I Challenged Fabio Valentini’s Trading Method (It Didn’t Go Well)

If orderflow information, as we mentioned earlier, it's the same data set telling everyone the same information, how how can it still have an edge then for everyone [music] or how can it still work for everyone? >> This is a really smart question. So, first of all, >> Fabio Valentini, you know who he is. A worldclass scalper, multi- top five finisher in the Robins World Cup and alpha researcher for multiple sevenf figureure corporate entities. Today, I want to learn one thing from him he's never shared publicly. How can your average retail trader build a strategy that actually works? >> If you identify which trigger is the most efficient in the short term and you use the trigger to align with the long-term bias, you have a model. >> But Fabio doesn't just expose how to build a strategy. He also lays down a complete step-by-step framework to prove that your strategy actually works and how to know when to abandon it entirely. >> Then you do [music] one thing called edge contribution. So you take a different information and you put in the model and you want to see if the quality of the signals [music] improve or not. So you can put absorption, you can put aggression, you can put exhaustion and you can test and see with this model of bias for this [music] specific asset which one works best. >> Do you have a systematic way where you say this is edge decay or no this losing streak is normal variance inside of the return structure of the strategy or the model [music] that you're running? >> Yes. Yes. I I I tell you exactly step by step how I do it. >> By the end of this interview, you'll have a complete framework you can use to take a trading idea, turn it into a strategy, figure out which parts are actually adding edge, pressure test whether the strategy actually works, and recognize when that strategy has stopped working. And Fabio [music] is even revealing one of his flagship trading methodologies live on the chart. Nothing in this video is financial advice. We're [music] here to study how an exceptionally successful trader actually builds, validates, and executes a working strategy. Now, let's get into it. Now, Fabio, a lot of what you do is already public. So, let's say you had to personally train a competent trader. >> Mhm. >> What What would you be able to teach him or show him that isn't already in your public material and YouTube videos? >> Mhm. I will start Brandon with uh data analysis and proper risk protection like the biggest struggle that I had during my career was building a process that let you identify what it's an actual edge. Okay. And how you can manage risk accordingly. We are humans. So the problem I think everyone have this problem. It's fighting your ego, fighting yourself. Closing the trade early when your hope tells you the market will reverse and you will go back in profit. No, we have this behavior of changing the stop- loss then the stop loss become bigger that the take profit or the inability to know when it's time to stop a strategy. You build an edge. Maybe let's talk about the opening range breakout strategy. Now it's worldwide used. I use it with volume. Okay, it's proven. It's sold it. You can test it for yourself. You can see that there is a drift on NASDAQ. So it's not something that I say there are proper research about this. Okay. So this is an edge. But what if in 2028 when a lot of people start using it, you start to see that net commission is not profitable. Okay. So all the validation side, it's what tells you it's time to stop. This strategy is not performing anymore. and everything that brings you to this level of competence. It's what you need to do before trading because you start executing a model when you know that what you are executing actually works. The other side is automatic the risk management. Okay. For example, nowadays platform one of these it's it's a function of automatic risk manager of deep charts is blocking the traders when you reach the maximum budget for the day. Okay. So let's say that you are you have a trend following strategy and you are in a really choppy day. The market is consolidating is not going anywhere. If you don't have a daily budget, you will continue to take trend following setups and if the market stays three days in consolidation, you can destroy 20% 30% of your account. But when the daily budget is met, I survive today and I come to fight another day. And this I think is one of the biggest takeaway for people that are starting and also for professional because let's separate the idea from the reality also when you are tra when you are trading for multiple years specifically if you manage big capital for yourself or for third party we are not immune for emotion we make errors okay and that's the reason you need a system that is keeping you accountable for today you had your idea you follow your plan. You lost money. Done. Go to train, go to a sauna, do something else, but don't spend another 3 hours at the chart trying to recover because you will only make it worse. And [clears throat] you know, Fabio, you would al you mentioned in the beginning the two things you would teach someone then was real risk management, proper risk management, and then also data analysis. Yeah. And I wanted to unpack that a little more in your data analysis. What does this look like? Let's say you have an idea of you see a new idea or maybe a new pattern or systematic sequence in the market. >> Yeah. >> And you say this looks interesting, maybe it works. What does it look like going from I see something that looks like it might work to I'm willing to risk money on this. This is an amazing an amazing point of view and topic that we need to cover Brendon because it's where majority of people fail. >> Yeah. >> So um before I was a completely discretional trader last years. Okay. And I came up to a realization that now with the upcoming AI and with all the tools that we have, it's just not efficient to spend days back testing. Okay. it's better to validate at least individual trigger okay with a statistics okay so [snorts] what I do I have three separate phase the first phase it's bias identification okay so I need to know on the intraday on the multi-day on the long-term where I'm going I will give you a practical example okay on Bitcoin to understand where I'm going I use onchain analysis that is like fundamentals but use on the blockchain data. The blockchain is public, so you can study everything. You can study what portfolio of more than 1,000 Bitcoin are doing at the moment. You can study what new portfolio, so retail are doing. You can study what Black Rockck it's holding in the portfolio. It's all open. Okay. So all this data I pull into one proprietary analysis model that I have and I build a bias from this level. We are supposed to bottom down the alving it's affecting the price. Okay. So I say I want to go long on Bitcoin. This is the bias identification. Okay. Then we go to what you ask the trigger. I need to know when to execute. No. And so the trigger is out. You can automate the trigger. You can see for example if an absorption works better than a demand. If a demand works better to be trendy than the order block. Okay. If the order block works better than the value area low all these test can be done in two ways. First way is for swing trading considering that it's 20 30 40 opportunities per year. I do it manually. It's really easy. It takes me one weekend to go all over the setups. No. So I have my trigger there and I have my checklist on the order flow. We are talking about hundreds thousand of setup. No. >> So automating the trigger and isolating this part and watching if I only watch at the trigger one-on-one risk-to-reward which one perform better between these four already from there you can understand which trigger is the most efficient in the short term. If you identify which trigger is the most efficient in the short term and you use the trigger to align with the long-term bias, you have a model actually. Okay. What's the problem with this Brandon? That the more you go in short term, the more edge can break easily. >> Okay. I've been really stubborn in 2025. I was trading a model that performed amazingly during the World Trading Cup that I engaged. Okay. After Trump the markets start to change I think you notice the new year session are more compressed you don't have this explosion and continuation you have explosion back inside the range explosion back inside the range so pure trend following momentum model like the one that I was using started to underperform >> and why have been stubborn because my analysis told me look Fabio don't continue to trade this because the edge the profit factor is deteriorating >> but considering that I use this model for multiple years to get amazing results. I convinced myself no let's try it another month let's try in another two months okay I could have saved a lot of money if when my data told me look switch this model because for the current market condition is not working okay would have saved probably six figures and these are all errors that bring them the trader to be able to adapt trader and machine nowadays work in the same way >> you know Fabio so a lot of traders what they'll do is come in with a strategy. >> Yeah. >> Gain some confidence in it >> and then it starts losing. Right now maybe it's edge decay. Maybe the strategy never worked. It wasn't properly validated. >> I want to go back to when you said the edge decays and it's time to turn off the strategy in that market condition. >> Yeah. How do do you have a systematic way where you say this is edge decay or no this losing streak is normal variance inside of the return structure of the strategy or the model that you're running? >> Yes. Yes. I I I tell you exactly step by step how I do it. Okay. So for the HDK we need to have a period of time in out of sample that is not respecting what were the results in the in sample. Okay. So you should start to notice an equity line that it's floating up starting to go sideways before. It doesn't go immediately down. Okay? It goes sideways. This sideways is making is making you reflect on the fact that net commission you are break even or you are slightly losing. Okay. Or you are doing profit but at a really slower pace and usually I use rolling quarter to evaluate it. Okay. Like I did with the world trading cup. Okay. As you saw the performance are not always the same. Sometimes the strategy perform better sometimes the strategy perform worse. And what I do I use the three months evaluation. Okay, with scalping because in three months I have more than 100 execution. So I have enough data to compare 200 execution with the previous quarter. >> Now if you do swing three or four execution, it can be a negative strike. >> Mhm. >> But 200 execution, 300 execution that flit down, you need to stop immediately the strategy. Otherwise, if you trade another quarter, you will eat the profit of the quarter before. Okay? So and and second of all I don't trade one model Brandon okay I had the years like 2023 where more than 6 months one of my model was in draw down a momentum model because the market started to compress again okay but I made I can I don't want to tell but I made money with crypto with my onchain model I made money with models based on the intraday I made model with the global money with the global macro So I think the real secret here is not being dependent on one strategy only because every strategy will meet a point where the market condition are not reflecting your idea. Okay. I think the problem of the majority of the trader Brandon is expectation >> is creating an idea of traders that can make profit in any market condition. The idea of 100% win rate. The idea of one strategy for life you study and you use it. I changed countless strategy during the years because I continue to improve on the option flow in the last years. The zerod option started to account for 60% of the notional volume per day. So when the market change when you go from a market that included zero DTE option and the volume yearbyear is doubling you cannot ignore your data if you are a intraday trader because this one it's relevant the edging activity the option flow during the day so I needed to start to study options I didn't want it it's not easy for me but I started to study option because it's necessary nowadays and even with this from a regulator perspective if you tell me Fabio can you guarantee guarantee that using all the best model you will make profit the next month specifically I cannot guarantee because the >> distribution of return is not linear no >> it's not each month you get your salary that's it will be amazing if it was like this it will be amazing because you could predict it but maybe you do two negative months one crazy month that covers more than the last six months then two break even months then one incredible and this is the biggest message. So during the years I understood that the secret Brandon is having validated edge on multiple markets diversified >> in a way that use smooth. They say the only holy grail is diversification. >> The only holy grail is diversification. You know, Fabio, the misconception or the mismatch in expectation like you say where if you really get into the data, strategies do not >> meet what you see a lot of people say they make monthly when you get into the data, right? So what specifically are you looking at numberswise then to validate a strategy and maybe that can give some people watching >> a better expectation to have >> when they're testing their strategy. >> So let's start from the going against the common expectation and going against the propaganda no on the quant system. I have one quant system the IVB it's negative risk-to-reward >> okay everyone say to you go one to 20 1 to 30 to 40 crazy risk-to-reward but you need to consider that to take this risk-to-reward you will have a really low win rates probably you will make also profit if you are good but you will have streak of losses of 10 20 trades okay so first of all it's understand that the model should have not only a good profit factor but a good recovery rate and a and a good equity line in terms of smoothness. >> I don't want to optimize for net profit between two strategy where one strategy make like this big swings and arrive to 10 million of profit and the strategy that is building consistently and arrive to 2 million of profit. I will always choose the second one because the second one is allow me to build more capital expecting less draw down. The first one it's a lottery >> because if I enter when is doing the downswing I will go probably in margin call for that account. You understand? So I don't want to say that it's not good to take some risk sometimes. I did I did in the bull run. I I put I stepped the feet on the gas and I loaded heavy because I was convinced about that analysis. But I don't see that if you want to build a stable okay portfolio this is the best way to go specifically with all your capital. >> So you would say then let's say looking at profit factor for instance >> 1.5 1.6 is that you would consider. Yeah, absolutely amazing. >> Specifically in thousand of because look when I for example I scalp I have single weeks okay where the single week profit factor it's above three. >> Mhm. But with 20 execution when you start to go to 100 to 1,000 if you if you find 1,00 execution okay sustainable profit factor at four five six >> probably you are making HFD or there is an overfitting uh some overfitting there. Okay, now specifically that everyone is vibe coding strategies. >> There are a lot of look ahead bias overfitting problems. Then you put the strategy in live marketing is not performing. Okay. But when you do proper back testing, when you do proper testing of the strategy, you can find out that 1.5 for example IVB is 1.4 four, okay, but over five years of data. Okay, I find it amazing because maybe it will you will not make lifechanging money with one trade, but you will have this equity that in terms of sharp ratio beat the S&P 500 because that's the final goal, Brandon. Okay. >> Do you think discretionary traders should still go through this level of validation? I think here we will see a lot of contradiction because I know amazing discretionary trader that use their actual market understanding pulling multiple data. I will give you an example. No, you have a really strong gamma exposure level. You have an absorption there. Okay, you have confluence from multiple side order flow fundamentals and options. and you go to this guy to tell look you need to back test every it's also difficult to get this kind of depth of data in the one minute if you are doing short term and they they cost a huge amount of money okay consider that only the feed for the CBOE is $30,000 per month so majority of retail cannot afford it okay so you can of course but I think still nowadays the experience of a trader the experience of a trader beat the algo method. Okay, if you take really experienced trader, >> so you know with the discretionary trader side, what I always struggle with is if a discretionary trader is struggling and it's not really psychology and it's not risk management, how do you tell them to improve? Because how do they actually attach what happens in the market? How can they actually attach an edge or a predictive measure to what's happening in the market and say you know what we have something here let's go for it. >> Are manual back testing or should they be using let's say a very strong chain of logic and reasoning about the why of a market and and why basically moves happen specifically. We we can cover this on multiple strategy because it depends random. Let me give you an example. One of the swing model that I was using also in the live of BNP Paribus is the um it's called the earning surprise. Let me explain really briefly for the audience what it is. Okay, you have the earnings of one stock. The expectation are neutral. the stock the earnings beat the expectation by 50%. So the earnings come bigger. What happens in the market? This is a measurable event and measurable edge. It's not something that I see different than you, different than him. We all see the same stuff. Okay. >> Once you have this event chain of logic, what do you expect a big buy aggression? What's the problem when you have a big buy aggression? Slipage. market participants are not willing to buy allin-one because they will pay a huge amount of slippage on their heavy position. So in this specific case, what happens that with a really good statistical validation, the market comes back to the gap created during the earning surprise and it goes back up. Okay, so you have an asymmetrical risk-to-reward >> because you cover below the gap. But in the stock market, what the stock market does usually specifically if you are good at picking the correct stock, okay, it goes up. This event is measurable. You can take the stock, you can take the last 5 years of data, you can see expectation beating 70% probability. Okay. Drift in 60 days after the expectation 80% of the time. Is this an edge or not? Of course. >> Mhm. >> Okay. Of course. And you know that this is a measurable event. But when someone trades something from gut feeling, no let's say for example what's your strategy? I wait for breakout of the trend line. >> Your trend line will be different than mine will be different than here. So how I improve a strategy if what I'm watching is not the same that you are watching. It's not the same that a machine I allow discretionality because I do analysis. But when it comes to measuring in order flow, if at this price level there is a 500 contract executed on the NASDAQ, you can see it. He can see a machine if I give this data can read it. >> Mhm. >> You know, so the main problem for discretional trader, I think it's the inability to get objective data sets. >> Okay. Also, let's talk about single concept engulfing candle. No, for price action trader someone considered the engy candle for a week above someone with a body above. So you see if everyone consider something different >> an educator can teach you a strategy you interpret in a completely different way and you will be lost. And that's the reason my propaganda at the moment okay from where in public in in YouTube it's always guys don't trust him don't trust me trust him get the working take a data set or if you are a discretionary trader observe what happened identify repeatable pattern in data >> try to understand why they happened okay and then execute this edge but executes this edge in simulation. Don't risk real money. This is the error I made. I started with a binary option. Okay. So, you can understand that now they are not even uh they are banned from Europe. No. >> Mhm. >> But I lost a lot of money in trading trying to guess what's the correct protocol to execute something that is measurable. >> You know, and the best the the biggest error everyone makes is thinking that this is only for algorithmic trader. It's not only for algorithmic trader because also for order flow trader. If you can measure the ability of a trigger, you can improve your entry. If you can improve your entry, you can change your equity line. Or let's say an example for an intraday strategy, if you change and you say, okay, I go to break even at one:1 risk-to-reward because you notice that when you enter the market either drift or it comes back and you save 20 stop-loss, 30 stop loss per year. It's a big difference in the P&L. specifically for a strategy with not many samples, one or two per day. No. >> Yeah. And you know, you bring up a good point with um your examples. And Fabio, you know, something's always bugged me too where people say an event is bullish or bearish, but do you think it's fair to say a signal is bullish or bearish until you have actually tested the outcome of it over multiple samples? For instance, the bowling bullish engulfing pattern is theoretically bullish, right? But if I went and tested it across ES 30 minute chart, let's just say 10,000 times and it has the opposite indication. >> Yeah. >> How do I now I have to update the theoretical model in my mind of what this signal is. So even with absorption right or um you know other orderflow methodologies how how what what do you think would you not call something bullish until you run this same level of validation? >> I approach the market as a spectator. I approach the market from outside. I don't bring with me my belief when I go in the market. Okay. So if the data tell me look actually this trigger is giving you a way better rate if you use opposite. >> Mhm. >> So when there is an absorption the market is probably to give you a continuation because at the end of the day what is an absorption? A lot of pressure that cannot break one level. But if it break that level you will have an acceleration. So don't use it to time the bottom but use it to put a stop order below. wait for the breakout and accelerate. This can be useful. Okay. So when we approach the market, we need to understand first of all that no one have the crystal ball. Okay. And that data are democratic. They say the same things I say to you, to me, Andrea, to other people. Okay? So we should be guided by those by how they reflect a behavior and in my process I test also signals that for the common word of trading should be opposite. For example, no when the RSI reaches an extreme everyone will expect a reversal. So you have a swing opposite RSI you go up. But in reality the RSI got build okay as relative strength index. So it's telling you that when you go down there is a lot of selling pressure. So you are trying to catch a falling knife try not to get cut and that's what I validated from multiple models. when when I went through the validation of the IBB for example I noticed that the first assumption that I made is that okay let's test when we break the beginning of the range of the NASDAQ what cames to my mind is a strong candle okay is bringing you a bigger expansion no a small candle with a lot of delta is bringing you a bigger expansion because If it's a big candle with a lot of delta, usually the market maker and the market likes to rebalance. So you cannot cover your position immediately. You will be taken out. All this reasoning, it's really technical stuff. But if you don't spend the time to test also manually, you will be always highting >> when you start when you approach the market and you start with two words. I think that I hope that you are There is no way around. So, you know, then would you say to be a successful trader, you actually need to understand the why of what you're doing works? >> Yeah. >> Or can you validate a model just through rigor, statistical rigor without really being able to explain the underlying market mechanic that is allowing the model to work in the first place? >> I think I think before testing a strategy, you need you need to understand why this edge exists. Mhm. >> For example, the stocks have an upside skew. That's the reason IVB exist and it's persistent. Okay. The commitment of traders data force big market participant to disclose their positioning on all the asset. That's the reason the drift exists. Earning surprise force big players to avoid slippage and to build a cluster of limit orders to get in the direction of the stock. Okay. The politician have anformational edge. If you ask me why you use this, I always know from A to B how to explain the edge to you. >> Recently there was one guy a market maker Mati after I came out came up with the video of uh the big trades. Okay. So the study of how the MBO market data can influence price. Okay. He came up with a with a a study using three years of data and he find out what I find out from experience. What he find out that if you see a big market participants it's one operator. Okay. So it's you maybe opening 500 contracts on NASDAQ. to open 500 contrast of NASDAQ. Probably you have an account of nine figures. >> I won't be doing it. >> Okay. So, you are a big guy. This this order here from a statistical perspective, it's influencing the market >> 5 10 and 15 minutes later. Exactly the time that takes me to enter a trade and go out. Okay. So, I'm using this window of opportunity to get a directional bias. of course is not the only thing and remember edge can deteriorate two years this one can avoid existing but I can explain why it exist okay and the personal experience the like Freddy or like discretionary trader is contributing also to the research part because someone like Freddy with experience of a market making firm can go to a quant and can say start to check gamma exposure levels with strong absorption with this threshold. How much adverse excursion, okay, and expansion they have. If you start to see every time you reach gum exposure level that is in this condition, okay, you have usually 20 points of average excursion and 60 point of directional escarion in your position. you know the average risk-to-reward and you know that the edge is validated then you need to be really careful not to go in detail about this edge because more people specifically hedge fund because we are retail at the end of the day okay we don't have big capital but the more hedge fund execute this hedge the more the hedgek brings the profit to zero because they close the inefficiency no >> so this also brings up something I wanted to know Fabio regarding confirmation, right? Which I have trouble understanding that um that concept, but with confirmation, how do you know if confirmation is helping getting you in later? And when you're actually crafting a rule set in general for a model or a system, how do you know what pieces and parts to put together? >> Nice. So, I always start with the bias first because at the end of the day, to take a position, you need to have a direction. >> Okay. specifically for trend following. I do mainly trend following. Then you do one thing called edge contribution. So you take a different information and you put in the model and you want to see if the quality of the signals improve or not. Now I don't want to get into technical because I'm not a quant but I have people that do this for me. So go through all the process. Okay. Once you input a new information there, you can then check how the metrics change by switching. So you can put absorption, you can put aggression, you can put exhaustion, and you can test and see with this model of bias for this specific asset which one works best. >> Mhm. >> And then you need to go over what you want to do. You want to create a completely automatic system. you need to bring to platform that allow you proper testing parameter permutation Monte Carlo simulation and everything or if you want to use the edge discretional like I do for one of my edge I have like imagine if I am at the same time the analyst the person that execute and the risk team >> so I go up with a daily plan this is the budget we have guys today we are going to risk let's say $5,000 of budget today $1,000 per trade Okay, five execution maximum. Worst case scenario minus 5,000. Best case scenario know $25,000 of profit. Okay, this is the bias we are going to keep. Why? For this, this this can be option flow can be profile framing. So the structure of the profile of the volume. Okay, which trigger are we going to use? We test this this this. So we are going to engage in this area in this direction with this one. Okay, I have a budget. I have a model. I have a statistical skew of why the model works. I can explain side by side why this not this, why this, not that, why we react to resume long on the put wall, what's the activity of the market maker there. All these things is managing to keep you in an environment where you are always the owner of the model you are executing. Because the moment you start to be the person that try to justify the model to himself you will get lost Brandon because you don't know what to change when things start to go south and things will go south because in there is no trader also if you take hedge fund okay so if you take medallion they have quantitative model they have best but they had good year average years and amazing years okay so you have a floating P&L at the end of the day when emotions start to get in you need to give yourself the correct reply of why you are changing this one >> why I'm re removing from the execution system the absorption >> maybe for the last three months didn't perform good maybe I see that the market always leave me unfilled so it's a waste of time maybe I notice that it's giving me a good execution point but lower risk to reward. All this testing Brandon is why people nowadays don't want to spend time. No, >> so they just want to have something ready package. I shared for free a lot of stuff the how the order flow work, how the option flow work, how the onchain analysis work, how the swing work. But still nowadays people even with all this material they want you to give them the strategy that will work forever. >> Mhm. >> And they leave a constant fight where I say guys I can give you everything I know. But even if I give you everything I know the market it's a dynamic creature. It evolve. The market that we are seeing now with Trump is completely different from the market that we had two three years ago. The market that we saw during COVID is completely different from 2010. It's a different market. Strategy that were working before are not working anymore. Imagine an algorithmic trader from a strategy created in 2010. >> No, >> that in the last 10 years lost money, continuing to execute because it worked in 2010. But if you think about this, this is what some retails are doing. >> Why? They see that the strategy maybe it's not rewarding them but what they do they come up to to say but I like the strategy it makes sense to me and I understood something doing education also that people always prefer simple explanation. >> Mhm. They want the explanation of do this and that really simple take this demand because they want to believe that trading it's so easy drawing a line on the chart like Brandon let's be honest if trading was drawing a line on the chart watching one data set trading five minutes from the sofa like they sell you in the info product space >> we will not have hedge fund paying billions in the research every year. >> You know, Fabio, I think your answer to this next question will put give a lot of perspective into what we're discussing here. >> Let's say you run a back test on simple model, nothing too complex >> imbalance of 300%. Right? Imbalance ratio >> gives for whatever the trigger is or however the top down model looks like. >> Great test >> validated looks amazing. Amazing. >> We love it. But then when you flip that trigger of the imbalance to 301%. Completely falls apart. Or go down to 299% completely falls apart. Just that specific setting for the model was something that seemed to have made money in the >> totally overfeitting. Like if you have an edge >> you don't have look there is a chart I don't remember the name of this chart but it's a 3D chart >> that is showing you like the profitability of the system based on the parameters okay when you have in this chart sharp peak to up it means that it only work in one specific parameter if the edge is real you should have a flat surface on the top so it's telling you look this is the confidence interval from >> an imbalance from 250 to 300 works amazing. 400 is too much. >> 200 is too less >> busy interval. But if an edge 300 works amazing, 301 completely destroy the edge. You just find that correct parameter for an asset and it happens a lot of time watching this constantly on Twitter like vibe coded strategy that are like equity line sharp ratio I don't know >> seven >> profit factor 24 and I was thinking like >> if you have a 24 sharp strategy if you have strategy that have a crazy expectancy with really uh short amount of time execution like on the one minute or the 15 seconds, you could really flip accounts at a speed that is crazy. Okay, you you could really build generational wealth executing this strategy for a short amount of time. >> But usually what I see then they put commission, they put slipage and the edge completely destroy. >> Mhm. >> Okay. Because sometimes they just do this simple error. They forget to account for commission and slipage and they always assume perfect execution. No one tick on slipage because maybe the strategy does one tick profit. Zero commission, zero slipage, one tick profit. You can make strategy that are crazy. No. >> Yeah. >> The error of people at the moment and the big risk is that this vibe coding is opening the door for everyone to test strategy. No, the problem is that you need specific operational procedure to test properly a strategy. Otherwise, you risk that what you see in the vibe coded test and what you see in the real market is completely different. No. >> Mhm. >> Because you don't account for overfeitting, you don't account for actual market condition and overfeitting. So that's the the point where we are at the moment. And another one to logically extend what you just answered. Let's say you build a thesis, >> put the pieces together. This is a long strategy. >> Just consider it bullish. >> Initially complete completely fails. But when you go when you short it, it actually for whatever reason works. >> Okay. >> How do you interpret this? Would you still abandon it even though it didn't match your thesis, but it did work on the opposite end? >> It depends because it sometimes it can you can lose consistent money for commission. >> Yeah. >> So the strategy can lose consistent. For example, I have some specific real accounts where I do testing. I pay a huge amount of commission, but this is the cost that I allow myself to then go with the hedge fund with big capital. And these are accounts where I just want to see the live execution that if you do in a platform you will never have the exact short-term execution accounting for slipage and stuff. No. >> So a strategy that consistently move lose money doesn't mean that if you flip it it's consistently profitable. It's it's a different topic because there are commission there are things that are outside the strategy. Okay. So I will abandon the strategy even if my thesis is correct. If I cannot find consistent edge there >> because my ego like if I start to go with my ego and say no my thesis is right. >> Mhm. >> And I execute something at the end of the day I'm shooting myself because I'm putting in a market an edge that is constantly eating. And then another problem of the portfolio is that you need to run analysis of the single model because maybe you see the portfolio going up. Mhm. >> You have six models. There are four models performing great and two models slowly losing money almost that you cannot notice. No. And you think oh the strategy it's break even is not performing. But you can see hundreds of execution the edge is decaying. >> No. >> Then in that case you have the final proof that this strategy needs to be needs to be changed. We have a lot of history of edges like this. Now I did want to ask you one final question here at the table and I have to ask so if everyone if orderflow information as we mentioned earlier it's the same data set telling everyone the same information how how can it still have an edge then for everyone or how can it still work for everyone? >> This is a really smart question. So first of all you cannot say that an edge works for everyone because an edge is is composite part okay order flow it's a data set option flow it's a data set okay commitment of traders is a data set the data set itself is not an edge okay so you take option flow for example with Freddy you have all the data of the CBO $30,000 per month This data alone without the ability to process, okay, without the ability to interpret, it's useless. >> Mhm. >> Because it's static data is like to say, look, you have all the pieces of a car. You have a Ferrari A12 superfast here. >> Mhm. >> But each single pieces that cover all the room, do you have the ability to assemble? Well, so with learn if the interpretation of orderflow or methodologies within it can be learned then if it's teachable and learnable in public wouldn't this destroy the edge as well you think? >> No. First of all uh you have infinite possibility of building a signal from a data set. Okay. So if it was only if order flow was when you see an aggression buy when you see an absorption sell okay it if this was the only signal that you can extract would be useless >> mhm >> but a data set I don't know if you know how many data you have inside one order flow candle of one minute >> it's it's >> it's impressive >> you can process for MBO you can process for bid and ask you can process for delta profile you you have just raw data that you can use to understand and to automate if you want the signal that you use in the model. Okay, if only was this. But what about the bias identification, the option flow, the profile framing? It's a mix of three four steps, okay, that create the strategy. And on top of it if you execute completely algorithmic strategy yes there are profitable edge in order flow okay just if the edge it's completely quantitative >> and it's on the extreme shortterm time frame okay this one can go through edge decay because if you share to an edge found you sell the edge I don't know to Jane Street >> with the capital that they have they will close the edge really fast they will squeeze all the money and then the edge is finished okay >> Fabio So, let's go over the last setup you took and how order flow can help you define bias and also potentially find opportunities. >> Let's go over it together, Brandon. So, uh Friday was a consolidation day. Okay. So, the main problem that people have during consolidation day, it's identifying the relevant swing. They don't understand when an impulse is finished, when they can take buy on the second one. So, they usually get confused. Mhm. >> This is a template from the tip team deep team that they did that is identifying the pressure area. So who is dominating and I want to guide you from the starting of the session till the end helping you to understand how you can refine everything here. So we open here during the session. Okay. And you can see that as soon as we open on the top this was the first drive of the opening. You see this red cluster area. This is an area where we saw previously strong absorption to the downside. Okay. So the buyers tried to push. You see this triangle? They got absorbed heavy and this area printed. So first of all, if I want to get a first confluence area of where to engage the market, this area here for me it's super relevant. What is the problem? The problem is that as you can see I don't have yet an identified swing point that starts from here. Okay. So what I want to do when we are choppy because we can see that the session started with an absorption from the sellers. We came down and then we had an absorption from the buyers. This is what I call the cage. Okay. Why the cage? Look on the right. This one was from the beginning of the session always balanced. This profile here is telling you that buyers and sellers are at fair value. So you don't have buyers taking control. You don't have sellers taking control. And I want always to engage when I have directional auction outside the value. Okay? Because this is where you can have directional movement. The first setup that I saw is this VWAP coming here. The VWAP is the volume average weighted price and it's where in terms of reload of the players you can see the maximum amount of pressure to the downside. Okay, I will mark the area and then I will show you also how I reason about execution. Okay, so we see that this area collapsed and now these sellers here told you look these players are stronger than this player. This is pure logic. Okay, the sellers took control. The next area that we had, okay, was down here. So I had all this movement from point A to point B to take my trade. Okay, this one. Then we go in technicals. The market reach here. This one. Okay, and once reaching the bottom, you start to see that the buyers take an important aggression to the upside. This one already tell you immediately from the deep swing. Okay, what's happening at the moment? The volume weighted average is engaging the price to the bottom side and all this delta candle color here are telling you that buyers are dominating this auction at the moment. Okay. Another information that I want to show you that when we came back here the market went back to the value area and you had another trade of confluence here because you are at value area low and you have you see this line. This is where the buyers took complete control. This is an absorption. So the sellers try to bring the price back and they failed again. >> Mhm. >> Let's build other contexts. I will make it even easier for you now to refine because I will remove everything that is not interesting for us at the moment once we finish the analysis and we will go on the delta. You know that the delta is the main force of the market. Okay. So when you consider taking a position like I mean reverting from the beginning that we saw, you want to have to your side a strong aggression. Okay? So you want to have on your side the sellers taking control or the buyers punching a wall. Can you see this outline of buy aggression that you had on the top? All this power was absorbed here. So they didn't add any reward. And we'll go also on this one. This one is what I call a mean reverting setup. Okay, on the one of consolidation we have on continuation of the move we have the same because as you can see the buyers are pretty aggressive at the top but the result is zero on this aggression here you also had a negative candle what does it mean that all this buy power got absorbed to the downside okay and I like to use also the delta candle as you can see to know when it's time to put my risk to zero let's say I enter here with a stop loss here when I see that the sellers take control I put my risk to zero here And then we go to the other setup that was the reload of the buyers. You can see here the same. You can see a lot of pressure from the sellers. This candle here no result. So getting completely absorbed and the buyers taking control. Now this one setup was impossible to take because was the bottom side. But when they come back to test this one is possible. You can see the sellers not making any result in this area. Now keeping this information here, I will go to the study that I've made during the years that is a proprietary model to understand when buyers or sellers are actually dominating the area. Okay? So we remove everything else because we don't need and we just put here the NASDAQ effort. Imagine it like this, Brandon. The NASDAQ effort is trying to simplify when a specific side is dominating the market. Okay. By studying multiple forces imbalance absorption aggression and path of least resistance. So if a number of contracts in contracts make easier to move the price up, it means that the book is thinner on the upside. So the market can explode there. And these are my confirmation that I use for my setup. Let me give you an example. I have this area here. When I see that the sellers take control from the box, I have also my timing this area here. So I could have tried an explosive position back to the next area here. When I come back to this area that already show me absorption. I don't know if it's visible, but you should be able to see that this is a stack of two cell control area of the sellers. This is an additional confluence that I use. And on the buy side position that we took, not the first one because it was too fast, but the second one, I have my timing and confirmation at the closing of this candle. This candle is telling me, look, it's time for you to go up. And also on position management, Brandon, how I follow my position. I trail my position below the last aggression. So I will be taken out below this low because the market I was expecting the market to continue. It take goes down and then resume. And if I want to enter, I can reload my position when we get back inside the area. But here I will be taken out from this. This is the NASDAQ effort I use. And the other one that we we saw before is to identify this wing. And then we have the signature of MBO orders. I show you here. This one you can set in multiple ways. You can set with MBO data. So to see if these 345 orders were from one single market participants or were from multiple market participant for example Brandon if I'm in a long position and I start to see that on this level you have all this aggression of the buyers but the price cannot advance instead of waiting maybe I will start to trail my stop loss below the last aggression because if this one fail the market will reverse if on this opposite side on the bottom side I start to see the buyers okay pushing the auction with 300 contract confirming the auction this is what I use when I want to trail my position okay so when I have in the body probably I cover below here why because I know that this player will come to reload the position exactly there look at this aggression at the bottom side let's go to the other setup that I show you also here how much was easy I don't know if you can see that Here there are 111 contact absorbed on the top. So the order flow is giving you just a data set random. You need to validate it with your bias and your idea. But this kind of data have countless application that you can make on the algorithmic and on the discretionary side. But if you tell me I don't know anything, I don't have a bias. Can I just buy okay a data set and be profitable? No. because it will just be the same of saying look I am a data analyst I I buy a data set and I become a data analyst you need to be a data analyst before no and also here you can see how much aggression the market supported here so you can see that the market was following up on the aggression and when they came back the buyers tried to push the auction higher but got absorbed again so this is what I usually use and also when I see huge contract like 167 it's a lot and then buyers pushing on the other side. Usually the market bottom down in this way. Okay, because it's a confirmation on the opposite side. Another thing that is really objective data that I would like you to to take a look to is the profile. No, you can measure this in price action. You cannot you can measure from the bottom swing to the top top swing and remove the price. Okay, why this one is useful? Because we can go directly in the analysis of who is dominating this auction here and we can see look here Brandon that we have the land of nowhere because this is balance and then we have what it's called discounts. Imagine you are in the Apple store. This one is the price of the new iPhone. If you go below you are discounted. You want always to mark this area. So when you take this area into account, okay, this area here, let's make with the marker so it's even easier. One and this one here, we can see that we have buyers aggression really relevant. A lot of sellers absorb and you have the value area low. Okay, when we go back to price, if you do proper testing, you can notice that price fill this level. And this is not for me for you. This is just executed order. Everyone can see it. Okay. So the buyer's aggression that executed from here was way more relevant that everything that happened here. So if you want to consider a buy, you want you want to have your confirmation but try to use your confirmation to buy from levels that are relevant for the order flow. Okay. So this one is just an additional layer that it's adding to a trader to refine the entry that is making >> and then the entire model as you said on the exit side just to clarify >> this would be more of a longer position holding >> this is uh scalp to intraday so it's taking shortterm position to get the intraday drift and then you can have also aggressive scalping aggressive scalping Brendon is not to take the main drift is taking position. Let me show you. For example, I use the order track. This one is watching a different thing of the market. Okay, this one is watching in each single candle. Who is dominating the migration of the value area? Okay, let me give you an example. The sellers are dominating on this one. You can see this big delta and you can see that the profile is pretty balanced. Okay, let's go on the next candle. We see a lot of sell aggression but candle closing by and value area still lower. When we have the confirmation that we are switching value from the closing of this value area. Why? We were in a downward drift. Value area was shifting lower. Value area now is switching higher. Here sellers were dominating. Look here who is dominating buyers. This one is what I use to load my one to one. So you will say I'm one is not this one is in the correct direction of the order flow with a really consistent method in risk-to-reward. Okay. So if you are good at understanding the main direction of the day you can scalp every single movement in that direction instead of taking one to five riskreward taking maybe three four position one to one. Why it's smart this because even consolidating market will give you profit. When you go one to five you can stay in one position for long term. Let me give you another example. This was the first one. No, we continue higher. You can see that the value area migrate higher. The candle still green. If this idea is validated, what I expect that the value area will protect and reload higher. So, I can put my position at the value area covering covering below going for another one to one. The same when I close this candle. And here is where I don't know if I would have taken a stop loss but let's no one to one also this one that goes against you maybe it will take your stop loss before okay but anyway you have opportunity where the market is compressing and you still go to take profit okay let me give you some sell example value is still switching higher but here the value is switching lower the sellers are dominating so an example for the sell side will be stop loss above the value area protected still machine gun to oneonone one the market continue lower. Okay, what I can see that this candle close here I take a target and I need to wait for the market to realign. This sell candle is still closed inside here. If we break below, I will make the same position. The market tell me no, we are not going down. The market flip from value area to value area high. My position here stop loss below the value area low. I will take another one to one another take profit. Okay, look here. We switch lower with this but it's a green candle. I don't take we stay red. And then we switch lower. We have another position. The same logic value area here. And then you say can you use the discretional analysis on this? Of course. Okay. So when you go inside the candle you can read who is dominating. Let me give you a perspective on this. This is the maximum the the point of control of the candle. No. Is which lower from the previous one? Yes. What does it mean that the value is migrating lower? Okay. Perfect. What do we have here? We have the next candle trying to engage and break this level. A lot of buyers activity, zero result. What does it mean? That there are a lot of passive order continuing. Also, if the candle is developing, when I see all this absorption, I can also try to go a little bit more aggressive and squeeze a one to two. Okay, but the objective of the model is not to build huge risk-to-reward. It's to build consistent equity and then maybe risk the equity of the profit that you made for bigger position. >> And you know, obviously with reward reward risk ratio and win rate have a symbiotic relationship in a sense. Obviously higher boid risk win rate goes down. So, we know a win rate isn't the most important thing, but with this one to one, sometimes one to two, do you have a um a a win rate that comes with this model? It's not always the same, but you can have also three four take profit in a row when you take a drift because let's take for example this one. You switch lower. First one position TP, second one unfilled. Second position TP, then you switch one another position long because you switch on this one, take profit. Okay, this one stop loss and then you start to go down. Take profit, take profit, >> of course. So you can have a positive drift. It's not distributed, okay? But you can have a positive drift. The win rate of this model is really high. But if you tell me, will you automate it? No. And I explain you why. because the automation don't take into consideration all the analysis and refinement of the position management. Let me give you an example. Let's say that you enter here. Okay, this candle close here. The model will leave the the risk here. But here I see that there is a big buyer absorption. So I trail my stop loss and I reduce the risk to zero. When I see that the drift is sustained like the one that we had in the closing range, I build my position. Let me give an example of position building. We flip higher. Okay, what does it mean that I can position myself here at the absorption or here with a stop loss below here? The market confirm I take my my first position to risk free because I can cover below this aggression and I can load another position and the extreme delta and the market continue the word then the market was closed but in days where you have drift Brandon you can build three or four rolling position with the initial risk of one. >> Yeah. So it's not linear the reward it's exponential but there are also negative days with this model for example when is extremely choppy you can have a win rate around 40 50% that is break even because it's 101 plus commission so it's not the holy grail unfortunately >> well does this only work in the in specific markets or would you think this isable >> took the time only to test NASDAQ but I think this is this is the nature of the market how the value migrate the actual data set of the orders I think it can work in other markets but I don't want to say it because I didn't do proper testing so I don't want to spread misinformation on this one >> right and so then basically your answer to that also opens up that just because a strategy works one market even in a highly correlated market like ES doesn't mean it can just jump right over to there >> I have models that are profitable in NASDAQ can lose harmonious >> and opposite they are correlate. The fact that it's correlated doesn't mean that the micro structure is the same. >> This is microructure pattern. Okay. And uh >> it can be different for a lot of assets. So also people taking profitable strategy on gold pretending that they can apply on NASDAQ can be profitable just madness because you are you are doing a big assumption. It's like you are saying uh you have been successful in trading so I also will be because you have been. It's it's not uh it's not like this. >> And when did you develop this model? >> This model here it's six months. It's got built in six months. Before it was not using the value area was using only the aggression and the big trades. But then I find out that the value area is adding a lot of value. >> It's it's a game of words but it's adding a lot of value because it's giving you how much sustain is the drift. The more the value area continue to drift up and you see building building building, the more you can build your position. And sometimes it's not like this with big order because they can be distributed in all the candles. So the next candle can retrace. You see a big order here and you close. But the value area is making sure that you keep all the drift as much as it last. And you can see also here these are directional move. Then you have also this choppy moves where you give back some profit of the market. I only trade trend following model at the moment. So it's normal that when the regime is not correct and the market is just not moving you will give some profit back of the market. It's part of the game. >> Now would you say that um with a model like this is the value area then the most important piece of this? Would you say I will say that value area associated with delta profile it's incredibly important for this model and the last tweak that I add to it is the NASDAQ effort area that is the one that I'm using also in the other model when I have okay a value migration that is going higher and also the area telling me look the pressure is skew on the upside I have all the information that I want pointing up at the same direction let me show you for example if we only filter for setup that have all the confirmation. >> We will take this value here, this value here, this value here, this value here, this value here and this value here. So we will take six trades, five take profit, one stop loss. Okay, >> here when we are migrating lower I will only take this one take profit also this one. So it's a it's a confluence of information without overfeitting because the risk is that if you continue like you know what they say Brandon they say >> if you stress the data long enough it will tell you what you want. >> Mhm. >> Okay. And you know because you were an algorithmic trader. So you you you know that if you stress but if you go on we go back and you start to see how the price these are real time how the price react to this area you can understand that the order flow is actually influencing the next drift on the short term. Okay. And not the holy grail. You can have really bad choppy moments. Now this is not an example because was an amazing buy opportunity drift. This is an example when you have choppy money. You have an amazing sniper short, an amazing sniper short, a stop-loss long. But look in this moment, one, two, three stop-loss that you take in a row. This is part of the market and it's part of how the market is developing. >> So would you say that's the biggest every strategy has a weak point? Right. So would you say that's the biggest Achilles heel for the system? >> All the trend following system suffer from uh consolidation regime and I accepted it. I accept not blame myself where I have a consolidation week. >> My goal there is to protect capital. >> So if there's consolidation and with this model for instance and obviously it's a trend following it'll struggle a bit. This is a necessary tradeoff right and >> it's it's like a physiological physiological risk for a businessman. No, if you want to run on the tourism and if you want to run on the tourism in the UAE, you need to accept that during the summer it's and you don't have a lot of tourists for two months per year. You know it, you accept it. For trader, there are cost that are commission. Can we be profitable with commission? Yes, but you need to have an edge that go over the commission. And this is the same for the trading space. No. And um this is uh of course the impact on commission on the short term is bigger but uh if you have a model that is keeping you in the correct direction of the market you can do. And also let me clarify this for everyone listening to this podcast. I don't want you to switch your brain off and think okay I will use this one. I want you to always keep your your attitude of I don't trust this model. I will test for myself. I will try simulation money. I will do my proper research and only when I'm 100% convinced that this is an actual edge, I will execute. That's the best approach that you can have. >> If if you were going to improve this if possible, >> I'm doing a lot of studies at the moment. like I spend a lot of money on quantum research and I'm studying bias system that take into account also the option flow pressure the edging activity it it will require more time but I think uh I can improve this a lot >> and you said this was developed six months ago how long does it actually take you to get to a validated model like this to start using >> this one I use at least three years of data of testing for validation and then the auto sample process requires three months of testing and watching because it's 100 of execution. So in three months you can collect maybe 300 trades >> but still I didn't plot this as a standalone strategy but as >> proprietary study and indicators because I always want people to use this in confluence with their model. I don't want that they abandon >> I want that they use the data set to improve what they already do. I don't want that they only rely on the data set because the big error that everyone can make is that you make indicators your god and you stop using your brain to go in the direction of the market. >> Y so if someone understood the entire model and everything that we just covered and they still struggled what do you think would be the problem they're having with applying this model? I think that uh every model if you take this for a specific moment in time can be profitable or negative because the edges on the long term. >> The the best advice I can give to people starting is not relying only on one model. So the diversification of the portfolio and the strategy that you use for example I show you also the mean reverting long-term setup. So trend following but on the long term. No, I have some days where this model doesn't perform and they have the long one. I have days where these two models doesn't perform and they have the crypto one. I will strongly disagree of relying only in one model Brandon because I've made countless testing and I didn't find one model that can be profitable in every market environment. So specifically for example if you engage prop firm >> where risk constraint is crucial. No because if you go underwater you can burn the prop. So you should have a diversification system or a system that at least keeps you out of the market when you say okay this one is not working for example look here >> trend following trend following you see that in all this area we don't have anything >> because the system is built in a way where it identify the best opportunity okay and it filter out when the market condition are not there if you tell me is it capable of doing this no otherwise I will be billionaire if every consolidation phase will be removed and they only take directional. It will not be a system. It will be a a money printer, you know. [laughter] >> Well, so you know another Achilles heel for trend following systems is the balance that comes with verifying the trend is in motion and not taking every signal that possibly comes. Basically, not wanting to miss the trend, but not wanting to jump in all the time hoping to get in so early, you know. So there's always that balance with early early entries with more losses, late entries but missing the trend. How would you say that? >> I have both. >> Yeah. >> This is my one daily shot. So it's my area of confidence to take my long position. This is my momentum model. No. So on this one I accept that if it breaks it's down. This is the the setups of the day. On this one is momentum. So let's say it breaks this area and it continue to print area. The momentum setup tell me Fabio you can be here but you can go here with this pressure. So as long as you keep a positive risk-to-reward and a sustain win rate why you should avoid the setup. >> Mhm. >> That's my goal. The problem is if you pretend to take a drift with a target here from the top because in this case you are overpriced on the asset. So probably the market retrace you close in fear and there is all these emotional chain of errors that you can make. >> So then with the trend following system you said you take partial profits up right? >> Yeah. >> And basically you take partial profits up until it reverses you get stopped out. Correct. Do you have a specific >> trailing? >> Yeah. Yeah. the trailing um you said you'll take at value. >> Yeah, for this model here I use two NASDAQ effort box to trail. So let me give you an example. Let's say that I jump in this position here at the first one. Okay, the first green box. >> Mhm. >> Once I print the second, I cover my stop loss at risk-free. >> Yeah. >> Okay. Let's say that I take this bounce. The market print a second box. I cover my risk free. Okay. It print another box. I cover my risk below the second box. Okay. It print another box. Okay. Here there is cost here. It print another box. There is cost here. Print another box. There is go here. It takes me out here. >> So it's not partial profits then. It's just trailing all the way. >> Just trailing. I I tested the partial profit. The problem Brendon is that specifically on NASDAQ that is an asset that can expand directionally all day. Mhm. >> Partial profit can take you out here and then you have 200 points of rally and these days where you have a really good rally pay out for all the >> partial profit that you are not taking. So it depends on the asset. If you tell me on the currency I take a lot of partial profit because the currency more mean reverting but NASDAQ have a directional skew up. So it's >> have you ever successfully deployed a partial profit strategy that uses partial profits? No partial profit. No, >> you you tested on the algo side. >> Yeah, it's it it always hurt gains more than it um reduced risk. >> It always uh too disproportionately though. >> I'm happy that you validated it because manually I I noticed that it's eating a lot of profit when you are right but at the same time when you are wrong anyway you take a full stop loss. >> Yeah. Yeah. It just risk went down small profits went down. Yeah, >> really small. So, >> never worked for me either. But Fabio, thank you so much for >> outlining this model. I think for the first time, I don't think I've seen it in your content. So, >> and I'm happy you showed it to me and you let me ask every question I could possibly want. Uh I know I'm uh I got a lot of questions. >> Thank you. >> Some people like it, some don't. Um but thank you so much again, IQ Capital. You can start your first challenge for as little as $1. Terms and conditions apply. Check the link in the description below for more details.