David Aferiat
Douglas Heikkinen 00:02
Hello and welcome to the power your advice podcast. The power your advice podcast is designed to bring financial advisor new ideas, why those ideas should be considered and how to implement them into your businesses. This podcast is brought to you by Advisorpedia the best place for advisors to grow in their minds and businesses. . .
David Aferiat 00:53
Hey, Doug. Good afternoon. And thank you for having me on your program, it's a pleasure to be here.
Douglas Heikkinen 01:00
Trade ideas gives technology generally reserved for hedge funds to the individual retail investor and financial advisor. That big statement is front and center on your site. Tell us what it means.
David Aferiat 01:14
Yeah, our arm. So our mission for over 18 years now we are an 18 year overnight success. And it's gone through its ups and downs. But our mission from the very beginning has always been the same. And that is to provide the technology to enable our customers and subscribers to make better decisions in the US equity markets. That's it. We do this generally by have we've always brought innovation and this technology with a sense of leveling the playing field for people who otherwise would never have had access to this, in fact, just the opposite to more traditionally, the institutions of Wall Street and those with the appropriate capital that could create and or come close to the kind of technology that we provide. But you have a couple of you have a couple of forces that have come together one, just the natural evolution of computation and processing abilities, and the capacity of databases and in stored memory and machine learning algorithm and AI to arrive at a sense. It's arrived at where we are today, which is at a relatively low cost, generate the computational power and the effort to sustain that and deliver it to you know, to people who who otherwise never would have been able to access this kind of information in the past.
Douglas Heikkinen 02:38
You say you're making AI accessible to professional traders? how exactly are you doing that? How does this thing work?
David Aferiat 02:46
Yeah, you know, and it's, we got to be very careful, I mean, or not, because the terms of AI and machine learning are being thrown around so egregiously. And so it's not that they start to lose their meaning. So it just means that Sure, you can you can, you know, the word salad that can come out of certain places, washes over you, and you think, okay, that's great, but like anything in life, you can need to scratch the surface and do a little due diligence on who's doing what, and how is it doing? You know, how is it doing? To understand a little bit more about what is a What do people mean, when they when they when they say AI, what kind of machine learning is going on, but to your question, how we make it accessible. The big terms before machine learning AI, if everyone recalls or words like big data. And in fact, you AI requires such an enormous amount of data to look at patterns and define and test scenarios that you see it, you know, plugged in, where there's in, no in a measurable amount of data to be to be analyzed, whether that's credit card statements, or whether that's stock prices. You have that enormous pipeline of, of data. And, you know, you have to have as a core competency, to intake all of that information. Strong competency in how to manage big data, how to set up arrays of servers in order to compartmentalize break down the task and then reassemble it up into new derived data, add some new derived information, essentially, make it simple, you are taking a lot of information, you're normalizing it and standardizing it and running tests on it and analyzing it statistically. And then pumping out the other end new information derived from that. And we have built the systems over the course of 18 years to do that. And it's it's it hasn't hasn't been in a in a bunker, you know, developed on the ground or, or away from anyone. Those customers who have been with us for so long have seen the progression of this technology. And so I'll just give you a quick example. We started as a simple very simple scanning tool. Enter in some parameters, choose a few indicators And, and watch us produce a list of everything that matched your criteria. Well, that was great for a couple of years back in 2003. And then we built on top of that a back testing capacity than more normal back tests will take one stock one indicator, go back 10 years or a lifetime, if you will, and show you all the times that this indicator would have been the right time to buy this particular instrument over time. That's great. But that's not the way we approached that task. Instead of having to know the stock a priori, we wanted our customers to simply say, I don't know what the stock is beforehand, I just know that this is a pattern I want to see, go and find me the stock that met this pattern, and then test whether this would have been a good, you know, a good pattern to make decisions on. And so that event basis, we call event based back testing, it was one of the first seeds that led to the machine learning AI that we have now. But that is that ability to take an event or any event, and then find the stocks in the past that would have met the criteria and then roll forward to determine would you have had an edge, if you had marched forward with this idea was the was the basis of our innovation for the next three, five, almost six years. And what it eventually came down to was a customer experience that kind of rocked us out of a bad place we were in with our technology. And that is the menus in order to produce a back test result that was favorable, were so many, it was like the carrot and the stick and we gave our customers nothing but the sticks and hid the carrot. Meaning you had to go through these screens to develop your own your own scan, and then run the back test and determine if the parameters you set were going to be good pull the lever and crank out the results manually. And then it would have been Congratulations, you have a terrible idea on your hands, you should avoid going into the market at all costs. And that's not what we're about either. Eventually, we had to come up with Okay, look, our customers are not doing a good job with this technology. How would we show the best version of it ourselves and then give that as a as a crutch or as a prompt to our customers. And that turned out to be, you know, the design flaw there was that there was no scale in it. We were doing all the work suddenly of creating this and we thought how can we add scale, create a system that does the work and finding what's the best combination of alerts for idea generation and risk management, the parameters for setting the trades that would yield the best profitability over and over again. So from that exercise, was really birth, this idea that we need to be running the simulations, we need to be running the scenarios and we need to be teaching the system to, to look at everything that's been created in terms of the degeneration and the risk management, change those parameters ever so slightly to see if there's a better result. And if so, keep them or out of whole cloth, find some new indicators to add and see if that improves the result. And and continue then to search and find and run these optimizations done overnight. And, you know, that may be the long explanation. But that is the essence of our machine learning AI, it is the optimization of over 70. Now, algorithms that we have defined for the AI and that it then optimizes and runs scenarios in order to get to the next day the biggest before they open with a list of not 70 but down to whittled down to about six or seven algorithms that from which it's going to look at the market activity that's locked.
Douglas Heikkinen 08:43
How do you nail it down and talk about the benefit of actively employing this technology to asset managers and stock traders?
David Aferiat 08:55
Yeah, so kind of like you know, what's what's the benefit of all this? You know, I've talked before about what our mission is to help people make better decisions. So the AI produces these results to help the lay out the guardrails for making a better decision. Here's an idea. And here's the entry where this idea becomes valid. Here's the exit the stop loss where this whole hypothesis fails. Because if it reaches that, it's time to get out and live to see another day. And then here's the profit target where you can start making more decisions about maybe Hey, taking half the position, maybe adding to it or maybe getting out and taking profits altogether over various timeframes. So that, as I said, helping making those decisions is what we do our special sauce, maybe the AI but the value proposition, what we're trying to do is to find these opportunities, where no one else is looking before and before anyone else can see them and serve them to our you know, to our clients. If you look at the market, like the depth of a sea at these different layers at these different depths of the ocean, You can see all sorts of different activity, there's what's happens at the top, which are the major market indices, and so forth, and everyone sees and looks at. But just below that level and depths below, you can see more interesting activity where sector rotations are occurring. And other news is happening in places where you just can't quite see it before even hits the news. And so, in the context of a registered investment advisor, or even a professional trader, we find the most the most relatability. And most of our semi professional we call an individuals who use the tool are often in this capacity, where it's about communication, and it's about the proper allocation, within a portfolio that may be invested in several buckets covering everything from real estate to, you know, to value into international, when you finally get down to that pot of money, which is this is an amount to engage, you know, in the market, and keep that excitement of as a proxy for decisions that are contained within this allocated set of money. That the the that especially in a registered investment advisor context with a client, you know, that kind of back and forth for some clients is very engaging, and keeps, keeps things keeps things engaging and relevant for the for the client with the advisor and for the advisor to be able to reach and communicate with those, that certain set of clients that are act that like to be actively engaged, it's not to say I'm taking my whole 401k and going in and tell me what to do. It's more look, we've covered everything that we need to cover for our you know, for a balanced approach. And with this pot of money, you know, let's engage in the market and learn a few things and see and talk about what needs to be talked about.
Douglas Heikkinen 11:46
Would you say that most people are trading in primary colors? And this allows sophisticated traders to see so many more colors and opportunities?
David Aferiat 11:58
Yeah, exactly. I would call it more orchestration Dolby stereo, I would, you know, it's, it's really the almost also the equivalent of like what you know, otherwise, you're watching like a football game without the commentary. I mean, it's sometimes it's nice to be in the stadium, but you miss a whole layer of, of, you know, of nuance. You know, I say this, on the eve of baseball season starting and my, you know, my wife has an uncanny ability to be sitting there and to say, seconds before the announcer, about any player, any particular situation, you know, you know, if there's a fly ball, he's going to go to third, and boom, it all this stuff happens. And so my point, aside from praising my wife is the fact that that extra layer of commentary is, is isn't is the advantage. And we aim to provide that by adding these layers of colors as you said, or orchestration. You know, and, and the Take, for example, you know, here's one example just today, I mean, the, our AI system found i n vo, n i n vo is in invo bio science. So it's a biotech firm. And, you know, when you catch something like this, that the entry price for this relatively low price stock was $6.31. And its last price now is at 975. Now, do we catch those all the time, you know, no. But more with more, you know, it's not a rare instance where we're just it. And so I envio came through a particular algorithm that we call breakout long. And offhand, I can't tell you exactly what the algorithm is looking for. But it's a long based strategy that is by its title here looking for breakouts to the long, you know, for on the long side, and over a longer period of time. And so sure enough, you know, these, the strategies, and when an idea is published, it's not meant to say you're either in it or you missed it, it's the chance to be able to take the suggestion from the AI and then decide and apply. It has its risk management parameters that it has already devised. And then it matches or doesn't match up with your risk management parameters, such that you can either say, this is an excellent trait to look at. But for the AI activity, I would never have seen this stock. And now I can assess if this move is already the bulk of where it's going to go from $6, you know, to $9 and it's over, then it's not for me. But if I can see in the pattern that you know what this is, this is the start of something that I can see over a longer period over a longer, you know, monthly chart, I can see that it was back. In fact, at these levels previously, it fell and now it's coming back. Is this going to be a longer term, you know, play those decisions, that conversation becomes more interesting, but because it's because the AI is found you something that is that is interesting. And it gives you this ability to now side and so we have for our institutional customers who We use the product outside of our and I'll say this, our large majority of customers are our retail, at least retail far as we know them to be, we don't ask for too much information as to who they are other than their email and filling out or the exchange agreements. But to our professional users of the technology, it's very much an input into like the decision committee, you know, in a portfolio fund or any any kind of fund where, you know, Monday or Wednesdays, you walk in with your with the ideas, and the portfolio manager sits and listens to the analysts. In this. In this instance, the AI is one of those kind of analysts to sit down and said, here's what it traded yesterday, is all this computational effort that we do. And what it's what it traded yesterday, is there a value in looking at it the next day, you bet there is, it's not just the intraday has to be there when it happens. It is often you know, the footprint are the canary in the coal mine that is looking at the say, not just this stock, but now taken in context with a lot of people do is look at the whole list of stocks that were traded by the AI and then start to look at the metadata around that. What sectors were these ideas from? We're, you know, we're over the last five days, you know, has 60 or 70% of these trades been in bio science or bio pharmaceutical preparation stocks? Or have they been in tech stocks? Or have they been in energy, you know, and or real estate stocks, you can start to see some patterns that the you know, that the AI collectively across its algorithms across the optimizations, you're starting to look at and find advantage. And now, if I may, I will bring back one other baseball analogy. I talked before about Wi Fi the AI about 70 of these algorithms. And every day it whittles them down an overnight process that we call a quantitative combine very much like, you know, athletes who would go through a sports combine and be assessed and rated and tested by the coaches and, and, and staff, you know, on a team before being selected. This AI does the same thing, it runs this combine and whittles down the list of 70 to about six or seven. And again, some names of these algorithms will appear for several days, and then they'll disappear. And then they'll come back or they won't. I've been around the business in an industry long enough to know that there are some borehole brokerages and whole advisories that have been built on just one algorithm. And they have a times for stretches of certain years, you know, years ago, lived and died by that one algorithm as to where and how to direct people's you know, money and around when it failed, it failed spectacularly. I'm recalling one particular company in Europe that that had an ugly, ugly end to use using an algorithm just one to you when it worked, it worked. And when it didn't, it failed. But the idea of running this combine is to let the AI determine where the decay is happening, where algorithms are moving away from the market or rather, where the markets moving away from them, right because the markets always evolving. And, you know, certain stock, certain algorithms will find their sunshine for a period of time, and then decay away. And when that happens, you can't put that in the hands of a human to be able to find that nuances when at the margin when the stock is when this season the algorithm is no longer producing the results that let it cut, you know, cut bait and choose another much like a baseball coach is going to fill out a lineup card. There's determinations as to term deceive, okay, who am I players are on a hot streak who my players are healthy and not injured. Who am I players actually, you know, lineup well against the opposing teams defense and the pitcher. So if those considerations go in, you have a lineup card that's done. And the team plays ball is is you know, a similar analogy as to what the AI is doing with its selection each overnight of the algorithms through which it's going to see the market. And that protects itself and our customers from mayhem that can happen at any moment, like in a black swan day. Take for instance, like, you know, when Brexit was announced, or presidential election or, you know, God forbid, bought the bombing of some country that can create these kinds of shockwaves. The AI, the AI is trained, if I may step back one more time, and the premise of what we're trying to do, how we deliver this technology make it accessible to our clients and subscribers. is what is the AI trying to do? And it's actually a source of great optimism that I have about being able to talk about the performance of the AI and to be able to relate to people how how it works. In any trading or investment environment. There's only four outcomes that you can possibly have. There's only four things that can happen. you'll either have a A small loss, a big loss, a small win, or a large win. And all the AI is trying to do is land in three of those four outcomes. We'll take a small loss any day, that's the learning of that's a, that's a learning event, we'll take a small win for sure. And we'll always take the big win, all we are trying to avoid is the large loss, and three out of four outcomes. That's great. Those are great odds. So if we're landing our clients into three of those four outcomes with every decision, that's a reason why we you know, the reason why we're growing so much, and the reason why we have the number of subscribers and clients that we that we do, is simply that
Douglas Heikkinen 20:44
You say that most of your clients are retail, or you think they are, why aren't more advisors adopting this? Is it just awareness? Something that they don't know about? Or what do you think?
David Aferiat 20:57
Well, for many years, we spun our wheels trying to ask that very question. And what I think it comes down to is that there are some advisors who, you know, advisors come in all different flavors, and some of them are motivated. And some of them have the resources behind them to have different motivations. Some of them, for instance, have the capital research and structure of large, large firms, where that frees them, to give them the ability to say, look, all the engagement in terms of research and where and how this money is performing, is outsourced to the resources of the firm that they represent. So like Charles Schwab, or you can have all sorts of, there's no limit to the data that you can access to see how well you're doing or how badly you're doing. And there's no shortage of ways that they can move, you know, put you into certain things. And so an advisor just kind of comes down in that context to just finding the next client and dealing with more relationship. And then there are other advisors who are much more hands on, and those are the ones that we find sooner than than the others, these advisors often have a oftentimes take a more active role in directly engaging the market on behalf of their clients. So an example is you'll have an advisor with a master account. And then a list of sub accounts. And the sub accounts are each client's portion of their of their of their portfolio, that under the Master Master account, is traded in or engages in the market, and then allocates those trades across to the to the sub accounts. And that's where the advisor is really engaging directly in the market. And that's where we find our tools, you know, leverage the most for those individuals, because we're providing an enormous amount of capacity and resource to, to land them in and help them enable them to make better decisions. Then, some committee from some fund, who, frankly, has covenants and restrictions that prevent them from seeing parts of the market that were able to see. And that's, again, is another distinction that's important to relate, you have funds that just can't, you know, go below the the s&p you know, after 200, or 250, or even 500. But there are more stocks on the market that than just the s&p 500 or the s&p 200. And it's in that context that we're not trading, we're not pointing at penny stocks, you know, or over the counter, but there's between those and the s&p 500 there's a whole healthy, you know, depth devotion that we help navigate that there are some funds can't can't necessarily, for multiple reasons, those covenants are in place, because they do not want to have over market impact with money that they have to move in and move out into a position. And it's for their own risk, you know, for at least larger funds, his risk profiles just prevent those kinds of behaviors. But that's why again, I say that advisors that are more directly engaged in the markets, use this tool and leverage what we're doing for them.
Douglas Heikkinen 24:04
Is this the future is now technology, you better start testing it thinking about it, are you gonna be left behind the the trend?
David Aferiat 24:10
Well, yeah, we had one of our advisors, you know, tell us that what you're doing is table stakes. Congratulations, you got a great looking product here and it works and nice little pony, you know, as you often said, and but that's the table stakes to be in the market these days, you might win if you have it, you're definitely going to lose if you don't have it. So I'm like okay. So, now we need to find and we do we have a wonderful pipeline of products that you know, that that we invest in quite heavily back into the firm to help bring to the light of day are these you know, the directions of where the the this technology is going to take us?
Douglas Heikkinen 24:53
Is there a way for advisors to harness the power of trade ideas to directly impact the ability to generate perform Amundsen Garner increased.
David Aferiat 25:03
Yes, yeah. So I, you know, a day in the life kind of routine for an advisor with this technology is twofold. It's, and we were actually launching a product that matches this kind of day in the life or journey, advisor journey. When I, when I described the following, we've seen sometimes so there's a an advisor, as part of the staff, there's an office manager, and there can be sometimes a young, more technology savvy analyst, if you will, who, you know, who sits in front of, you know, trade ideas, and basically kind of says, Here's something that's, that's, you know, that's just come through, or here's yesterday's results that we want to flag and, you know, it creates a small write up, that says, you know, our research found this stock from yesterday. So tomorrow, it might be our research found iron vo, and, you know, here's the, here's the reasons why the stock is, you know, it's interesting to look at, it's crossed over its 200 day moving average, it's broken out, or it's, you know, it's now filling a previous gap from several months ago. And it stands in relation to not just the technicals, but fundamentally, you know, the impact of the vaccines and the pandemic is brought more attention to this, etc, etc. But you have a nice little narrative that you can find within trade ideas, both on the technicals as to why it's doing what it's doing. And even, you know, we have a whole subsection that you can, when you once you find the idea, we have, you know, all the additional research on news that you can look at the you know, the position of the insiders that are in the company that are trading the stock, what positions are there are they taking, and then you can see even additional stocks within the sector, that are also, you know, moving like this iron, vo, or outside of the sector, but they have the same technical movement that are also crossing over their 200, day moving average are also doing the same things, even if they're not in this particular industry. And all of that can be used to create quite a narrative to do what to engage with a client, and not just all their clients, because this tool isn't for everybody. And this tool isn't for every advisor, but for those advisors who engage directly and want to grow a u m. And, and on that basis, like to engage with their potential clients or existing clients then who like also to be engaged in the market, you've got a storyteller, you know, here that that is not only finding the ideas, but is acting like your co pilot, you know, giving you guardrails, for the decision making process of about when this trade is live, when it should, when the thesis fails, and where a profit target for the first iteration of assessments can be had.
Douglas Heikkinen 27:59
It sounds like the tool is constantly evolving, where do you see trade ideas in the next five years.
David Aferiat 28:05
So we are technologies pointing into ever, in two ways. And I didn't actually mentioned this with the you know, so as I mentioned, we're coming out with, we're coming out with a newsletter that does more of these kinds of narratives, narrative, you know, articles that we're talking about. And those products are geared away from the GUI interface of our technology, which the platform and this is part of the direction where we're going and we'll be having more informational products. So and on a newsletter basis, we'll you know, we'll we will have, we're about to publish, we call it trade ideas, relative strength index, where we have an index of five stocks that are performing in the input from an algorithm in a particular way. And, you know, every week or so every month, and depending on what you want to choose, and the length of time that you want to be made aware of, you'll know that it will, we will publish where certain stocks cycle out of that index and where other stocks cycle back in and have that as a as a newspaper or as a newsletter, rather, in a newsletter form that advisors can can then take and say, Look for all the power of trade ideas, I just want to be able to focus it down to some consistent ideas. And it's a layer, it's an extra, I should say it's an option for what kind of engagement people want to have with trade ideas. We recognize that the platform we have and the technology is visual, but it's very interest tons of information on it's like a dashboard of you know, flying an airplane. And if you don't want that much instrument instrumentation, then you can, you know, let the copilot fight or you can use this kind of synopsis newsletter product that we're that we are launching soon. That the simple index that allows you to kind of consume the data in a much more You know, way that doesn't cause so much into, you know, indigestion from eating so much.
Douglas Heikkinen 30:06
And if people are interested right now, where can they find you guys?
David Aferiat 30:09
Sure. So that's trade dash ideas.com. From there you can, you can see, we have a whole list of resources, both educational, and informative, in the sense of educational. Our Youtube videos are webinars that we have several times, we have over 12 hours of programming each week, where we're talking about the markets and the functionality of our tools and examples that come through. If you didn't want to, you know, if you wanted to do the first toe in the water experience with tre that is you if you come to trade, dash ideas comm you can sign up for what is our trade of the week, it's a free email that we send out on Monday mornings. With a description that I just described, like the one we're coming out with, with our relative index. It's a one stock idea with a narrative of here's the entry price where the stock with a trade is live. Here's the exit price where the whole thesis fails. And here's our profit target to keep in mind, so that you have the guardrails to make a fully informed decision around one idea that's selected. It's actually it's sourced from the AI produces a list and then we have our data scientists and market you know, market guys make the final decision as to what the trade of the week is going to be so that we don't repeat too many times in one sector. And so we take another considerations. And we we choose one and and off it goes.
Douglas Heikkinen 31:40
That's great. David, that was super interesting. We really appreciate you joining us today. Thank you very much.
David Aferiat 31:46
My pleasure. Thank you.
Douglas Heikkinen 31:48
For everybody Advisorpedia our producer Jakie Beard and the power advice podcast team. Please follow us for our latest updates on Twitter, LinkedIn and Facebook at Advisorpedia. This is Doug Heikkinen