How to Create Millions of Trading Strategies a Day and Keep the Best Ones | Miguel Jiménez

[00:02] algorithmic trading using a tool like Strategy One. I did a master's degree in statistics with the R software, which was a programming software, so I working with data. I even

[00:16] backtested a strategy and made 5,000 trades manually over 8 months. So I said, this has to be able to be automated somehow. I kept searching and searching until I came across Strategy One and that

[00:30] . The heavens opened up for me and I said, "Wow, this is exactly what I need." In the last mining I did to extract strategy, I mined 60 million strategies over 3 days. If I have to give any recommendations or

[00:46] advice on algorithmic trading, the simpler the better. make money in the financial markets by letting an algorithm trade entirely for you? This is a long-standing discussion within the

[00:59] believe that it is impossible to leave everything in the hands of a robot, that a robot cannot an experienced trader. But the truth is that many traders don't want to realize artificial intelligence, robots, and tools that allow

[01:14] traders are advancing by leaps and bounds. There will come a day when want to accept how the industry is progressing, and you will see that there are fewer and fewer sense to continue making decisions as human beings because there are

[01:29] do it better than us. So to talk about this, today I've brought here a trader who has been using algorithms and robots for many years and screen at Darkwun X, earning around 20%

[01:43] trading manually, doing it through robots. And not only that, but it doesn't does it without the need for programming using a single tool which is tool, as he will explain to us, is basically test thousands of strategies

[01:57] tool generates them, and what he does is investigate the best ways to strategies, the most successful ones, or the ones that work best for him, and then apply them in the real market, check, monitor how they work, and if they work well, continue

[02:10] have the expected results. As you can see, it's something completely different from what we leading him to generate very attractive returns. Miguel from the Winning Trading Strategies channel to have him explain how

[02:23] this process works, what it's like to stop trading manually, to put your entire future this case, a robot that has been created with a tool. How does he manage to Because think about it: what Miguel does with this tesquan is something that is impossible

[02:37] for a human being to do manually. It would take a human being months to create a single strategy that is robust, that is possibly winning, or at least that seems so, working daily and very hard. But instead,

[02:49] he generates, as he will say in the video, tens of thousands of strategies in a few obviously beyond the reach of any human being, as you can see. And my question makes sense and if you can actually make money this way, or if it's all just

[03:02] something I'm missing because it really seems incredible to me. So we're going to what his data is, how he selects the best ones, and how he ask him everything that comes to mind about how this thing of creating

[03:16] tool like Strategic One works so that we can all learn a trading options. I hope you enjoy the interview and good luck! And traders, as you can see, Miguel has an educational community of traders where he

[03:29] Quant and the automation of trading strategies to try to gain you want to be algorithmic traders and are builders, and trying to create strategies completely

[03:41] left below in the video description so you can go directly to everything it offers. Believe me, you'll like it a lot. Now Miguel, how are you? Hi Victor, how are you? Great,

[03:55] thank you so much for inviting me to the channel. On the contrary, man, seriously, thanks for met, as we were saying, there at an event we did at Ola Traders about , more or less. Or it was last year, more or less. And we were chatting about

[04:08] say the same thing, and we talked about it now, we wish we could record certain talked about fatherhood or children, I don't remember very well, but it was something like that and think you can get a lot out of those kinds of recordings, right? You're

[04:22] in the whole world of trading, right? And suddenly instead of talking about trading people think more or less the same, that we analyze risks, I don't know what, I don't know, also wanted to bring you here because I think you can contribute a lot to people.

[04:36] that I find very interesting, and that's why I'd like people to get to know you. So, if you'd like, you can provide some context: who you trading style is like.

[05:06] has its advantages, it has its disadvantages, like everything else. But well, that was, let's say, the niche in which I went on to you know what's going on, Victor? In the world of trading, and

[05:21] especially algorithmic trading, you open one door and arrive at a room with four more doors; you explore the different ones, and it's an endless path. And well, as I said, I specialized in all this

[05:36] algorithmic trading stuff, strategy automation. Also, well, I'm quite involved in one of the specific parts of this software, which is a wizard-like feature that allows you to integrate

[05:49] an external strategy into the Strategy One ecosystem, right? And well , that's how I started. But well, my origins are like I

[06:02] think almost 80% of the people in trading. I started with cryptocurrencies, trading. I started with cryptocurrencies, I started by burning through accounts, in fact I also participated in that NTF game boom, I don't know if I was involved in all of that,

[06:19] NTF game boom, I don't know if I was involved in all of that, uh, and well, it's all a learning experience, it's the path we're taking and well, when you're burning through accounts you also realize that training is necessary. And at that time I did

[06:32] training is necessary. And at that time I did some training, a master's degree in Forex and financial markets and such, and from there I also started to learn a bit more on my own, and well, following the path that almost

[06:45] passed since then, approximately how many years have gone by? Well, look, my first contact with the market, I remember it was in 2015,

[06:57] because, you know, a coworker, I remember he was laid off and I remember he was laid off and with the severance pay he started getting into stocks, but he also approached it in a very

[07:11] interesting way, or at least it seemed that way to me at the time, and that is that he didn't look at whether the stocks went up or down, he only looked at the dividend. Okay, give him a certain dividend, and with that dividend he would do his calculations and

[07:25] . Excellent. And then, based on that inherited experience, I started to open my first account, I started to do some things, and soon after, as I said, the

[07:41] started to be able to operate well with Binance and Coinbase. In fact, I still have open accounts with these companies, and that's how I got

[07:53] how do you get there? So, since then , are you into crypto? You're all in world of programming, of algorithms. Do you have any employment connection to that, or how does that work? Well, my

[08:06] to that, or how does that work? Well, my academic background is in biology, the old biology that was done, but I specialized in oceanography, in I specialized in oceanography, in oceanographic ecology, and I

[08:20] oceanographic ecology, and I also specialized in the analysis of pico-nanoplankton distributions in the water column. So, why am I saying all this? Because what I worked with was a very

[08:34] specific device called, and still called, flow cytometry, which called, flow cytometry, which selected each particle in an attractor, and in the end, it handled massive amounts of data, so I

[08:47] had a lot of background in statistics, statistics, attractors, camin, all this kind of history. In fact, I later also did a master's degree in statistics

[09:01] using R software, which was a direct programming software, background in working with data. What happens when, after burning through my accounts, I started doing this master's degree I was telling you about? In

[09:17] doing this master's degree I was telling you about? In this master's program, you would backtest the strategies you devised using software you know perfectly well, Okay. I even did a backtest of a

[09:31] strategy that I remember calling the cloud because it was based on the Ichimoku cloud and all that, and I did 5,000 trades by hand over Holy [ __ ], that's a good thrashing, huh? It was brutal. I think I was the one who did the most operations

[09:47] , that is, the one who did the most operations in this academy. What's going on ? There came a point during the backtesting process when during the backtesting process when you were saying, wow, maybe we could

[10:01] improve the strategy this way, or that way . Of course, if you make modifications in the middle of the process, the backer test is no longer valid. So I said, this has to be able to be automated somehow. It does

[10:15] n't make sense to do it this way. And that's where, let's say, the seed was planted in me to ask, how can this be automated? And then I started looking for different software. Well, uh, I remember that the first one I used was

[10:32] Alfa Alpha Trading, I think it was called, by Almodobar. And then I kept searching, I kept searching until I came across Strategy Quan and that was a turning point for me

[10:44] . It's like the heavens opened up and I said, "Wow, this is what the heavens opened up and I said, "Wow, this is what I need." And that's where I started to delve deeper. That's great, man. Excellent. So, in the end you get into this

[10:57] whole world, because you were also, let's say, incentivized by the data you were working with, and then you reach a Because you have no other choice; if you want to go further, you have to search, right?

[11:11] brutal. Yes, it was a natural thing, not planned in that sense. It is true that my background was a statistical, scientific background, very much about applying the scientific method and such

[11:26] because that is my training. In fact, I did a doctoral thesis at university, I had several important projects and so on, but of course, there comes a point when you get into this world, uh, the time that ends up being

[11:42] dedicated to the processes is brutal. And I was saying, this is not possible manually, at least not for me, no, no, no. Total, total. And then it was a way to start looking for, well, the alternatives that technology

[11:59] offered at that time. 100%. I don't know if you think the same, but I've at that hybrid point. I'm still there too, but in the one where they do backtesting, that is, they use more objective data and information, so to

[12:13] trading mechanics, but they do it manually. So, their job is to collect data manually and do backtesting. So, that's great. It seems to original trader, doesn't it? The one that executes it manually, but the

[12:26] algorithmically is insurmountable, because if you have a strategy that is already mechanical or objective, really objective, why not automate it? Know? So, why stay at that hybrid point? No, I do n't know if you think so if you don't see it the same way.

[12:38] n't know if you think so if you don't see it the same way. Look, I agree a lot with some of the things I've heard you say many times, and that is that you can make money in the markets in many ways. I know

[12:52] discretionary people who make a lot of money, bad at discretionary trading, but who are pretty good at

[13:06] , right? So I can't criticize anyone or tell anyone that it's better to anyone or tell anyone that it's better to do it one way or another. Now, do it one way or another. Now, the testing, verification, and

[13:19] analysis capabilities of an algorithmic trader are impossible for a manual trader, algorithmic trader are impossible for a manual trader, even if they dedicate many hours of their day to study and analysis; it's simply impossible, materially impossible. I'll

[13:34] give you just one example. In the last mining I've been doing to develop a strategy, I also posted it on Telegram. I mined 60 million strategies for 3 days , already. So

[13:49] , that's it. Yes, yes, it's crazy. Clear. That 's enough, isn't it? Yes, yes. And out of those 60 million strategies, I kept 400, and those now need further

[14:01] analysis. A manual trader in a lifetime cannot analyze 60 million strategies; they do n't have enough time or ability to do that. It's impossible and probably pointless because, in the end, what

[14:14] I've seen in the world of manual trading is that manual traders don't just follow certain patterns. The experience of looking at

[14:28] patterns. The experience of looking at charts, of seeing situations that aren't just price action, that aren't just how a particular candle is moving with speed, there's something there, an intangible that I

[14:45] think sometimes escapes us, that can even escape them , but I think that's what really makes the difference between a successful manual trader and just another manual trader , let's put it that way.

[14:58] manual traders who make a lot of money. The ones that earn the most are manual ones mechanical. That is to say, well, actually there are some who do in some markets, but in general everything I see is a bit more artistic, a bit more

[15:11] learned intuition, something, as you say, that they see, that accumulates with in the end makes them take more risks in certain positions, that makes them inefficiency, but it doesn't have to do with mechanically executing something, you know?

[15:25] know some who do that and have won a lot, and are up there at Most are more artistic when they are discretionary. Yes, but even these that are more mechanical, how to put it? The

[15:39] training, the flights, the flight hours they have, for example, in terms of their risk aversion, is something they have had to do little by little. little by little. And that, hm, that also has an

[15:55] ethereal component, let's put it that way, which I know escapes me at least, and I escapes me at least, and I admire them tremendously. Sure, just like you, I know many of them and I have complete

[16:09] admiration for them because, of course, we see them when they're up there, but the path these people have taken

[16:21] Absolutely, absolutely. And besides, we only see this, which I think is an interesting topic are watching this. I'd like you to put this in the comments and I want to know your opinion because I recently tried to convey a message regarding the fact that

[16:34] the traders who make the most money that I've seen are discretionary traders who trade profitable ones, that is, numerically speaking, are algorithmic traders. In the achieve profitability through algorithms because they provide you with a series

[16:47] access profitability in a more sustainable way, because you have basic tools that help you at least not to lose money. And that discretionary ones, which I've seen earn a lot and do well, are more

[17:02] polarized in the sense that many people fail in that environment because it's because you have to adapt to the real-time environment, you have to see many learned intuition of things that can't be seen, they are intangible. Therefore,

[17:16] become profitable, but when they do become profitable they are also more profitability, right? Because they have that extra ability to accumulate take off. Do you see it that way too, or have you thought about it? Yes, I agree with

[17:32] you on this point. Hm. Do you know what's happening? There are certain structural inefficiencies in algorithmic trading that are very

[17:44] simple and just by ignoring them you're not going to get rich, but you can earn, let's say, a return, okay? return, okay? And also, the opportunity cost

[17:57] of pursuing those kinds of strategies is very low. low. In contrast, in manual trading, in this trading, for example, of small caps, right? Just accessing that type of

[18:09] market already entails a much higher opportunity cost in terms of accounts, account size, even equipment, etc., etc.,

[18:21] right? And you also have to keep in mind that you're going to have a number of months, if not years, where you'll be losing money every single month. And of course, that's why I say, the opportunity cost is much higher. In

[18:36] contrast, in the algorithmic one you can start with an account, for example, start with an account, for example, like those from Darwinex, right? For now, at €45 a month, well, that's not really much money. And from there you can

[18:51] even start making a profit with incredibly simple strategies. Furthermore, if I have to leave any recommendation or advice—and I'm nobody to give recommendations, but anything I can contribute to the

[19:07] simpler, the better. That's a golden rule that I think people don't value enough. day, I was talking about, for example, Open Trange Breakouts,

[19:23] strategies that have been around for 30 years, that still work, or Conors' TPS, which also works today. These are basic rule-based strategies that people don't find nowadays because, I don't know, they aren't given

[19:35] enough publicity or they don't look for the literature, in books, whatever, and they very complex strategies that have many layers of analysis and decision-making that generate a lot of branching when in reality, as you say,

[19:48] on these types of simple, replicable strategies to make it massive, right? And from here, from here, and you do it with a fleet, right? From systems. with a fleet, right? From systems. Sure, sure. And what's more, I

[20:02] believe that in algorithmic trading there are two types of people. There are simply, uh, give me a robot that I can put in and it will make me

[20:14] rich and have the Lamborghini already parked there at the door, you know? And then there are those who get so caught up in such a level of study, research and so on, that they start adding layer upon layer upon layer of

[20:30] complexity that either makes them create real complexity or they never make any progress, you know? And I have indeed encountered this over time, and that is why, for example, in some of

[20:45] the training courses I conduct, one of the first topics I cover is the issue of degrees of freedom and complexity, so that they realize how easily complexity can be introduced into a

[21:03] system, often even unconsciously, and how this brings us closer to one algorithmic trading, which is overfeiting, right? Over-optimization. So, I'll give you an example even with Strategy One and

[21:19] even with Strategy One and show you how, without selecting all possible options, the combinatorics give us a number that is greater than the give us a number that is greater than the

[21:33] put that in that image, it says, and this isn't putting all the possibilities you have, if you put them all, the number is so gigantic that it doesn't make sense. So, uh, it's something that

[21:47] impacts people, it's very important to see the importance of also guiding searches very well, of doing very simple things and above all of planning the work you are going to do. Hm, totally. The other day I had a

[22:01] conversation, let's see what you think about this, uh, regarding Strategy Quant and uh it had to do with uh I spoke with an algorithmic tr who said that for him the Strategic One is that many people use it as uh uh generating

[22:16] creating strategies without knowing where to look. And he defended the idea that the interesting thing is always to have a prior hypothesis, that is, to think, what inefficiency do I think I can exploit, and from there

[22:29] use this tool to filter, right? To search, to save yourself the work of but already oriented towards whether that makes sense or not." Do you see it differently? Look, there are different ways. That's a

[22:44] valid way, okay? But for me, it's not the only way. But for me, it's not the only way. Why? Because that way you're subjecting your search process to the constraints of your own

[22:57] understanding, while you have a tool that can bring up possibilities you would never have thought of, okay? So, it's a valid way, it's a way I use, for example,

[23:11] it's a way I use, for example, when I always say that with Strategy One you have three ways to search for strategies. One is this one we're discussing, which would be to introduce that hypothesis

[23:25] into the Strategy One ecosystem through something Wisard, which is the tool that allows us to do that. That would be one. Another would be data mining through genetic algorithms, and a third would

[23:37] brute force, What's the difference? Brute force will search you within the entire universe of possibilities, and the genetic algorithm will search you within a

[23:52] slightly more limited area, and when it finds something good, it will try to improve it. Okay? So, those would be the three possibilities. For me, none of them are bad, okay? What is true is that the configuration you have to do with

[24:07] the tool for one or the other changes drastically, that's for sure. imagine a trader wants to use Strategy One and wants to do what you do in terms of trading. The, let's say, the, how do you say this? The

[24:20] learning process, or the time it would take, for example, to develop the the best for them or which one to use. When have you calculated, more or less, starting from zero. Uh, if someone starts from

[24:34] Uh, if someone starts from scratch, uh, I think in 4 months, 3 or 4 months they can have a solid foundation to start working with. You know what the problem is? The problem we also have is that we live in

[24:48] Yes. Hm. And also, uh, Strategy One, uh, when you first open it, the first thing you see, because they've highlighted it in a

[25:00] different color, is the start button. Everything is in shades of blue and the button Yes. And they also have a pre-loaded configuration, so people see it and click the button and then they start

[25:14] seeing that, as you say, they start generating those strategies, like these wonderful curves and such. Uh, the vast majority of it is worthless, right? But people stick with that. I agree with the detractors, not only of

[25:30] tool, that it might be a risk. But that doesn't that it might be a risk. But that doesn't mean That a trader, whether a novice or a complete beginner in algorithmic trading, starts with a builder might not be a

[25:45] starts with a builder might not be a good option, as long as it's done good option, as long as it's done in an organized and guided way, and above strengths, weaknesses, and risks are explained. But these same strengths,

[25:59] weaknesses, and risks can be applied not only to builders but also to programming languages. In other words, if I say that Multichart or Python isn't the only valid tools for algorithmic trading, look, they are

[26:15] their strengths, weaknesses, and also their risks or limitations. The key is understanding them. The advantage of a builder over, for example, a

[26:27] programming language, from my point of view, is the learning curve. The learning curve for a programming language is much steeper, and not everyone is willing to go through it, while

[26:42] with a builder it's much shorter and psychologically more you can achieve more things more quickly. Fast, right? So I think the feedback is better there. Absolutely, absolutely. Also, we're already in

[26:57] the age of artificial intelligence, in the sense that obviously I wish everyone could program and have the ability to areas of their world. That's fantastic. But it's true that little by little

[27:09] we're seeing how that aspect of programming, of writing code, can tools that will allow you to do that directly, like a chatbot where I want this." And the chatbot or the AI ​​itself executes

[27:23] that code. So perhaps builders start to make more sense when compared to pure programming in trading, because I think the future of the algorithmic trader may be moving more in that direction, right? I do

[27:35] n't know if you see it that way. Yes, yes. Look, the issue of programming, I think it's at a real crossroads with the arrival of Artificial intelligence is not just a new phenomenon; it's that

[27:50] we already know there are AI systems that program AI systems that program better than 99.9% of the best programmers in the world, right? I think it was yesterday or the day before, I was reading

[28:05] I think it was yesterday or the day before, I was reading a study that showed Gemini, in terms of programming, was ranked fifth among the best programmers in the world. So,

[28:19] n't it? What's happening now? I think we also have a great opportunity right now. In fact, the other day, in one of those conversations we

[28:33] those conversations we sometimes have among traders and friends, we were talking about how we're approaching a world where what's truly important a world where what's truly important won't be knowledge, but rather,

[28:47] how to put it? It will be imagination, the ability to imagine possibilities, because if you have the ability to imagine something, there will be a tool, an

[29:01] AI, something that will make it possible . So, companies won't... Looking for someone who's really good at programming. The company is going to look for someone who's really good at imagining a solution to a problem,

[29:16] for example. So, in the world we operate in, in the world of we operate in, in the world of trading, uh, for example, uh, in the trading, uh, for example, uh, in the

[29:31] tools section that I had to create because of the sheer number of tools because of the sheer number of tools people are developing, sometimes to solve a problem, other times simply to make their lives easier.

[29:45] The thing about trading is that it's insane. It's insane, and there are things you think, this could be commercialized without any problem. And in fact, there are projects that have come out in just two months in the community that

[30:02] what do you think, man? but it's a huge opportunity for those people who have the ability to imagine solutions to problems. Yeah,

[30:16] a bit of lateral thinking, right? The fact of saying, I have a problem, the market, for example, in the case of trading, where everyone... We see the to exploit inefficiencies, how to improve them, that's what will

[30:28] make all the difference. Furthermore, I think that many people, regardless of whether you use a builder, do it through have a barrier to algorithmic trading because they always associate it with

[30:41] someone who has gone through stages of data science and programming. But in tool of algorithmic trading. Algorithmics goes beyond simply being a programmer, because if it were, programmers would all be traders and

[30:54] making millions. It's not like that. There has to be a way of thinking, a way of approaching the markets that is more logical, more focused on exploiting specific inefficiencies. That's what

[31:06] algorithmic trader or not, how you manage risk, which is... What can have a builder that gives you millions of strategies, but if you can't actually exploitable and which ones aren't, it's useless to have that many

[31:20] strategies because you don't know which ones will work or when they'll stop working, you know? Yes. And there's also a very important aspect to this, which Yes. And there's also a very important aspect to this, which

[31:36] discussions, and that's when you approach the world of algorithmic trading. I think it's also very important to have a plan of what you're going to do, and in that sense, let's say, my background in working

[31:52] with null hypotheses, with the scientific method—but the scientific method, scientific method—but the scientific method, not seen as something abstract and scientific method is simply that you say, "I have an observation, I see

[32:06] something in the market, from that observation I come up with a hypothesis, well, I think that come up with a hypothesis, well, I think that happens because of this, and now I create a series of experiments to see if that actually holds true or not."

[32:21] that actually holds true or not."

[32:36] The advantage we have today is that between builders and artificial intelligence you can do that whole process in an automated way, but you have to be organized, you have to know what you are looking for and why you are looking for it. Let me give you an example. Um, one of the questions I'm most frequently asked within the field of Strategy Quan, for example, uh, for those who don't know it, Strategy Quan works with a series of filters

[32:50] need to be configured. Because? So, as I 've mentioned, I can come up with 60 million strategies, but now I have to, well, filter all those strategies to keep the ones that

[33:04] might actually be interesting. That's done with those filters and those settings. In fact, those configurations are really Strategy One's secret sauce , okay? Therefore,

[33:17] anyone can use it, and to get the most out of it, what filters should I use?" But the question itself comes with the

[33:30] error. There is no single filter, that is, there is no standard filter. Each project will need a configuration. Because? Because depending on what you're looking for, you'll have to include some things or

[33:44] others. That's why that preliminary phase of stopping, thinking, and writing down on paper or in a notebook, " this is what I have observed, which may be a candlestick pattern," is so important. This is my hypothesis and this is the experiment I am

[33:58] my hypothesis and this is the experiment I am going to do. A very quick example. I see a candlestick pattern, which is when there are three consecutive bearish candlesticks, and the fourth is bullish. Okay? Well, that's my observation. What

[34:13] after three bearish candles there is one bullish one. What is the experiment? I'm going to do a backtest on this asset and see, I'm going to create a little robot that will

[34:28] do the backtest and now I see the results, I analyze those results and results, I analyze those results and now I verify if my hypothesis was true or not. That's it . It's that simple.

[34:41] add something there, that's super interesting, the fact of being able to use that scientific method, as you say, parameterize it with a kind of order guide so that it makes sense, right? You start from something, a

[34:54] hypothesis, and you deduce it until you reach your goal, right? Let's see if the results prove your hypothesis true, okay? That's extremely important. because I've debated this many times, I don't think there's a

[35:06] right or wrong opinion, but I'm just going to tell you the one that I'm sold and that convinces me. Which is the point, of course, in the end when in your hypothesis there might be a rebound, there is a fourth bullish candle after three bearish ones, from

[35:19] will manage to survive the passage of time is the one who understands why it specific asset. For example, the fact that you 're in, I don't know, let's imagine in a raw material, right? Imagine yourself in the pig and you know that there is a

[35:33] that bullish candle make sense in certain contexts after three bearish candles. When you add to that hypothetical deduction of basic information data, like candles, a further part that has to do with

[35:47] why the world is expressing itself in that way, that's when you can better avoid the black swans, in why suddenly an inefficiency stops being exploited and that's what Constantly. Do you see it that way, or do you think it 's unnecessary? But

[36:03] . Look, it's true that in certain situations you can infer the reason why, but look, you can infer it in some

[36:16] cases, we might think that's the reason , but many times we're not even sure that's the reason, okay? But there is a lot of information about the market

[36:31] that makes it behave the way it does , which is not accessible and will never be accessible to us. So, since we don't have that information readily available, we can see the effect it causes,

[36:47] but we don't really know where it comes from. So, this is where statistics come in, okay? Because of all this statistics come in, okay? Because of all this process that I have mentioned,

[36:59] process that I have mentioned, we would have one final and perhaps the most important one, and that is to see if the result we have obtained has a . soundness from a statistical point of view or if it

[37:13] from a statistical point of view or if it can simply occur by pure chance, by pure variance, etc., etc., right? But even at that point, which might be the most tedious for someone without statistical knowledge, artificial intelligence makes it incredibly easy because you can simply pass it all the data and say, "Hey, so, tell me if this makes statistical sense." You have to tell it, "No, I want you to make me a student tea and then I want you to run it through Mogoro off and I don't know what and do something to it, right?" right? But even at that point, which might be the most tedious for someone without statistical knowledge, artificial intelligence makes it incredibly easy because you can simply pass it all the data and say, "Hey, so, tell me if this makes statistical sense." You have to tell it, "No, I want you to make me a student tea and then I want you to run it through Mogoro off and I don't know what and do something to it, right?"

[37:54] guide you on what tests you can do. In other words, it makes our lives much easier. But to answer specifically, in some cases yes, but in others no. And dismissing those others

[38:10] because we don't know them, I think, makes us miss many valid opportunities. Another thing is how long that will Another thing is how long that will last. Of course, if you know the origin, it's

[38:23] easier to anticipate possible changes. But, as you rightly say, this type of strategy where the origin is known or inferred in a more or less

[38:35] reliable way, which we all know, I know what this one you're talking about is, the one about the pig, you know the one about the gold, about Friday,

[38:48] but you start looking and there are six, seven, you ca n't find any more, while the universe of possibilities we have on the other side is enormous. Yes, yes, that robustness methods to determine whether, even without knowing the original idea, it

[39:03] sense, the market eventually already discounts that, I understand, right? It's like, Sure, absolutely. Clear. Look, the other day, for example, I was talking Clear. Look, the other day, for example, I was talking

[39:16] and he said to me, "But Miguel, do n't you turn off the robots in the news?" Um, because there was another colleague who was doing it, and I was telling him, "No, no, but not because I'm braver or more risk-averse, no. It's

[39:32] because when I did my research, when I did my backtest, I didn't stop it when there was news; I ran it as planned. And if I do a backtest as planned. And if I do a backtest going back 10 or 15 years, in those 10 or 15 years there must have been

[39:47] not only similar news, but even more news than there might be today, but there must also have been different market regimes, there must have been black swans, and so on. That's why it

[40:01] 's so important to do a long-term backtest, and that's why what backtest, and that's why what synthetic data, and so on. Okay. Um, ah, there are many things I still

[40:18] but [laughs] also, because of this synthetic data thing." I really want to ask you now. I'd also like to ask you—and this will happen, I say this I want to ask you this. Um, also about what criteria

[40:31] be robust. Um, for example, when you use Strategy One and something comes up , could you give us a general idea of ​​what you think are the most relevant things that can lead to increased robustness? But before that, you

[40:44] like to show your track record on Darkwunx, I don't know if you have it there. Um, you have it there, yes, to share it. Yes. Um, go ahead. I have several Yes. Um, go ahead. I have several Darwinex. Um, the one I'm working on the most,

[40:59] although it's true that I've neglected it a bit lately because of the time I have with the community and all, uh, is PDCL, it's public, I mean, I've been working on it, I mean, I've been talking about it for a while now.

[41:13] About 10 days ago I was at All-Time Highs , I'm in Darwin's Gold, uh, now I'm in a Drawdown, and well, this is also something that's good for people to know because, look,

[41:27] people think that once you reach Darwin's Gold it's all uphill, but trading is like that, it has its drawdowns, I've had my drawdowns, now I'm in a drawdown, I'll get out of it and then there will be

[41:43] another drawdown. Look, I often mention this, the important thing isn't whether you go into a drawdown or not, because you will, the important thing is how you recover from those drawdowns and, above all, that it's

[41:56] a contained drawdown. I think my record in PDCL is over 2 years. I think the maximum drawdown is around 13%, when the average they always say in Darwin is around 20% because of the

[42:11] risk engine they use. So, it's good, right? It's contained. Uh, well, I'm getting on with it, doing

[42:23] my thing. Then I have another one, but that one's for experimentation, and putting a lot of things into it. Okay core strategy and then the lab, right? Kind of where... Totally,

[42:35] totally. And I wanted to ask you about the track record, because in the end, a a track record, right? Whether it makes sense to do it or not. In this case, the reason for creating that track record is a matter of wanting, as

[42:50] example, managing third-party capital. It's a decision to maintaining your strategies over time and to have a calling card for people. What's the reason for the track record? Look, the track

[43:05] record... in fact, the trash record that PDFL has, if you analyze it, you can clearly see that there are two stages, right? There's a first stage where Darwin,

[43:17] well, he's doing... It goes up, it goes down, but it doesn't quite find up, it goes down, but it doesn't quite find its footing, well, for the same reason: because it was a stage of trial and error, of trying things out, of

[43:31] analysis and so on. And then there came a moment when I said, well, people were moment when I said, well, people were asking me for training, people were asking me to help them with certain things, and I think it's

[43:44] important to have a cover letter, as you say, or a resume, if you want to call it that, to say, "Hey, this is what call it that, to say, "Hey, this is what I do, here's

[43:58] proof of what can also be achieved because, well, I also have many students who have reached the goal, right?" But anyway, regardless of that, having a short cover letter is good.

[44:12] cover letter is good. I'm not exactly a stickler for the rules either, if you want to call it that, in the sense of saying, no, you always have to demand a record, eh, if you want to train someone or something. I think

[44:26] having a record is a plus, but after so many years I've met people who don't have one and are magnificent trainers, or who have one but don't want to make it public, you know?

[44:41] Well, there comes a point when I say, well, look, there's n't say anything else. Total, total. And with respect to your track record at Darwines,

[44:53] I wanted to ask you about the strategies you use there being strategies created exclusively for that environment, for Darwines, because I've come across let's imagine, right? I don't know how you do it, I want to ask you about that. Imagine you

[45:05] work with Darwinex, you work with, I do n't know, with Axi, with uh Fondeo, with, I don't know, with different uh environments uh even, I don't know, bull, I say this because it's not are different environments where you can squeeze things out in some way; that's the

[45:17] replicate what you do in one environment in others by adjusting things. You do that in that you adapt your strategies and your portfolio to each of the platforms; you do it the same way on all of them. How do you do that?

[45:30] all of them. How do you do that? I create them in a generic way and then I create them in a generic way and then adapt them to each environment. Uh, and this is something that is also asked a lot. Look, when you perform a backtest, you

[45:43] Look, when you perform a backtest, you have to configure the data on which you do the backtest depending on the broker or the environment in which you are going to put that strategy to work. When people start out, they think that the

[45:57] data is the same for all environments, but it's completely different. In other words, the candlesticks might be the same, but the configuration of things like the point of value, the pipside, and the minimum stop distance, and there are even brokers where

[46:15] , for example, the NASDAQ has two decimal places, while another has one. So, in some cases, for example, the So, in some cases, for example, the minimum position size is 0.01, in

[46:28] minimum position size is 0.01, in others it is 0.1. Of course, all of that affects your risk, your stock distance, everything. So, all of that needs to be everything. So, all of that needs to be adjusted. If you don't adjust it, you'll surely be

[46:41] operating with a risk that isn't the risk you studied or the one you created the robot with. Look, it's not just about creating the robots, there's behind the scenes, right? Completely. Yes, yes, that's it. First you

[46:55] have the base and from there you can tweak it, modifying it so that each one adapts and also re-evaluating, because sometimes a robot can be profitable in a certain environment and in another environment,

[47:08] simply, for example, because the spread doubles, that robot is no longer profitable. So, we need to analyze it. That's super interesting, because and that means that a strategy that might have a solid but minimal advantage

[47:23] in one environment, then the spread increases a little and that breaks down. And yes, yes, it really does fall like a stone. I have charts like this as the spread rises from a certain level, for example, look, this is

[47:37] something very interesting and I'm going to take this opportunity to comment on it. Gold, uh, in recent years has had an average spread of one. Since October of last year it started to rise and now we have had for months that the

[47:54] average spread is at five. In fact, I've seen as many as 15. What's going on? I am convinced that many of the convinced that many of the strategies that are in gold and are having a

[48:07] strategies that are in gold and are having a hard time are doing so because of this; they have lost a lot of advantage due to the brutal increase that we have had and that is being maintained in gold. In fact, one of the reviews I

[48:22] have for this weekend, for example, is of my gold strategies, to re-evaluate them, because of course, it's one thing for you to say, well, there's a month when, because of a war, because of an extraordinary event, the price of

[48:39] spray goes up a lot, right? For example, during Covid, right? There is also a peak there, but we have been holding steady for several months now . So, it's something that needs to be looked at, that needs to be analyzed, because I'm sure that many of

[48:53] the strategies that are suffering now are due to this change that has occurred Completely. These are the regime changes you were talking about earlier, right? backtest, who wants to work with past data to apply it to

[49:06] the present, always needs to have representative market data first; But it also needs the ability to adapt to what you say that has never happened in gold, or at

[49:18] be smart and adapt to it, because otherwise you're going to keep fighting, chipping away at the stone, obsessed with the idea that it has to work out, and the strategy will Yes, yes, yes, totally, totally. And speaking of

[49:30] you normally use? Well, look, the truth is that I have a pretty diversified portfolio, okay? I have gold, NQ, for example, I mean the one from the PDCL, okay? I have NQ, uh, DAX, gold, I have

[49:49] uh Gen Dollar, uh, I have too, which one was the other one? I was just looking at it a little while ago, you know

[50:01] ? Ah, yes, I also have pound pound dollar, that is , I have a little bit of Forex, I also have indices, uh, and well, and the portfolio is uncorrelated, diversified and well, all this, right?

[50:16] And it's also uncorrelated in time frame, isn't it? It's true that my main timeframe is H1, but I also have strategies in H4 and I also have daily strategies, okay? And the thing is, some really cool strategies come out in the daily news

[50:30] . Clear. Yes. That's balance. The thing is, people often don't have the patience to let it work, right? Total, total. That balance, right? We know that the more augmented the system,

[50:42] more robust they are, but of course, they are so few that many people don't want to wait for them to run or prefer the adrenaline rush of right? It's difficult to maintain that balance. Totón

[50:56] . And speaking of that, as I was saying before, uh, is there any advice or idea you can give to people who want to do estates one or do algorithmic training programming to look for strategies in backtesting and

[51:09] the criteria that I think can increase the probability of robustness or of being transferable to the real world? Yes. Well, look, we were just Yes. Well, look, we were just talking about one of the threads that

[51:22] was opened as a result of a masterclass we did in the community; it was a super interesting topic. And, well, to put it another

[51:34] way, there's one thing I honestly love, and that's breaking those supposed rules of algorithmic trading that are set in stone, right? I like to try them out and see if they really work or not, right?

[51:49] So, in that sense, one of the things that are practically set in stone in algorithmic trading is that the Shar ratio, the drawdown return, the profit factor, all of these are metrics that

[52:04] improve the bigger they are, right? If this were the case, gentlemen, everyone would be making money in algorithmic trading because we would simply have to look for strategies with a high SHAR ratio, with a large drawdown return, with a

[52:16] low drawdown, and we would have, well, we did a series of studies, also working with our friend Yaome Anolí, which showed that the vast majority of these metrics do not correlate in future

[52:33] profitability and are also super-related with each other with respect to the profit factor, sorry, the net profit. In other words, if you have a strategy that gives you a net profit that makes money in the

[52:49] backtest, these metrics will most likely be good, but these metrics do not guarantee that the strategy will make money in the future , so what they are telling you is that they are

[53:02] worth very little. We need to look for other metrics. What metrics do I like? Okay, metrics that I like, for example, are a certain minimum number of operations for the sake of having, uh,

[53:16] statistical solvency, okay? A statistical significance. Well, obviously you have to make money, but once you have a high net profit once you have a high net profit , you hardly need the rest of the metrics

[53:31] in this regard. For example, there's one thing I love: winning streaks versus Okay? This is something that is very important to me,

[53:44] and I'm going to share with you what is also something very important to me. also something very important to me. Obviously, there is always a comparison between the insample period and the autosample period. In other words, what is the auto-

[53:59] sampling period? The insample period is the period in which you train the algorithm to see how it works. It's like the syllabus you give a child to study and then you give them a test on that syllabus. You do the test on

[54:13] the autosample data, okay? Obviously he has to win in Insample, he has to get a good grade on the exam you gave him, but he also has to get a good grade on that other exam that I didn't know about, but which is related, it's

[54:29] the same syllabus. Okay? Well, but it's not just about getting a good grade. For me, there is an important piece of information here, and that is that the slope of the curve in insample and in obsample must be similar or greater in obsample.

[54:46] Okay. That's difficult. That's difficult, isn't it ? That's difficult to achieve Don't believe it, don't believe it. The problem is that anymore. Most people say, "No, I

[54:59] want it to have a higher or similar drawdown return than it has in Insample," but maybe they don't stop to look at the slope of the

[55:11] right? And also because many people fall into that overfeiting we were talking about earlier in Insample, it's very easy for the curve to become a bit less attractive in Outsample, you know? Not as obviously beautiful as it was

[55:24] in Insample. So, you do a work forward for that, for example, or something similar, or is that something you don't do? Yes. Look, yes. I mean, because this is the same old question, why do we do a work forward?

[55:38] Because, of course, at Strategy One, for example, we have such a battery of robustness tests that many people fall into the trap, from my point of view, of saying, "Since I have 18 tests, I have to do them all

[55:52] now." And they don't stop to think about what information each robustness test gives you, and why What do you do, right? For example, if I have an asset, like gold, as we were talking about before, where I want to see how the

[56:08] strategy behaves under stress levels due to the spread at a specific moment, then it might make sense to do a Monte Carlo, which is a a Monte Carlo, which is a randomization with a spread range,

[56:22] randomization with a spread range, well, between 0.5 and 8, for example, which would be appropriate in the current market conditions. Well, that might make sense, but perhaps if you're with a broker that doesn't

[56:35] cause sleepage, it does n't make sense to do a Sleepage Monte Carlo, even if it's available. Well, the same applies to the World Forward and the World Forward Matrix .

[56:49] What information does the World Forward give you? And does that information apply to your process, And does that information apply to your process, your research, your project? your research, your project? If it applies, yes, and if not, then no. In fact,

[57:03] most people don't know this, but you can use the World Forward Matrix or the regular World Forward in two different ways. You can use it as a methodology for optimizing the strategy, but

[57:18] also You can use it as a filter. For example, in the last project we did in the community, I had two World Forward Matrices. One was a filter, a World Forward that

[57:35] I call a " sudden death" matrix, because if it doesn't meet those conditions, you directly eliminate the strategies. Then eliminate the strategies. Then you have another work forward matrix that you can use

[57:48] to optimize what's left, but it depends on the project. I like the World for One Matrix, evaluate whether the edge is consistent across different time splits. That is,

[58:05] I haven't just captured a moment of positive variance in the I can cut the data into any pieces I want and still see that

[58:17] the strategy doesn't degrade. Absolutely . It's what you can apply that the interesting thing is knowing when and how to use them, and really provide you with

[58:30] available, because otherwise you'd never finish and I'd give you extra information. absolutely. Of course. It's just that there are many people who take Well, I'll do it because I have it here. Yes, yes, yes. Absolutely, absolutely. Hey,

[58:45] end. I wanted to ask you, we've already talked a bit about trading, about people have gotten a good idea. I also wanted to ask you about part of what you do. You also have your channel, Winning

[58:59] all the links to below, okay? So they can find you. Uh, you're also in that educational phase you mentioned , right? In that phase where start offering that kind of knowledge. Is there anything in that

[59:12] also benefit your trading? Because I've spoken with Many traders... I don't know, the point is this. I've been talking to many present and future of the retail trader

[59:24] see it the same way in terms of creating communities, because this is such a solitary field that if you don't connect with other people, it's difficult to scale mentally, financially, and in general. It's difficult because you focus

[59:37] only on yourself, you get stuck, like a fool, just looking at what you know and meet other people, and they might provide you with some information that helps you, as a collective intelligence, to scale your trading. Are you

[59:50] that process been like? Yes, yes. And I totally agree with you, Victor, and in fact, that's what happened to me. I mean, I spent many years giving training, even a master's degree in

[1:00:05] even a master's degree in algorithmic trading and such, and in the end, there were so many students who said, "Hey, Miguel, man, fantastic, We really enjoyed the training, just like that, but what now?" Right, of

[1:00:17] course. And that's where QX Trader came from, right? It's the QX Trader came from, right? It's the community I set up at the school, and I set it up because I saw there was a real need in the community

[1:00:32] a real need in the community to feel supported, accompanied, because when I started talking about QX Trader and all that, a lot of people would say, "Oh, well, an academy, right?" And I'd

[1:00:46] tell them, "Look, inside you have training, in fact, all the training I've been giving in the master's program, it's all free, just for joining,

[1:00:59] okay?" In other words, it's not like you pay to enter and then inside you also have to pay for... No, you don't have it there. But education, while have it there. But education, while important, is the least important thing in the

[1:01:13] community. What's really important is that you're going to meet a lot of that you're going to meet a lot of people who have gone through the same problems as you, who have walked the

[1:01:26] same path you're walking, and have access to people who are have access to people who are managing millions of euros, people managing millions of euros, people who have an incredible track record or who

[1:01:39] are making a living from passing funding challenges, and be able to directly say to them, "Hey, so-and-so, look, I do n't know how to do this." We have a feedback section, which is precisely about this topic and it's incredible.

[1:01:54] about this topic and it's incredible.

[1:02:33] with the most satisfaction. So, as I said, I created So, as I said, I created this community. There's training, there's a whole list of people who need help and get feedback,

[1:02:48] people who share what they're achieving—they've funded, achieving—they've funded, withdrawn, reached their goals, different stories. And then there's also a projects section where people in

[1:03:01] the community organize themselves and create their own projects, their own workflows. And I also give

[1:03:13] masterclasses, live sessions, we do live Q&A sessions for people who are just starting out and people who are just starting out and need that live contact.

[1:03:27] Well, we're also working on organizing an in-person meet-up because, as you said, people need that contact, and well, we're really happy about it. There are already more than 300 traders

[1:03:41] in the community, which I honestly never would have imagined in my life would work so well, and people are

[1:03:53] asking for more. Many people want to come in, but the truth is that I come in, but the truth is that I only open for a few hours every few months, only open for a few hours every few months, to be honest. Well, what we

[1:04:07] could do, and also out of deference to you and your community, is, if you want, we can open the you and anyone else who wants to join this

[1:04:23] adventure. We really appreciate it, Miguel. If you'd like, here's the link below. I'll talk to you later; let me know if you want. And if anyone

[1:04:35] part of this project, which I think is spectacular, there's the link, okay? I, Miguel, truly, it's a pleasure. Uh, I think what you've Because when you create something, a community that feeds on itself,

[1:04:50] creating things there, that begin to be a living organism, right? They , I've seen this many times, that's what good communities are like where you at the bottom and then it's a cycle that feeds you, that's awesome. And I

[1:05:03] answering everything openly, because many people have these algorithmic trading, and being able to bring in profiles like yours. It's always a pleasure, man. I thank you from the bottom of my heart and I hope you have been well. Well, the

[1:05:16] truth is, these conversations feel like home, and it's a pleasure to be here. I want to thank you again, Victor, for the invitation. A big hello to your whole community, and we're here to

[1:05:32] help in any way we can. The point is to help, because when you truly help, wonderful doors open. Totally, man, We're in touch. Big hug.