Matt: “But what do you use with Nurp to help identify a profitable strategy? Like where does it come about, like, hm, I’m going to look at a Crossing RSI Divergence or these SMAs need to cross over here? Like what is it, the recipe? And I know that you probably can’t give the exact one because everyone’s trying to make one themselves, but what are certain things that can help somebody understand what’s a good strategy and what’s a poor strategy?
Abhay: Okay, yeah, that’s a great question. I think, uh, so those two are kind of different questions in a way. Uh, the first question is how do we come up with strategies? It’s a matter of finding patterns in the market. What we’re really looking for is a consistent, repeatable pattern, and that just comes from us being Traders ourselves. So I’ve been a Trader and investor for more than 5 years, and Jeff even longer, as you know. So it just comes from knowing certain things when they happen again and again and again. We find that pattern, and then we develop a strategy to use that pattern effectively. The second part of that is, though, uh, risk management. Because even though you might find a consistent pattern of behavior, it’s never going to be 100%. So it’s, I don’t think you, I haven’t come across, I don’t think you would have either, an algorithm that has 100% win rate, right? So, um, any sort of, same with any manual Trader as well. So any sort of pattern that we’re finding, or which is effectively an inefficiency in a way in the market, then that’s not going to be 100%. So we have to think about different ways for the times that that pattern doesn’t actually play out, how do we manage that risk? So we look at all the different things around drawdown, we look at different strategies to be able to scale into trades, whether that’s Grid or Martingale, um, again to make sure that we end up on the profitable side despite not winning the first entry on the trade. So that’s the early stage of strategy development. And then once the base strategy is in place and we have a beta version of the algorithm, from there on it’s the quants that come in and help really optimize it. Because the same pattern is not going to behave the same way on every pair, for instance when we’re talking Forex or even crypto. It’s not going to be the same on Bitcoin and on Ethereum. So, the quants really help optimize and understand map Behavior at the asset level, whether that’s a Forex currency pair or a crypto, understand behavior, and then fine-tune the strategy to that particular asset. So, in a way, it’s like tuning the engine of a car to boost the performance to the maximum.”
Matt: “Okay. So, from my understanding, the base layer, which is the first strategy, is finding the engine of the car. So, we’ve identified we want to go with a V8. Then the quants come in and they put all the fueling and everything, all the parts that are needed to make this V8 run efficiently. So, you guys would essentially be, would go out there, find a basic strategy, and then you’d have Quant come in and then maximize a basic strategy to become a super complex profitable strategy. Am I far off here?”
Abhay: “Yeah, pretty close. The base strategy itself also has to be profitable because if the base strategy is not profitable, it’s not likely that with more optimization it’s going to become profitable. And so, the first thing really is, is the base strategy profitable without any optimization? We just run it. Is it still going to produce a good result? If the answer is yes, then we optimize it to produce a great result. That’s the goal.”
Matt: “Okay. And then, you’ve got you and Jeff or whoever’s involved, you’d find the strategy. It’s profitable over a certain amount of time. What’s that amount of time that you consider to be profitable?”
Abhay: “Yeah, that’s a great question. Um, so far, at least in our algorithms, we’ve primarily focused on the last 3 years because even that’s industry standard. Because even other firms who go back further in time understand that the current market state is a lot similar to the previous 3 years than it is, let’s say, the last 20 years. However, at the same time, it’s good to always go back further and see certain events and certain black swans and how the strategy performed against it. So, right now we’re working towards moving to a 10-year timeframe where we go back and test a minimum of 10 years while putting a higher weighting on the previous 3 years. Very interesting.”
Matt: “Wow. So, three years is basically the point where you deem it to be worthy or profitable, and that’s what you backtest each of these strategies over the time of. Now, since we’re talking about products and different strategies, the most popular one in Nurp that I feel that maybe is my opinion but this is what I think it is, is the Fed bot. But there’s several others, and I know that you guys are working really hard. Are there more in store for the community and potential clients?”
Abhay: “Yeah, absolutely. We do have a few projects in the development pipeline right now. And we, in fact, expect another one to launch next quarter which right now is in testing. And I don’t want to give away too much, but in testing, it has outperformed the Fed. So, really, really excited for that one. If things go well, then next quarter, we could potentially launch.”
Matt: “And you, again, we don’t have to unpack too much here because I know that maybe there’s some type of I guess timing with releasing of software, but you said performance here and you weighed it to the FED bot. When you say performance, are you weighing it solely on its performance of taking x amount of profitable trades or is the performance risk adverse or is it a combination of the both?”
Abhay: “Yeah, that’s a great question. It is a combination of both. However, what clients generally care about is profit at the end of the day. So, we put a heavy focus on that. But we also put an equally heavy focus on risk management. So, we figure out, well, what’s a reasonable drawdown level for this strategy? Let’s say it’s a high-profit based strategy. Then we figure out, well, clients would probably be comfortable with about a 25% drawdown. Let’s test how many times that drawdown was breached in our backtesting. And then based on that, based on even factoring one or two instances of losing 25% of the account, where do we actually end up at the end of the day? So, we test for all of these different permutations and combinations. And right now, the strategies we have in the pipeline, I mean the one that I was talking about, we’re very close on. But even another two, um, that we’re in the more early stages on, we’re now thinking about having multiple configurations where in the set files itself we let’s say provide a high-profit configuration and a low drawdown configuration. So, it suits the different types of clients because some clients are happy with a much lower return as long as they’re taking a much lower risk for that return where some clients expect a higher rate of return but are more willing to take a lot more risk..”
Matt: “So, the one that’s launching, and again, no pressure, you don’t have to say too much. The one that’s launching next quarter, is it similar to the FED bot in terms of its strategy that makes it want to trade? For instance, whether it’s a moving average cross is what is triggering the trade? Is it different from the FED bot or is it similar?”
Abhay: “Uh, it’s quite different. Yeah.”
Matt: “Wow. Okay. The other unique aspect about this one, which I don’t want to say it will definitely launch next quarter because we want to be fully comfortable with our testing before we, you know, take it forward. However, the other unique thing about this one is it’s inherently got hedging built in. So, if you’re on USD Cad and you’re going long and you’re in a long sequence, it will not take a short sequence until the long sequence has closed. However, with this strategy, it can recognize when things have changed on that pair and it can be in a long sequence and then start another entirely new short sequence at the same time. And eventually, both would close out. But while both are open, they hedge each other for drawdown. So, it keeps the drawdown lower on the account because of the hedging attribute that it has there because the FED bot would just run the trades and then scale in until eventually it could start closing out old ones. This one you’re saying will take opposing positions to its original one when things have shifted and kind of hedge out any potential drawdowns, right?”
Matt: “Exactly, yeah. Well, it will continue the long sequence and then it may even take another short sequence. So, you now have two sequences that are running in parallel in opposite directions on the same pair. Whereas with FED, you could only have one sequence in One Direction because it’s quite difficult to keep track of these Martingale based sequences. However, with this strategy, that’s been figured out. So, the big positive is because of those two sequences running in opposite directions, they hedge each other out and it lowers the drawdown significantly on the account.”
Matt: “And how many people do you typically have working on a given strategy? For instance, this one, how many people are actually, you know, picking and prodding at it to see what it does and how it’s performing? Is it just you by yourself or?”
Abhay: “No, we do have a full-scale team in place. We have four that are full-time on the product team right now, with another two roles that we’re hiring for right now. So, we will soon be six plus another, um, another four or so that we work with more part-time, project-based work. So, almost 10, I would say.”
Matt: “Wow, interesting. Okay. So, I feel like we’ve basically covered up everything that I wanted to talk about. One thing that I would love to do in the future with you is actually pull up certain charts that you’re looking at and basically give us some projections. And I know nobody’s a genie, but it’s pretty interesting to see someone’s thesis of where a particular asset could head in the future. But we are running out of time, as you can see at the top of the screen, we have 6 minutes left. So, um, I just wanted to thank you for your time hopping on here, and I’m pretty sure that everybody in the community and potentially to soon come into the community will find tons of value here. And is there anything else that you’d like to share or cover before we jump off?”
Abhay: “No, I think we covered a fair bit of ground. Thanks for having me. We should do more of these in the future, and be happy to jump on charts and show things as well. That I think is a big differentiator with us at least from what I’ve seen in the industry because we don’t just hand out software to clients and then leave them by themselves. Because we’re active Traders and investors in the market, we continually give them updates in the community, we alert them of potential risks we see in the market, and potential changes that we’re making on our own accounts. And then there’s also a support team in place to help them with any setup related aspects or if there’s an error on their terminal and they can’t figure out what exactly it is, we’re there to help them along the way. Because our North Star is ultimately that clients are profitable and successful. Because if they are profitable and successful, then that’s a measure of success as well.”
Matt: “That’s a great measurement. Great measurement. Well, awesome. Thank you, AB. I look forward to connecting again in the next episode. Thanks a lot.”
Abhay: “Thank you. See you.”
Please visit Trading Strategist Talks Algorithmic Trading | with Abhay to watch the interview on YouTube!
The post Trading Strategist Talks Algorithmic Trading | with Abhay first appeared on Nurp.com.
