YT YouTube2Text

YouTube2Text → Transcripts → 4 Institutional Scalping Strategies Market Makers Keep SECRET (Automate Prop Firm Trading)

4 Institutional Scalping Strategies Market Makers Keep SECRET (Automate Prop Firm Trading) — Transcript

by Chart Fanatics · 14,343 words · 2,275 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00I made more than 30 million euro for the
0:02bank. 90% of institutional traders were
0:06going like in one place to get their
0:08trading strategies ideas. The place guys
0:11is the social science research network
0:14and I want you guys that you steal this
0:16like steal as much as you can because
0:18this one was something that really
0:20impressed me when I started.
0:22>> He traded over $15 billion in total
0:25volume at one of Europe's largest banks.
0:28Introducing the one and only Matteo
0:30[music] Conte.
0:30>> There is an industry paper called VWOF,
0:33the holy grail of day trading [music]
0:36systems. The fact that the vast majority
0:38of institutional traders are using this
0:40type of algo amplifies these directional
0:42moves.
0:43>> While trading at the banks, Matteo ran
0:45over 400 plus strategies at one time.
0:47And in this episode, he reveals the
0:49exact place that institutions find their
0:52edges. And this is in a nutshell the big
0:55difference between retail and
0:57institutional trading. What is the
0:59pipeline every single institutional
1:02traders follow? Every single trading
1:04strategy needs to have these three boxes
1:07filled. So you need to know why you're
1:09opening. You need to know where to
1:10close. And a third condition is having
1:13like these exact edges and strategies
1:15have been used to get consistent prop
1:17firm payouts [music] by moving towards
1:19automation. Matteo reveals all of this
1:22and so much more in this special episode
1:24of Chart Fanatics. Every single
1:27institutional trader that I know at
1:29least once if not every single day they
1:32send an execution order to the market.
1:34They often use
1:36>> I guess starting off like where should
1:38we begin? What are we covering today?
1:40>> Yeah. So let me start with
1:43um some facts like obviously uh I
1:47started trading just like the majority
1:49of the people that are watching is most
1:51likely like being 16 17 years old
1:54wondering what trading actually is. How
1:58do you trade like an institutional
2:00trader? But like online is very hard to
2:03find proper information, especially like
2:06fresh information straight out of the
2:09trading floor.
2:10>> And um to get to know what is the actual
2:14process, the step that I went through
2:16was going to university, studying 12, 14
2:20hours a day because like it's an
2:21extremely competitive type of job. So,
2:24you really need to have top grades just
2:26to book yourself an interview with banks
2:31or hedge funds. And if you're lucky
2:33enough like I was at the end of my five
2:36years of university,
2:38I got a seat on the trading floor. Uh I
2:42joined the bank in 2018 getting my own
2:45books in 2019 and since there like
2:49there's been a process of learning how
2:52institutional traders actually trade and
2:55it is very much different from what you
2:58usually see online. Um the very first
3:01thing that I want to say is that the
3:03first difference between retail and
3:07institutional trading is that retail
3:13chase trades.
3:16So all retail traders go after
3:21trades.
3:23On the other end, institutional traders
3:25go after validated trading strategies
3:36trading
3:40strategy. And this is in a nutshell the
3:44big difference between retail
3:47institutional trading. And this is as
3:48well the reason why there is
3:50inconsistency between the profitability
3:52of retail traders versus like having
3:55institutional traders that make money
3:58basically every single month.
4:01We are going through what is the
4:03pipeline that every single institutional
4:07traders follow.
4:08>> Mhm. which
4:11so pipeline
4:16and this pipeline starts from getting
4:18the idea for a trading strategy. And
4:22this is just the beginning of this long
4:24journey to end up to actually having um
4:28our trading strategies trading life
4:31because once we have the idea, the next
4:33step is defining a set of rules.
4:39This set of rules needs to be so
4:42specific that you can feed it to a
4:46machine
4:47and to feed it a machine it needs to be
4:50encoded.
4:54Now I imagine at this point the audience
4:56might be thinking oh two things. One is
4:59what's the difference between chasing
5:01trades and validated trading strategy?
5:03To them that might sound like the same
5:04thing.
5:05>> Yeah. Well, chasing trades, you might
5:07see like one of the usually like they
5:10point at one specific setup. They say
5:13this is a setup where like this uh can
5:16work out well for you, can make you a
5:18lot of money 70% of the times, right?
5:21But if you look for validated trading
5:24strategies means that is a strategy that
5:27if you apply it systematically every
5:31single time is going to win 70% of the
5:35trades.
5:36>> Do you see the difference? So more so
5:38rather than the idea of a trade, it's
5:40actually validated with data with actual
5:43specifics that you can look at and go
5:45and have not certainty because don't
5:47want to give across the wrong impression
5:49but almost it's the closest thing you
5:51can probably get to certainty within the
5:52markets.
5:53>> Yeah, you can you can understand what
5:54are the probabilities of success and
5:57this is the main difference. It's not
5:59just about because like a good strategy
6:01can have 70% win rate, but if you only
6:05trades that strategy as a single trade,
6:09>> it doesn't guarantee you that you're
6:10going to be right that that you're going
6:14to be on the right side of that trade in
6:16that specific moment. On the end, if you
6:18apply that strategy over 100 trades,
6:24statistically you will win 70% of them.
6:27And this step of going from a set of
6:30well- definfined rules to a piece of
6:32code, it has been for a long time the
6:35biggest barrier to entry for retail
6:38traders to start trading like an
6:41institutional trader. Lucky for you guys
6:44uh in 2023 something u magnificant
6:48happened which was like having Chpt and
6:51the other large language models going
6:53mstream which completely remove this
6:57massive barrier to entry and
7:00now that you have the ideas or you can
7:04get the ideas we will get there where to
7:06get the ideas like an institutional
7:08trader
7:10you will know shortly how to that the
7:12rules that every single training
7:14strategy needs to have,
7:16>> you can encode it leveraging the power
7:19of your CHP code or your favorite large
7:23language model. And then the next step
7:26is to back test.
7:32So here what is back testing? Back
7:35testing is taking your set of rules and
7:38replaying history. So every time your
7:41entry conditions were satisfied, you
7:45would have like a simulated trade back
7:48in time.
7:49>> And obviously you don't back test it
7:51manually like I've seen doing it online
7:53every now and then, but you leverage the
7:56fact that you translated it into a piece
7:59of code to just replay history. And this
8:02one is going to give you
8:05one single equity curve which can be a
8:08pricing then rising if the strategy is
8:11not good enough. But as an output what
8:14you get out of it is a set of statistics
8:17as we were saying that uh can tell you
8:21if you are on the right path. Like for
8:23example, if you have a positive net
8:27profit, what is the win rate of your
8:30trading strategy over the past let's say
8:32five years?
8:33>> Um what is the average trade win? So how
8:36much your strategy is winning every time
8:39you place a trade? What is the draw down
8:43of your strategy? So which draw down is
8:47the where draw down is the peak to
8:50valley of your communive P&L when you
8:54have a rough batch
8:55>> and all of these statistics are
8:57necessary to put you in the map.
9:00However, this one is not the last step
9:02because then we need to validate.
9:11>> Can you read it?
9:12>> Yes, of course. Yes. Okay, the
9:14validation is another extremely
9:17important step in institutional trading
9:19and usually is divided in two phases.
9:22One is the split test.
9:25>> Mhm.
9:26>> So splitting your data between in sample
9:29and out of sample.
9:32So in sample
9:34versus out of sample.
9:38Why is this one important? This one is
9:41important because when you work on your
9:42set of rules when you are developing
9:44your trading strategy putting code you
9:47should always work with in sample data.
9:50So let's say that you have 10 years of
9:52data here. You should purely focus like
9:56let's say we split this data
9:59>> in 80%
10:01in sample and then you have a remaining
10:0420% out of sample
10:08in your in sample data is where you're
10:10going to work on your rules adding
10:12removing parameters fine-tuning the
10:14parameters
10:16until you find a strategy that looks
10:18like it's behaving correctly in sample.
10:20Yeah,
10:21>> let's take a break for a minute there,
10:22guys, cuz a quick word from our official
10:24platform sponsor, Ninja Trader. If
10:27you've been following along on this
10:28channel, you know how much time we spend
10:30breaking down charts and market
10:32structure. And Ninja Trader is an
10:34awardwinning futures trading platform
10:36built for exactly that. You get advanced
10:39charting, depth of market tools, and
10:41market replay so you can work through
10:43historical sessions and practice in a
10:46fully simulated environment, all without
10:49risking real money. Here's how to get
10:51started. If you're new to the platform,
10:52use the link in the description below to
10:55download Ninja Trader and explore it in
10:57the simulator. You get two full weeks
11:00for free to test everything out. And if
11:03you're already trading prop, choose
11:05Ninja Trader as your platform at
11:07checkout. It's the best platform for the
11:10job. So, go check it out and let's get
11:12back to the episode. And afterwards you
11:15feed the same rules like frozen in time
11:19to unseen data. So that if your strategy
11:23is still performing good out of sample
11:26means that you didn't overfit it.
11:29>> Mhm.
11:30>> Because overfitting is when you add too
11:32many rules or parameters just to make
11:35your equity curve look nice in sample.
11:38But as you apply to unseen data, you
11:41will see something like this, like an
11:42equity curve that gets destroyed.
11:45>> And this one is going to be the first
11:47one. If you have a good result both in
11:49sample and out of sample, it means that
11:52you didn't overfeit and there is a high
11:55chance that your strategy will work well
11:58with live data as well.
12:00>> And this one is really to put it a very
12:02high level, we uh just to give a little
12:05bit of context. And the second step is
12:10the Monte Carlo analysis. Mhm.
12:13>> Monte Carlo analysis is something that
12:15like when you look at the internet looks
12:17extremely
12:19difficult to understand but uh what is
12:22really about is just taking all the
12:24traits of your back test and then either
12:28reshuffleling the order of the traits
12:30>> which is what is known as Monte Carlo
12:32reshuffle
12:34>> and basically like in this way you
12:36create multiple version of the past
12:40multiple equity parts so that you can
12:43understand if your equity curve in your
12:48back test was looking good just because
12:50of the sequence of trades or is that
12:53because you really have an underlying
12:56edge
12:56>> and the
12:58so let's just show let's say that this
13:01one was the equity curve of your yeah
13:05the original one
13:09>> and then you start changing the order of
13:11trades right But you will always start
13:13from the same point and you will always
13:16end up to the end because the traits are
13:18always the same.
13:20>> However, you the path to get there is
13:23going to be different and then you
13:26repeat this process
13:29basically thousands of times [snorts]
13:31and the bigger is the dispersion
13:34from the top equity line to the bottom
13:37or the weakest is your edge.
13:39>> Yeah. on the dread like when you
13:43see that reshuffleling the order of the
13:45trades
13:47you have a very tight distribution
13:50around
13:52the average it means that no matter what
13:54is the order to which your trades took
13:58place you still were going to have a
14:01nice tight distribution of the outcomes
14:03>> you're showing the edge is very strong
14:05>> exactly
14:05>> regardless of how the trades are
14:07shuffled so when we say shuffled to for
14:09the audience
14:10in terms of that it's not changing the
14:12strategy or anything. It's literally a
14:13case of let's say the original uh data
14:17was a thousand trades and of those
14:20thousand trades like the average losing
14:22streak let's say was seven and the
14:24average winning streak was five but it
14:27shows you the peaks as well. So the peak
14:28losing you know number of losing trades
14:30in a row was 15 in the original test and
14:33then the most wins in a row was 15 just
14:36for example random numbers. But then the
14:38the reshuffle was essentially putting
14:40all those trades into a random order and
14:42it might be a case where you know the
14:44the peak losing number of trades in a
14:47row was now 45.
14:48>> Exactly.
14:49>> And therefore you know to your point the
14:51further that's how when the further away
14:54those lines will get.
14:55>> Exactly. uh which means then the edge is
14:57maybe isn't as strong versus let's say
14:59if that number was 16 instead of 15 on
15:02on the reshaw or 14 that's when they're
15:04going to be closer together. So showing
15:05you that regardless of how that data is
15:08redistributed and all those trades are
15:09mixed up it's actually seeing a
15:11consistent data.
15:13>> Yeah 100%. And this one like I really
15:15like the reshuffle because it's like
15:17something very simple that literally
15:20doesn't require a lot of computing power
15:23but already like helps you to visualize
15:25like how dependent were you to that
15:28order of trades and gives you a first
15:31idea of the distribution of possible
15:33draw downs because like this the second
15:37uh um common version of the Monte Carlo
15:40is the bootstrapping and resampling.
15:43So, which is the classic
15:49>> spaghetti charts. I'm not saying because
15:51I'm Italian, but it's like it is the
15:54classic spaghetti charts, right? Where
15:57you have multiple equity lines.
16:01And um the difference between the uh
16:05reshuffling and the resampling is that
16:09the ending point of the equity curve it
16:11is different. Right? And the reason why
16:13is that is because you don't simply just
16:15reshuffle the trades
16:17>> but some of the trades appears twice. So
16:21they can appear multiple times.
16:23>> Some others don't appear at all. And
16:25that's how you reproduce
16:28multiple version 10,000 20,000
16:31simulations of the past.
16:33>> So that's where it comes back to your
16:34validation. So it's not just a trading
16:36strategy, an idea is then validated
16:38through these stress tests if you will.
16:40>> Correct. And what is cool as well is
16:43that on top of that doing something like
16:47this type of Monte Carlo you do get as
16:50an output like a distribution of
16:54outcomes.
16:55>> So like you can understand what is the
17:00expected profit of the strategy? What is
17:03the expected return if I run this
17:06strategy 20,000 times? What is the
17:11probability of getting a draw down
17:14larger than $10,000?
17:16>> Mhm. And all of that all these possible
17:20scenarios they are given by the
17:23distribution of outcomes of the
17:26simulation which can be used as well
17:29like during live trading for example if
17:31I know that after 10 trades
17:35>> okay after 10 trades I should expect as
17:39maximum draw down
17:44$10,000
17:46if If I see that there was only a 5%
17:50probability of getting a draw down
17:52larger than $10,000, like that one would
17:55be a red flag for me saying, "Okay,
17:57maybe I need to step back and stop the
17:59strategy because it's very unlikely that
18:01after 10 trades, I get a larger draw
18:04down than $10,000." And all of these are
18:07yard sticks that are provided by this
18:10validation phase that you can use to
18:13give you awareness while you're trading
18:15live.
18:17And once all of these steps are
18:19completed, so you checked that you
18:24didn't overfeit it with uh where the
18:27simplest version is running like a
18:28comparison between insample results
18:31versus out of sample. You did your Monte
18:34Carlo to see how strong your edge
18:36actually is running 10,000 20,000
18:40simulation. Only at that point we reach
18:44the final phase which is live trading.
18:53And note that
18:56with live trading is not that distution
18:58trader sits in front of the monitors
19:02waiting for the entry conditions to be
19:04satisfied.
19:05>> Just like as we did like we encoded the
19:08strategy. So like the live trading is
19:10actually all about automation
19:16where you still leverage your machine to
19:20just monitor the data that are fed to it
19:24and every time that your conditions are
19:26met opening or closing your position on
19:29your behalf. So at that point like the
19:31question might be like so all this work
19:35to have your machine trading and indeed
19:37like what is your job right like your
19:40job as an institutional trader it is
19:42doing this research
19:45>> to make sure that the strategies that
19:47you're trading live they do have a
19:50positive expected returns they do have a
19:53positive value they most likely than not
19:55they're going to have a positive P&L at
19:57the end of the Okay. And your job is
20:02only one monitoring the performance
20:10and two monitoring the risk.
20:13So this is what you're doing while
20:16you're trading strategy are trading
20:17live. And given the fact that they are
20:20automated, you can go back and restart
20:23this process of going from trading
20:26strategy idea rules and code back test
20:29validation and eventually having the
20:31strategy trading life. Futures traders,
20:34it's time to hear about Apex Trader
20:36Funding, the largest futures firm in the
20:39industry. They have completely changed
20:41the game with their new evaluations. So
20:44far, they've already paid out over $600
20:47million to traders around the world
20:49[music] just like me and you. Now, not
20:51only that, Apex actually holds the
20:53record for the largest payout to
20:56traders, which is of course with
20:58Jadecap, no stranger to chart fanatics
21:00[music]
21:00with $2.5 million in a single payout.
21:03And the person before that was Trader
21:05Kane with $2 million in payouts. Now,
21:08what sets Apex apart? Apex, you can get
21:11up to 20 accounts which no other firm
21:14offers. Now, not only that, they have
21:17extremely clear rules with their revamp.
21:21There's no MAE rule, no 5:1
21:24risk-to-reward rule, no hidden rules, no
21:26gray areas, clearly defined draw down
21:29models. So, you have two options now.
21:31end of day draw down and intraday
21:33trailing draw down which means no payout
21:35denials, no payout reviews, no
21:38discretionary decisions and no
21:41surprises. What sets them apart as well?
21:43Traders receive 100% of approved payout
21:46requests up to 20 accounts as I
21:49mentioned and one day to pass. Now the
21:51other thing that I love about Apex, not
21:53only can you get 20 accounts, but
21:55there's different account sizes. So they
21:57have 20k accounts all the way to 150k
22:00accounts. So you can get 20 of any of
22:02the account sizes that you choose. Not
22:05only that, you have various trading
22:08platforms from wealth charts to trade of
22:11to rhythmic. The choice is yours. So
22:14make sure you use the code CF [music] to
22:17get up to 90% off on evaluations right
22:21now using the link in the description
22:23below. Trader Apex today. Let's get back
22:26to the episode.
22:28>> Look at this. Like usually the re
22:32classic retail traders like just as the
22:35idea and trades live with real money
22:40the idea right after
22:42>> on the other end like as an
22:44institutional trader you need to follow
22:46this pipeline.
22:47>> Yes. And once you have a strategy that
22:49has been validated, that is proven that
22:53can make you money in the future, you
22:55finally trade it live.
22:57>> So this as you to your point, like most
22:59people will learn an idea maybe from a
23:01YouTube video or just generally from
23:03some somewhere. Very rare nowadays that
23:06people are coming up with their own
23:07ideas, but that's possible too. And but
23:09straight away from there, they're
23:11looking at uh going live most of the
23:13time. At most someone might create a few
23:16rules. Someone might you know back test
23:17a tiny bit but you know not the level of
23:20detail that is necessary to really have
23:23confidence and edge within the markets.
23:26In terms of the first two which I know
23:28we're going to focus on like the idea
23:30generation and rules but when it comes
23:32to the idea generation and those rules
23:34um and then getting encoded. So then the
23:36encoding side of things is where again
23:39as you mentioned that language models
23:41can you really help with that for the
23:43average retail trader. But in terms of
23:46that from that point on a lot of like
23:48the back test side of things is
23:49automated. Obviously you're you know
23:52giving the inputs to make it all happen
23:54but the back testing and the validation
23:56that's all an automated like being done
23:58for you process to give you that data
24:00for you to review. Is that correct?
24:02>> Yeah correct. like uh that that's the
24:04big advantage. Like nowadays even the
24:06average retail trader has access to so
24:10many welldesigned tools that allow you
24:13to run back test and run like
24:16simulations.
24:19My favorite one I'm not affiliated or
24:22anything like that but my favorite one
24:23personally is multi charts because like
24:26the programming language is extremely
24:28simple. It's very easy to understand
24:31what the code is doing which is key to
24:33understand what you're doing if you want
24:35to make money in the market as an
24:37institutional trader and then like the
24:39back testing infrastructure is very
24:42solid.
24:42>> So you can rely on the back test.
24:46you have the possibility of run um
24:50simulations
24:52using your back test like classic Monte
24:55Carlos and really like is not a matter
24:58that you as a retail trader you don't
25:01have the tools it is a matter of really
25:03like having access to this process
25:06getting to know which are the steps and
25:09the boxes that you need to check before
25:12trading your strategy life and you say
25:14that I I love that you said that it is
25:17nowadays is harder that retail traders
25:21come up with their ideas. uh because um
25:24today I really want to focus on the
25:26first two steps like uh where do the
25:29idea comes from as an institutional
25:31trader and I want you guys that you
25:33steal this like steal as much as you can
25:35because uh this one was something that
25:37um really impressed me when I started
25:40that uh so many institutional traders
25:42were going like in one place to get
25:45their trading strategies ideas.
25:47>> The place guys is the social science
25:50research network. interest.
25:53>> Okay. Why does this matter? Because 90%
25:58of institutional trading strategies or
26:01strategies applied by institutional
26:04traders on any single trading floor are
26:06coming from research papers.
26:09>> Mh. And uh the feedbacks that I got like
26:13when I started like this journey of
26:16showing like the behind the curtains
26:18obviously social trading to the broader
26:20public was that yeah but you cannot use
26:24research paper because it's hard to read
26:26them or like the edge is already
26:29decayed.
26:30>> Fair enough. But you don't copy the
26:32paper. You look at its findings
26:35>> and you look for the underlying reason.
26:38the why that strategy can make you or is
26:41supposed to make you money in the future
26:44>> only triggers the idea that allows you
26:46to frame the rules for your trading
26:49strategy and this is very important I
26:52will write it down that
26:55the goal is
26:57you don't copy
27:00the paper
27:05but what you do
27:10is analyzing the findings
27:16and you look for the why,
27:20the reason why your strategy is supposed
27:23to make you money in the future.
27:26Um
27:28I brought you four examples that uh I
27:31thought was a good like exercise to go
27:34from research paper to an actual
27:37testable strategy.
27:38>> Mhm.
27:39>> Uh which we can cover one by one. So
27:42let's write it down again that 90 to 95%
27:46of the trading strategies ideas
27:50for institutional traders are coming
27:52from research.
28:00And in particular like there is two
28:02types of research. There is academic
28:04research.
28:05>> Mhm.
28:08which is research conducted by
28:10university professors or researchers
28:14within the academia
28:17or industry
28:22which might be hedge fund managers or
28:26traders which conduct some specific
28:29research around trading strategies like
28:31momentum mean reversion or strategies on
28:33gold. um all of that and they publish
28:36them online.
28:39Website like the social science research
28:42network is like this giant collection of
28:45all this research which is 100% for free
28:49where anyone can just go and their job
28:53is to dig and trying to find something
28:56interesting. M
28:58>> again it's not a matter of reading 40
29:01pages because just like for the encoding
29:04you can leverage your machine leverage
29:08the help of large language models to
29:11crack this uh these uh papers trying to
29:15understand okay what is this research
29:17about what are the findings why is it
29:20supposed to make money you don't copy
29:23the research one to one but it is an
29:25extreme important starting point to
29:28define your trading strategies rules.
29:33Which are the rules?
29:35Let's move there and then let's have a
29:36look at the examples.
29:38Every single trading strategy
29:43needs to have three things.
29:45>> Mhm.
29:46>> An entry.
29:49So the conditions on why you should open
29:51a position,
29:54an exit, why you should close the
29:58position after it has been open.
30:00>> Mhm.
30:01>> And position sizing.
30:08Every single trading strategy needs to
30:11have needs to have these three boxes
30:14filled. So you need to know why you're
30:16opening. You need to know where to close
30:18and generally is having a take profit
30:22stop-loss. And a third condition,
30:25one that I
30:28find very often across trading
30:30strategies is having like an exit linked
30:33to time. So, for example, closing a
30:36position after 1 hour or at a specific
30:40point in time like 3 p.m. And this one
30:44is satisfied if you don't hit your take
30:47profit nor your stop-loss.
30:49>> Okay. And the third component, position
30:52sizing, which is something that I feel
30:55like is not discussed enough across uh
30:59uh retail traders because it's really
31:01like knowing how you size your trading
31:04strategy, your position can make a huge
31:07difference in both reducing the risk of
31:10blowing up your account.
31:11>> Yeah.
31:12>> And uh improve the risk adjusted return
31:16of your trading strategy. We will have a
31:18look at some very simple examples on
31:21that one as well.
31:22>> Tradzella is the all-in-one platform
31:24built to turn you into a profitable
31:26trader. And at the center of it is an AI
31:28co-pilot [music] that does the heavy
31:30lifting for you. It watches every
31:31session. It learns exactly how you trade
31:34and your strengths and your weaknesses,
31:36the patterns you keep repeating without
31:38even knowing it. And it tells you in
31:40plain language what's working [music]
31:42and what needs to change. The moment you
31:44close a trade, your AI captures it
31:46automatically. Fully sync to your broker
31:48or prop firm. Zero manual entry. You can
31:51replay any trade tick by tick, pip by
31:53pip, to see the full picture. Back test
31:55your strategy on years of real data
31:58before you ever put up real money behind
32:01it. Your AI co-pilot ties all of it
32:03together, surfacing insights you'd never
32:06find on your own and guiding every
32:08decision you make going forward.
32:10Hundreds of thousands of traders, one AI
32:13powered platform, [music] real lasting
32:15results. Now, if you want that AI
32:17co-pilot and see your trading excel, the
32:20link for Tradzella is in the description
32:21below. Use the code CF20 for 20% off
32:25your yearly subscription or CF10 for 10%
32:28off your monthly subscription. If you're
32:30a trader, you have to know about Chart
32:32Academy. Chart Academy is the first
32:34all-in-one free trading education
32:37platform. That's right, absolutely free.
32:39No catch, no upsells, no links to sell
32:42you something. All free. proven
32:44world-class traders providing an A to Z
32:47masterclass on their strategies, their
32:49concepts, and their processes that make
32:51them the best. From futures, options,
32:54crypto, stocks, and forex, everything is
32:57covered on chartacademy.com. Whether you
32:59want to fix your trading psychology or
33:01learn from a market wizard directly, you
33:04can do it all on Chart Academy. Not only
33:06that, by just spending time on Chart
33:07Academy and bettering yourself as a
33:09trader, you can get points that unlock
33:12exclusive discounts, free trials, and
33:14free accounts with your favorite prop
33:16firms and trading softwares. Don't take
33:18my word for it. Chart Academy is live.
33:21The link is in the description below.
33:23Now, let's get back to the episode. But
33:24now that we have all of this, now that
33:26we know that the rules needs to have an
33:28entry, exit position sizing, we need to
33:31have a reason why our strategy
33:35is supposed to make us money. We can
33:38move forward and have a look at the
33:40first example.
33:40>> Definitely. Let's do that.
33:41>> Should we do that?
33:42>> Yeah, sure.
33:43>> So, let's start from the first paper.
33:46And the idea here, I really want to show
33:48you how to go from a research paper to
33:51an actual tradable strategy. So let's
33:55start from the first paper which is
33:58market intraday momentum.
34:03So market
34:06intraday
34:11momentum
34:17and this one is a paper from 2018
34:21a paper that you can find on the social
34:23science research network 100% for free
34:27and the reason why I wanted to take this
34:28one as first example is because at the
34:32end of this example You should be able
34:33to understand why one of the most
34:36popular retail trading strategies, the
34:40opening range breakout
34:41>> actually worked so well over the past
34:44six years.
34:47As I said, like the first thing that we
34:50need to know about any single paper is
34:52the findings
34:59and the why.
35:02why this strategy is working, why they
35:05found these specific findings within the
35:08paper. So the findings of this paper is
35:12that overnight gap
35:21plus the first 30 minutes of trading
35:24from 9:30
35:27to 10:00 a.m. It reveals the imbalance
35:31between buyers and sellers and it
35:33predicts the performance
35:36of the last 30 minutes of the regular
35:39trading hour.
35:42What does it mean? It means that if you
35:44had a return a positive returns between
35:474 p.m. to 10, you should have a positive
35:51return even in the last 30 minutes of
35:54trading.
35:55>> This is what they analyzed in the paper.
35:57This is the finding of the paper.
36:00>> This is true.
36:00>> And then you need to ask yourself, okay,
36:02but why?
36:03>> The reason why is because after the
36:06market is closed, there is all a new set
36:09of information that gets released.
36:11>> Yeah.
36:11>> Like uh news, earnings,
36:15global positioning.
36:16>> Mhm. So
36:19news
36:21plus earnings
36:26plus global positioning
36:32which creates all of these orders
36:34between buyers and sellers and these
36:39orders collides in the first 30 minutes
36:42of trading revealing the imbalance.
36:45>> Yeah.
36:55And this imbalance if there is a lot of
36:58buyers over sellers the market is not
37:02able to digest it
37:05just in the first 30 minutes and this
37:07imbalance is carried on in the remaining
37:10part of the day and that's why you have
37:12an intraday momentum.
37:13>> Yes.
37:14>> Is it clear?
37:14>> Yeah. Yeah.
37:16And uh now that we know the findings, we
37:20know what they did like okay you check
37:23what is the performance in the first 30
37:25minutes if there is more buyers more
37:28sellers and that one allows you to
37:31predict the last 30 minutes.
37:34You might be tempted to try to develop a
37:36strategy taking the entry,
37:41the exit, and the position sizing of the
37:45strategy and replicate it. But if you
37:47try to replicate it, you're going to
37:50have a very nice equity curve up to 2018
37:57followed by a very poor equity curve
38:00because you would be victim of the alpha
38:03decay.
38:04>> Yes.
38:04>> However, we don't copy the paper. But we
38:08can take the fact that the imbalance
38:11gets revealed a finding that they have
38:15in the first 30 minutes of trading
38:18setting up an entry of a classic opening
38:21range breakout. But now we know why it
38:24is supposed to work because overnight
38:26there is all the set of informations and
38:28not all the players all the market
38:31participants are going to trade until
38:34the market is open.
38:35>> Yeah. And we can say that entry
38:38condition
38:40open a long
38:44if the price
38:49closes above the high of the range
38:53between 9:30 and 10:00 a.m. Right. If
38:56the price closes above the high of the
38:59range, we open a long position.
39:01>> Mhm.
39:03>> Exit. You could say for the long the
39:06same like putting a stop loss at the low
39:10of the range.
39:14So would be somewhere down here. Right.
39:17So here we have our
39:22stop loss
39:27and for
39:29the takerit
39:32a classic reward to risk ratio either
39:35between one or two. Okay,
39:39this needs to stay simple because we
39:42need to have the control over what we're
39:44doing.
39:49take profit. Let's say we have here the
39:51eye of the range. So would be let's say
39:54if we take one to one the TP would be up
39:58here.
39:58>> Mhm.
40:00>> And we know that we are going to open a
40:02position only if the price will reach or
40:06will close above the high of this
40:10specific range.
40:12And we can add even an additional
40:14condition like closing the position if
40:17we don't hit nor the stop loss nor the
40:20take profit let's say at
40:233:30 Eastern time. So a very simple set
40:27of rules but very well defined. Yeah.
40:30>> That allows us to translate it
40:33>> into a tradable trading strategy.
40:36>> Last one that is missing is position
40:38sizing.
40:38>> Yeah. What I always suggest during the
40:41development phase is to use a fixed
40:44amount of contracts of one.
40:46>> Why? Because makes it simple to develop
40:50the strategy. However, this is not what
40:53I would suggest to trade live. Because
40:56when you have a fixed amount of
40:59contracts like one, if you have a very
41:01wide range,
41:03>> Yeah. So high volatile days like you're
41:06going to put at risk something like
41:08$30,000. Yeah.
41:09>> On the end in narrow days you're going
41:12only going to put at risk let's say
41:14$1,000 and academia
41:17research again with the two very
41:20important papers and this is something
41:22like it is done so often on the trading
41:25floor is to make your position sizing
41:29conditional to the volatility of the
41:31instrument.
41:32>> Yes. The two papers I'm talking about is
41:35volatility managed portfolios and the
41:38second one is the impact of volatility
41:40targeting. What is the volatility
41:42targeting? Volatility targeting is to
41:45have your position sizing where you
41:48always put at risk the same amount of
41:51money. Let's say $10,000 per trade.
41:54>> Mhm.
41:56[clears throat] But the amount of
41:58contracts that you're going to
42:01trade with is going to be based on the
42:03volatility of the underlying. So like if
42:06you have a very wide range.
42:08>> Yeah.
42:09>> So a highly volatile day, you're going
42:12to use a lower number of contracts still
42:15risking 10,000.
42:17>> Yes.
42:17>> Okay. M
42:18>> on that trend like if you have low
42:20volatility day with a very narrow range
42:23you still put at risk $10,000
42:27betting or like trading with a higher
42:29number of contracts. And in this way,
42:32you don't have that highly volatile days
42:34are going to dictate the outcome of your
42:37trading strategy, but overall, they're
42:39going to smooth out
42:41the outcome of your trading strategy,
42:43improving by definition, the risk
42:46adjusted performance of your trading
42:48strategy. And this is like something
42:50like that is never discussed across uh
42:53retail traders but it's such a small
42:56simple change that can highly impact the
42:59performance of your trading strategy.
43:02And this one is just one version of the
43:03strategy. But given the fact that right
43:05now we're just brainstorming on what is
43:08the right strategy that we should test.
43:10We could have a version where we have a
43:13short version. So to go short if we
43:15break the low of the range.
43:17>> Yes. However, this one I tell it from
43:19the experience is that when you develop
43:23a strategy based on the imbalance
43:26between buyers and sellers in the first
43:2830 minutes,
43:30that's works better on long go only
43:33strategies because once you start
43:36breaking down the level,
43:38>> yeah,
43:38>> you have a lot of buyers stepping in,
43:42>> removing that imbalance so easily on the
43:44trend like on
43:46>> trending case like it's hard that more
43:50buyers steps in and you have this
43:52drifting effect making this strategy
43:54fairly effective.
43:55>> Mhm.
43:57>> We didn't copy the paper. It's not that
43:59just because the paper saw that uh we if
44:02we have a good performance in the first
44:0430 minutes we're going to be long in the
44:07last 30 minutes. But we took the concept
44:10of taking this imbalance gets reveals
44:14itself in the first 30 minutes to start
44:16defining
44:18properly the rules of a strategy that we
44:21can actually translate into a piece of
44:23code.
44:23>> Mhm. And that's like key on this process
44:27of going from
44:30research
44:31>> to a testable strategy because now like
44:35even if you're trading from home an
44:37opening range breakout now you know that
44:39the reason why your breakout is actually
44:42has been working fairly well over the
44:44past five years six years and not that
44:46this paper is from 2018
44:48>> like it is because you have informations
44:51accumulating overnight Yeah.
44:54>> Imbalance revealing in the first 30
44:56minutes and that's what is pushing the
44:59price in your direction.
45:00>> So now you start getting like more
45:04knowledge, more awareness around the why
45:07your strategy should work and the why
45:10you're doing what you're doing. Mhm.
45:12>> This is what I really find fascinating
45:15about this process of justifying what
45:18you're doing based on leveraging the
45:20work of people that are much smarter
45:22than you at the spent.
45:23>> So, well, they say like your why is so
45:25important, right? And just in life, let
45:27alone in terms of trading like knowing
45:29why you're putting a strategy to use or
45:31why you're planning to trade a strategy
45:34is so important. You know, like you said
45:36originally, a lot of people have maybe
45:38an idea, but usually that idea is just
45:40something they found, you know, come
45:42across and there's no validation, no
45:44trying. But this is not only got a
45:46thesis behind it like an actual academic
45:49thesis but taking the understanding of
45:51what that's representing as you
45:52mentioned you know with the uh overnight
45:55action and all these orders and
45:57participants waiting to step in and
45:59because they're waiting that then
46:00creates as you say an imbalance in price
46:02and our job a lot of the time in terms
46:04of edge and profitability is because you
46:07notice an imbalance in price and so you
46:10know it's showing you that full process
46:11step by step as you say and [snorts] as
46:13you say the opening range breakout is
46:15very popular. I'd say over recent years
46:17in particular. But one thing that is
46:19cool about um this and over opening
46:22range breakout like in general is that
46:25they are part of a structural limitation
46:28of the financial world in a way that
46:30there is so many players that are only
46:33allowed to trade after 9:30 Eastern
46:36time. And that's why they don't place
46:39orders before because they can't
46:43>> because if they were allowed to place
46:45the order before like this
46:48market anomaly the fact that there is
46:51like the reveal of this imbalance in the
46:54first 30 minutes would disappear. So
46:56knowing why it is working like let's say
46:58things changes and this uh
47:02usage fund or some pension funds they
47:06start trading after or like before 9:30
47:11a.m. the effect of this one might
47:13disappears and knowing that it's working
47:15because of that
47:16>> they lose that inefficiency.
47:17>> Exactly. M
47:18>> so like uh
47:20gives you a complete different
47:22perspective over what you're doing and
47:24why you're doing and why you're making
47:25the money.
47:27>> Next example. Do you have any other
47:29question on this?
47:30>> No no this one this one makes sense
47:31100%. you know, like I said, because
47:33it's a a very widely used strategy,
47:37maybe not properly tested or data
47:40collected, but um that's I guess one of
47:42the interesting things is that
47:45>> yeah, like
47:45>> uh if it's a if it is a validated
47:48strategy um even if you didn't I guess
47:51that's a strange one, right? Maybe an
47:53idea is validated not by you, but
47:56someone's obviously validated it to be
47:58be able to publish or someone
48:00potentially has validated it, but you're
48:02just using the idea of it. Uh just the
48:05concept that you've learned online, you
48:07could still be profitable, but you might
48:10not have the confidence that you need.
48:11>> Yeah. Exactly. Long long
48:12>> term. Exactly. Like if you on top of
48:15that like you get the confidence over
48:18what you're doing
48:19>> like it changes your perspective on
48:22trading completely because like you
48:24might get like let's say we develop the
48:27strategy. By the way guys the academia
48:29is saying that the first 30 minutes
48:32>> are the range that is optimal to take
48:36advantage of for any opening range
48:39breakout.
48:39>> Yeah. So and that one is like it's not
48:41monte they saying that I invite you to
48:44develop a strategy based on the rules
48:46that we discussed before to confirm that
48:49obviously the long only type of strategy
48:51works better and it's not just a matter
48:54because of the beta of the market that
48:57obviously if you have long strategies
48:59and the market is just going higher you
49:02get a better performance because that
49:04strategy performed well even in 2022
49:06when the market was bleeding and the
49:08other thing is like the risk toreward
49:12ratio to use is one one or one to2 like
49:16that one is like the optimal that you
49:19can find for
49:19>> for this strategy
49:20>> yeah for this strategy for academia
49:23>> but again the the the goal is to gain
49:26the independence of going and testing it
49:31yourself like trying to get the idea
49:37frame it with entry
49:39exit and position sizing
49:41>> and then use large language models to
49:45get versions of your strategy and try to
49:47test it yourself because then like you
49:50gain independence from anyone like you
49:52can go on the social science research
49:55network. Spend the full day finding all
49:58the strategies that do you want testing
50:00as many strategies as you want and if
50:02you keep digging you find gold.
50:04>> Of course that's how this game works.
50:08not only at home like in every single
50:10trading floor on earth.
50:13>> The second strategy or better the second
50:16paper I wanted to still be somehow close
50:21to something that is still
50:22understandable across all the retail
50:25traders
50:27which is a paper from the industry this
50:30time.
50:33>> So the previous one was purely academic.
50:36Mhm.
50:36>> This one there is an industry paper
50:39called VWOP, the holy grail of day
50:43trading systems.
50:49The biggest difference between this one
50:52and the other one again is that this one
50:54is based on research and there is a
50:56beauty in that as well like
51:02is very practical.
51:03>> Mhm. Because you have the price, right?
51:11So we have our price and our time and
51:15from the price and the trading activity
51:18at any given point in time you can get
51:21the VWOP which is the volume weighted
51:24average price.
51:25>> Yes. Say differently,
51:28the VWOP is the average price of
51:31everything traded
51:40up to any specific point
51:45weighted by the volume.
51:46>> Mhm.
51:47>> Is it clear?
51:48>> Yes. Yes.
51:49>> Cool. So this one is our VWOP
51:53and this one
51:55is our price or the stock or of the
51:58index.
52:00Again we need to go back to which were
52:02the findings
52:04of the paper and we will need the why.
52:10The findings of the paper was that if
52:12you go long when the price is above the
52:16VWOP and you go short
52:20when the price is below the VWOP,
52:26>> closing the position when the price
52:29crosses back the VWOP. So
52:33as exit we will have and we will repeat
52:36this later but as exit we will have that
52:38when the price is equal the VW whoop
52:43we close the position we can get very
52:46good returns like we can get
52:48systematically good returns and what the
52:51research that they did was on QQQ
52:55which is an exchange traded fund.
52:58>> Mhm. tracking the performance of NASDAQ
53:01100
53:07and they used a time frame of 1 minute.
53:11So instrument they use QQQ on a one
53:14minute time frame. Every time that a one
53:18minute bar was closing above the VWOP
53:23was getting longer and they were closing
53:25the position when the price was closing
53:28back
53:29>> below the VWOP.
53:30>> Yeah, exactly. With the VWOP. So exit
53:33the position. Let's put it with a and
53:36the same like going short
53:39>> on the flip.
53:40>> Yeah, exactly. on the flip when um one
53:42minute bar was closing below the VWOP
53:47and closing the position. This one would
53:49have been actually a losing trade but
53:51closing the position when it was
53:54crossing back
53:56>> to the to the VWOP line.
54:00The returns that the fund was actually
54:01like quite impressive. You can find it
54:04in the paper. Something like in the
54:05order of 671%
54:08over 5 years. But we need to go beyond
54:11the headlines. We don't care about the
54:13headlines. What we care is like
54:15understanding what they've been doing
54:17and why. Why is it possible that when
54:22the price crosses below or above, we
54:25have like these directional moves that
54:28allow them to generate 600% over five
54:32years. The reason why is behind the VWOP
54:39Argos.
54:43So every single institutional traders
54:46that I know at least once if not every
54:49single day they send an execution
54:51algorith or they send an execution order
54:54to the market. They often use VWOP algos
54:58which is a type of algorithm that tries
55:01to track and get the same execution or
55:04as close as possible to the VWOP price.
55:07>> Yes.
55:08>> Okay. So it's like a so it is a volume
55:11participation algo. If there is a lot of
55:13volume traded at a specific point,
55:16>> the algo will send more orders at that
55:20specific point in time and that one
55:22causes some heavy directional moves
55:25because you have a lot of buyers or a
55:29lot of sellers that try to track the VO
55:33price and in the moment like you have
55:36this crosses between above or below the
55:38view price. The fact that the vast
55:41majority of institutional traders are
55:43using this type of algo creates like
55:45amplifies these directional moves.
55:48>> So now we know that it might make sense
55:51to follow a logic that is related to
55:55this um uh vivop logic. They did the
55:58research. They already tested a strategy
56:00a version of a strategy using a one
56:02minute time frame.
56:04>> Yes.
56:04>> So we could do exactly the same. What I
56:06would change as first thing
56:12to get a little bit of context and a
56:14little bit of different ideas is that we
56:16don't need to deploy the same strategy
56:18on QQQ
56:20>> but we could might decide to test the
56:22strategy of ENQ so NASDAQ 100 futures if
56:27we want to use the same underlying so
56:29NASDAQ 100 but you could test it you
56:32could test the same on ES you could so
56:35S&P 500 future
56:37You could test the same on crude oil.
56:40>> You could test the same on single
56:42stocks. Like again, you're writing the
56:43code once and then you can apply that
56:46code to whatever instrument you want,
56:48right? As long as you know that you have
56:50institutional traders using this
56:52specific algos on that instrument. Yeah,
56:54>> like this logic is supposed to work
56:57across all the instruments, right?
57:00This is the first important takeaway.
57:01And then like obviously you need to
57:04define your entry.
57:08So again it's going to be to go long
57:12when the price closes above the VWOP
57:18to go short
57:22when the price
57:24closes below the VWOP.
57:26>> Mhm.
57:29exit.
57:31We can decide to simply have one exit
57:34condition.
57:36>> I'm guessing the VWAP is uh plotted on
57:38from the open.
57:40>> Yeah. Yeah. That one is another thing
57:42that you might decide like you can
57:44decide to use uh the VWOP from the open
57:48>> or you can decide to use the previous
57:51day VWOP incorporate as well. Those are
57:54like the type of parameters that is up
57:56to you to use. What I strongly suggest
57:59is to use it from the open.
58:01>> Okay? So to start uh uh considering like
58:05tracking from 9:30 a.m. If you're
58:07trading US or some markets here in
58:10Europe like starting tracking from 9 and
58:13then like uh
58:15after you get some data in start like
58:19checking what is the good uh level at
58:21which uh um or better and and then like
58:27trying to understand even during the
58:28back test what is the number of minutes
58:32or hours that you need to have which is
58:35the optimal that allows you like to
58:38benefit the most out of
58:40>> the most volume I imagine because it's
58:42from the start of the day.
58:43>> Exactly. Exactly.
58:45>> And uh lastly like obviously like uh the
58:48position sizing
58:55[snorts] and just like we discussed
58:56before always use
58:59one contract in development process. Uh
59:02if you are testing for stocks, you can
59:05might decide to use a percentage of your
59:08portfolio. Let's say you have a $100,000
59:10portfolio to allocate 10% of it just to
59:13see how it would behave or again like uh
59:20having like uh your position sizing
59:23conditional to the volatility of the
59:26underlying instrument just we have seen
59:28before. But this is again in the very
59:31first phase development phase
59:33>> always one contract and then you can
59:36test different version. Yeah, different
59:38version like afterwards once you make
59:41sure that
59:42>> the strategy that you Yeah, exactly.
59:43that you have on end is actually
59:46pointing to the right direction of
59:48profitability.
59:50>> Next one.
59:50>> Yeah, we can. Yeah. So
59:53now we have seen two examples
59:55>> and both of them I decided to take them
59:58because is topics that the classic
1:00:02retail traders release a lot
1:00:04>> and on that right now we start like
1:00:06going like with the last two examples
1:00:09>> um with stuff that is one of them more
1:00:12academic which is the one that we're
1:00:13covering now.
1:00:14>> Yeah. And the difference between this
1:00:17one and the previous one is that this
1:00:20one is the results of a collection of
1:00:22research papers.
1:00:23>> Okay? So like multiple interviews.
1:00:25>> Yes. Don't know if some of you already
1:00:27heard of it, but this one is about the
1:00:30post earnings announcement drift.
1:00:36The logic of the post earnings
1:00:38announcement drift
1:00:42is that you have the release of the
1:00:44earnings.
1:00:44>> Yep.
1:00:46>> Which is set by this vertical line
1:00:50earnings.
1:00:53And what has been found over literally
1:00:57like 60 years of research. This one is
1:00:59one of the very first inefficiencies
1:01:03properly documented. Okay.
1:01:04>> And still alive today.
1:01:06>> So it's like to the ones I say ah alpha
1:01:09decay. Yes there is alpha decay but
1:01:12there is always a different angle you
1:01:14can approach with always basing your
1:01:17your knowledge from research.
1:01:20>> You have to get creative essentially.
1:01:22>> Yeah a little bit like you need to
1:01:23that's why the advantage of testing
1:01:25different things and say okay this one
1:01:28>> is not working on this specific type of
1:01:30stock. Yeah. but maybe can work on a
1:01:33different type of stock.
1:01:34>> Okay,
1:01:34>> which we will see with this one as well
1:01:36where there is like a very good insight.
1:01:38But like you have the earnings so until
1:01:41the earnings day the stock can do
1:01:43whatever
1:01:45let's say that is rising in price then
1:01:48there is the earnings release.
1:01:51What the post earnings announcement
1:01:53drift is saying on a very high level is
1:01:56that if there is a positive surprise in
1:01:59the earnings and as
1:02:02>> a consequence a positive price reaction.
1:02:05>> Okay. Yeah.
1:02:06>> For the following days and weeks the
1:02:10stock keeps drifting higher.
1:02:14>> Would it be the same in reverse?
1:02:16>> Exactly. And the same in reverse. If
1:02:19there is a missend earnings or a
1:02:22negative price reaction the day after of
1:02:25the earnings, there is a negative drift
1:02:28that keeps going for the following days
1:02:30or weeks.
1:02:31>> Mhm.
1:02:34>> And here are [snorts] the findings
1:02:40of the research which split it across
1:02:42three papers. So the first paper is from
1:02:471968
1:02:49>> and is an empirical evaluation of
1:02:52accounting income numbers.
1:02:55Empirical evaluation
1:02:57say [laughter]
1:02:59numbers.
1:03:01Okay. And uh this one was the very first
1:03:04um observation that uh where they saw
1:03:08that if there was a positive increase in
1:03:10value the following days and weeks the
1:03:14price was keep drifting higher and this
1:03:16one was like its purest form of the post
1:03:19earnings announcement drift
1:03:21>> or drifting lower when there was a
1:03:24negative price reaction.
1:03:25>> Yeah. And this one was the first
1:03:28contribution to this uh uh price
1:03:33anomaly.
1:03:34>> Yeah.
1:03:34>> The second one was a paper from 1989
1:03:40which is delayed price response or risk
1:03:43premium in full was pied the delayed
1:03:47price response or risk premium question
1:03:49mark.
1:03:51delayed price response
1:03:54or risk premium. And the biggest
1:03:57contribution of this paper was noticing
1:04:00that this drift
1:04:03>> Yeah.
1:04:03>> was lasting for around 60 days.
1:04:06>> Okay.
1:04:07>> So 60 trading days which is
1:04:10approximately exactly 3 months
1:04:13>> Okay. Yes. So including weekends.
1:04:14>> Exactly. Exactly. Um the third paper was
1:04:18a paper from 2006 called comparing the
1:04:22post earnings announcement drift for
1:04:24surprises calculated from analyst and
1:04:28time series for
1:04:29>> that's a very long yeah but but the main
1:04:32point of this one was the introduction
1:04:35>> of the analyst
1:04:38>> consensus
1:04:39>> and the general obviously we'll probably
1:04:40have links in the description and we've
1:04:42shown them on screen but for just the
1:04:45general general idea of how much detail
1:04:47these papers have. Like one of these
1:04:49papers, how how long and detailed are
1:04:52they?
1:04:52>> They are like 30 to 60 pages.
1:04:56>> But that's the key of taking the paper,
1:04:59downloading it,
1:05:00>> feeding to the machine and then
1:05:02extracting exactly what is the finding.
1:05:05Okay, what is the reason
1:05:08>> why?
1:05:09>> Because for some people that might not
1:05:11sound like a lot, but in reality, you
1:05:13know, that's on one topic. There's
1:05:14literally one one thesis on one topic
1:05:17that's a 30 to 60 page.
1:05:20>> But but that's and that's like
1:05:22>> and this particular one has multiple
1:05:24papers.
1:05:24>> Yeah. And you need to you need to start
1:05:26changing like the perspective that it
1:05:30fits a lot is because many people spend
1:05:33so many hours studying that that the
1:05:36finding is reliable and you don't need
1:05:38to spend months on just to understand
1:05:42the finding. So in reality, you don't
1:05:44need to because of these papers and and
1:05:46their availability as well is that you
1:05:48don't need to trust it in the or you
1:05:50don't have to worry about trusting it
1:05:52because the research has been done.
1:05:54>> Yeah.
1:05:54>> To your point, it's like okay, you just
1:05:56got to find the angle, you know.
1:05:57>> Exactly. Like what where it's at now.
1:06:00>> Yeah. 100%. Like research has been done.
1:06:03Often these papers are published in
1:06:05financial journals. So like they are
1:06:07peer-reviewed across different
1:06:10academics. So like they they they say if
1:06:14it turns up being published on a journal
1:06:17means yeah what they found is actually
1:06:19true.
1:06:21>> And uh then like it's up to you like to
1:06:23take like each one of these elements.
1:06:26Yeah.
1:06:27>> And combining it on a working trading
1:06:30strategy like again on the first one we
1:06:32have the price reaction. So if we have a
1:06:36gap up on the following day
1:06:38>> Yeah.
1:06:39>> from open to close.
1:06:40>> Mhm. um you might have like the first
1:06:44layer of the post earnings announcement
1:06:47drift. The second one you know that it
1:06:49lasts for 60 days. The third one is
1:06:51introducing the concept of the analyst.
1:06:53>> Yes,
1:06:57>> the analyst what they do is before the
1:07:00earnings are released they give an
1:07:02estimate of the earnings per share.
1:07:07The insight that they had was okay let's
1:07:10not compare just on the price but let's
1:07:13see what is like the drift if is
1:07:16stronger the drift if the actual earning
1:07:19per share is larger than the estimate of
1:07:23the analyst.
1:07:24>> Mhm. So if we have a larger or stronger
1:07:29positive drift in this case and if the
1:07:33earnings per share is below
1:07:36the estimates from the analyst. So if
1:07:39there is a miss in the earnings is the
1:07:42negative worse. So it is more strong the
1:07:46negative reaction.
1:07:48>> And today this is like the standard. So
1:07:53not just using the price reaction but
1:07:57keeping consideration what is the actual
1:08:00earning per share versus the expected
1:08:03earning pressure as well. Okay,
1:08:07>> these the insights from the papers. Now
1:08:09the why why is it that not the
1:08:14informations are reflected immediately
1:08:16on the price? Because we need to keep in
1:08:19mind that if you go to any university
1:08:22course they will talk to you about
1:08:26deficient market hypothesis. Yes,
1:08:29>> the efficient market hypothesis is
1:08:30saying that as the news are released,
1:08:34all the informations are reflected
1:08:38immediately on the price.
1:08:40>> Mhm.
1:08:41>> But
1:08:42this one like in front of us, we have 60
1:08:45years of research. They say no, it's not
1:08:47the case because
1:08:48>> if that uh was the case, you wouldn't
1:08:51have this drift over the following days
1:08:54>> interest
1:08:54>> or weeks, right?
1:08:56>> Mhm. And the reason why is because
1:09:00one
1:09:01there might be a low coverage of the
1:09:04stocks.
1:09:05>> Okay.
1:09:06>> So not all the analyst
1:09:11might cover properly all the stocks. So
1:09:15we can say that information travel
1:09:18slowly.
1:09:18>> Yes.
1:09:20>> And lack of awareness almost.
1:09:22>> Yeah. Exactly. like if especially like
1:09:25if we focus on
1:09:28small cap
1:09:32or midcap
1:09:34it's not that these stocks are as much
1:09:37as follow as mega cap so Tesla invidia
1:09:42Apple
1:09:43>> because they're like the information is
1:09:45out it is reflected right away indeed
1:09:47like newer paper from 2021 show that if
1:09:51you try to apply I post earnings
1:09:53announcement drift strategies to mega
1:09:55cap you basically have no edge.
1:09:58>> However, if you apply it to
1:10:02small cap or midcap,
1:10:04the price is not reflected immediately
1:10:06because of the analyst having a smaller
1:10:10coverage across the stocks. And the
1:10:13second
1:10:15reason is liquidity constraints.
1:10:23constraints.
1:10:27What do I mean with that? Yes, maybe
1:10:29there is a stock that is followed by a
1:10:32fund manager that would like to build a
1:10:35larger position,
1:10:37>> but maybe it's not able to create or
1:10:40allocate all the capital that wants to
1:10:42allocate on a single day because there
1:10:44isn't enough liquidity. So it decides to
1:10:49spread the execution of the order over
1:10:52the following 10 days.
1:10:53>> Yeah.
1:10:54>> Or over the following couple of weeks.
1:10:56And that's what generates
1:10:58>> the drift.
1:10:59>> This drift on the upside or the
1:11:01downside.
1:11:05If we now try to take all of this and
1:11:08translate it into a trading strategy, we
1:11:11need our entry
1:11:15which again can be
1:11:19if we have a positive price reaction. So
1:11:22the price increases
1:11:25>> and
1:11:27we have a bits on the earnings. So the
1:11:30actual earnings per share is larger than
1:11:34the expected earning per share. Then we
1:11:38open a long position at the opening of
1:11:40the following day.
1:11:41>> Okay.
1:11:45>> Exit again we take it from research
1:11:48suggesting that 60 days is optimal.
1:11:51>> So let's use 60 days position
1:11:56taking them as starting point.
1:11:59>> Yeah. as suggested starting point and as
1:12:02position sizing.
1:12:08>> You could do something like either a
1:12:11percentage of your portfolio. So
1:12:14allocating for every single stocks that
1:12:16you're monitoring let's say one to 2% of
1:12:19your portfolio or a specific dollar
1:12:21amount or again like in this specific
1:12:24case you could do as a position size
1:12:29conditional to either the price movement
1:12:33or like the deviation between the
1:12:36earning per share the actual earning per
1:12:38share with the from the expected earning
1:12:41per share.
1:12:41>> [snorts]
1:12:41>> And you could try to
1:12:44check and and you could try to test
1:12:47these different variation to understand
1:12:49which one has the best risk adjusted
1:12:52returns.
1:12:52>> We in this particular one cuz like with
1:12:54the VWAP uh it was un it's quite easy to
1:12:57understand where would your stop be
1:12:59right and where would you look to exit
1:13:00that trade as we already saw with the
1:13:03diagram. Uh and same with the opening
1:13:05range breakout with this one in
1:13:07particular. What would that look like?
1:13:08because you would essentially maybe be
1:13:10the low of the previous day.
1:13:12>> No, not necessarily like um that one is
1:13:14needs to be tested
1:13:16>> like uh for that purpose like given the
1:13:19fact that here we have um
1:13:21cross-sectional type of trade
1:13:24>> and monitoring let's say we're
1:13:25monitoring 20 to 30 stocks.
1:13:30>> The simplest starting point for an exit
1:13:33is just focusing on time. So you take
1:13:36the time which was suggested from the
1:13:38research and focus on that. Afterwards
1:13:40you might add an additional layer for
1:13:43risk management purposes where you say
1:13:45okay let's put like
1:13:48let's try to put like 5% lower than the
1:13:52opening of the day or like if we have a
1:13:55gap up like using the closing before the
1:14:00earnings got released.
1:14:01>> Okay,
1:14:02>> as a stop-loss for example. But these
1:14:05are just like examples that you
1:14:07>> would you would you think that these are
1:14:09things that you would when you get to
1:14:10the sort of encoding stage and so on you
1:14:12would have these sort of rules so you
1:14:14can do your back test and everything.
1:14:15>> Yeah I would start for the big test I
1:14:17would start exactly like with this
1:14:19simplified version. to just using a time
1:14:21exit of 60 days.
1:14:23>> And then like I would once I divide
1:14:26between in sample data and out of sample
1:14:29in sample is where I would add different
1:14:33rules like for example what would happen
1:14:36if I include a stop-loss.
1:14:37>> Okay.
1:14:38>> Would improve their risk adjusted
1:14:41performance.
1:14:42>> Okay.
1:14:43>> There is where like you test
1:14:46people. Exactly. Because um why do you
1:14:49do that? Why do you do this split?
1:14:51Because you don't want that you just c
1:14:53fit your strategy. Then you notice that
1:14:56if you add a 1% stop loss,
1:15:00>> it is the best possible results that you
1:15:02can get in sample but then like you
1:15:04apply it out of sample on unseen data
1:15:06and in that way like you understand did
1:15:08I over fit or was actually
1:15:11>> the best possible version.
1:15:13>> Understood. And so it really is all a
1:15:16process that you need to follow to make
1:15:18sure that uh um you're not fooling
1:15:21yourself,
1:15:22>> you see.
1:15:24And um yeah and one one interesting
1:15:28point as well like obviously as I was
1:15:30saying
1:15:32this one is a strategy that nowadays is
1:15:34working
1:15:36the best on small cap and midcap
1:15:39>> but not necessarily like you need to
1:15:40focus on small cap and midcap in US. You
1:15:44could take the same strategy and
1:15:46applying on the European stock market,
1:15:49>> right? Because you might find that in
1:15:52Europe
1:15:54there is a better performance across
1:15:56>> I can imagine maybe the analysts are
1:15:59even less maybe.
1:16:00>> Yeah exactly like there is less coverage
1:16:02there is less interest there is less um
1:16:05>> maybe even the liquidity aspects in
1:16:07terms of the constraints.
1:16:09>> Exactly. So like see you're funny like
1:16:12getting there. You're getting there.
1:16:13You're getting there. Like all of that
1:16:15like is all points that you really
1:16:17start. This is the new way of how an
1:16:20institutional trader thinks of like
1:16:22connecting these dons thinking exactly
1:16:24just like you did right now thinking
1:16:26yeah there there is less liquidity in
1:16:29Europe so maybe I could apply
1:16:31>> this strategy on European names rather
1:16:33than just focusing on US names
1:16:36>> and um and um on the one thing that I
1:16:40want to mention if you want to try to
1:16:42develop strategies on the postix
1:16:44announcement drift that the short side
1:16:48is becoming less effective
1:16:51>> for one simple reason that
1:16:55when a company has negative results
1:16:58>> Yes.
1:16:58>> they tend to pre-annon announce it.
1:17:01>> Yes. Mhm.
1:17:01>> So that one is making a huge difference
1:17:04nowadays that you have the CEO
1:17:07saying yeah the numbers are going to be
1:17:10>> poor. [laughter]
1:17:11So like
1:17:12>> you want to get ahead of it.
1:17:13>> Yeah. doing doing a lot of risk
1:17:16management
1:17:18>> EPR before end and that's why you do
1:17:21have
1:17:21>> try and soften the blow.
1:17:22>> Yeah. And that's why you do have already
1:17:24like
1:17:26>> already some down
1:17:26>> a negative price reaction before the
1:17:30earnings is announced and that's why you
1:17:32>> so those gaps won't be as dramatic
1:17:34potential.
1:17:35>> Exactly. Exactly. And even the
1:17:37consequence might be not as dramatic.
1:17:39>> Yeah. It might be a 30-day thing or or
1:17:41just sideways because it's already
1:17:42priced. Exactly.
1:17:43>> So essentially you're trying to the
1:17:46thesis of this the foundation is that
1:17:47it's something that's in a surprise to
1:17:49the market almost.
1:17:50>> Exactly. versus uh if they're announcing
1:17:52it then it's already going to start
1:17:54getting priced into them. if they start
1:17:56putting some announcement of yeah it's
1:18:00going to be a bad so like doing some
1:18:02management of the expectations that's
1:18:05where like the p tends to die out and
1:18:08this is a tendency that the cos tends to
1:18:12do a lot
1:18:14>> before negative numbers I've seen some
1:18:16cases even in positive numbers that they
1:18:18started like saying yeah not the numbers
1:18:20are going to be much better than what
1:18:22we're
1:18:22>> so then it wouldn't be valid for that
1:18:24>> and that one exactly when you will have
1:18:26um
1:18:26>> so I guess like even though it's the CEO
1:18:28it's almost in the same category as the
1:18:30analysts right if the analysts are
1:18:32covering that oh you know this is a well
1:18:35one they're covering it a lot and they
1:18:36might be even covering it saying hey
1:18:38this is going to be negative it's going
1:18:39to be this giving their thesis uh again
1:18:41it will start to get priced in it will
1:18:43start to be expected uh so similar
1:18:45thesis if the CEO is talking about it on
1:18:47the positive side as well
1:18:48>> it's not going to be as much of a a
1:18:51shock if you will in the market and you
1:18:53probably won't see as strong of a gap.
1:18:55Maybe we still get a gap, but not as
1:18:57strong. Sometimes you might not even get
1:18:58a gap because it's already been spread
1:19:00out there to the masses and
1:19:02>> Exactly. Exactly. But again, on the
1:19:04positive side, we see it rarely, but can
1:19:06happen. It is something that you just
1:19:08need to be aware of.
1:19:08>> It's almost like because the CEOs don't
1:19:10want
1:19:11>> a very drastic negative uh impact on
1:19:16their stock because that won't come
1:19:17across well. But they love no doubt a a
1:19:21surprise positive announcement because
1:19:23then you know it's all in the headlines.
1:19:24Oh, the stock's up 10% today.
1:19:26>> 100%. 100%. And it is just something
1:19:29that it's a good to know if you're
1:19:32developing strategies like that.
1:19:34>> What's interesting as well though is uh
1:19:36these strategies so far, they're not
1:19:38complex, you know, they're not like uh
1:19:41something that's overly hard to
1:19:44understand. It's not even hard to
1:19:45understand necessarily. And a lot of the
1:19:48time when people talk about institutions
1:19:50and institutional trading, they
1:19:52automatically probably assume that there
1:19:54has to have crazy technology, uh, crazy
1:19:57information, insider information and so
1:20:00and you know the thoughts instantly go
1:20:02there versus actually I can trade like
1:20:05this really as you said the difference
1:20:07being that rather than just being an
1:20:09idea and a strategy that you may have
1:20:11learned from somewhere this is uh backed
1:20:14up by years of data, years of evidence
1:20:16and research to give more confidence and
1:20:19certainty and the framework to then
1:20:21build upon and just uh again find the
1:20:23angle for today if it's like the the
1:20:26last example being 60 years old still
1:20:28works today.
1:20:28>> Yeah. 100%.
1:20:29>> But it's just about adapting as you
1:20:31mentioned maybe it's the markets maybe
1:20:33rather than large cap you're going
1:20:36medium to low uh or small sorry um and
1:20:39then it might be instead of the US
1:20:40market you're moving over to European
1:20:42market could be the Asian market. Um, so
1:20:45these slight tweaks that again aren't
1:20:47complex. It's just for taking the
1:20:48thesis, the idea, the foundation of said
1:20:51strategy and concept um, and then
1:20:53validating it for today.
1:20:55>> Like 100% like complexity
1:20:58doesn't mean more profitable. It is
1:21:01something that you learn quickly on a
1:21:03trading floor, especially like once you
1:21:04have like a incredible infrastructure
1:21:07because you need to know what you're
1:21:10doing. Like complexity actually means
1:21:12fragility more often than not because
1:21:15you have many points where your strategy
1:21:17can break like they I made more than 30
1:21:21million euro for the bank
1:21:23>> and 80% of it were out of extremely
1:21:27simple strategies.
1:21:29>> This is so key like I cannot say exactly
1:21:32what the strategy was doing but
1:21:36it wasn't complex at all.
1:21:37>> Mhm. Fourth example is related to
1:21:41another extremely well-known anomaly
1:21:43which is the overnight
1:21:50market anomaly.
1:22:00>> Have you ever heard of this?
1:22:01>> No, I haven't. Not this one.
1:22:03>> Okay. So what is interesting about this
1:22:06one?
1:22:14Let me Yeah. Picasso
1:22:18my Bangok.
1:22:22So the market closes at 400 p.m.
1:22:25>> Mhm.
1:22:26>> Market closes Market opens at 9:30.
1:22:31closes at 4, opens at 9:30,
1:22:36closes at 4, right?
1:22:38>> Yep.
1:22:41>> There is a paper from 2008
1:22:46which is called the return difference
1:22:48between trading and non-trading hours
1:22:52like night and day.
1:22:55What they found in 2008 like across
1:22:58multiple index, across multiple equity
1:23:01index, across multiple stocks
1:23:04>> was noticing that the
1:23:08returns
1:23:09once you analyze what contributed the
1:23:12most on the performance
1:23:15of for example the S&P 500 which has
1:23:18been like amazing and constantly rising.
1:23:20>> Yeah. What they noticed was like that
1:23:2290% of these returns was coming from
1:23:26overnight holdings.
1:23:28>> So regular trading session 9:30 to 4
1:23:32basically flat and then overnight gap
1:23:36flat overnight gap and again that one
1:23:40was in 2008 around 90%. If we break it,
1:23:45if at home you do exactly the same like
1:23:49getting a strategy which as entry
1:23:55it opens a position at that goes long at
1:24:004 p.m.
1:24:02And as exit,
1:24:05you close the position. So you exit at
1:24:099:30 a.m. when the market's open. Okay,
1:24:14you will see that if you do this and you
1:24:16apply this strategy on let's say NASDAQ
1:24:19using NQ contracts
1:24:21>> from 2015 to today like June 2026 you
1:24:27have again 90%
1:24:30of the returns that are coming
1:24:33>> over
1:24:33>> from this strategy from holding the
1:24:36position overnight and only 10% if you
1:24:41were just holding
1:24:42during regular trading hours. It is
1:24:45something like extremely fascinating
1:24:48that has been there like literally for
1:24:51decades and not just on NASDAQ but
1:24:54across multiple indices and again
1:24:58everyone at home can just do this simple
1:25:01test
1:25:03>> and the reason why
1:25:08is that 90% of the returns of equity
1:25:13indexes
1:25:15is coming from overnight gap,
1:25:19>> right? And the reason why
1:25:25>> we have multiple
1:25:28line of thoughts. There is multiple
1:25:30theories. There isn't like a single
1:25:32theory that is like saying this is this
1:25:35is the exact reason why. But the first
1:25:38one is about
1:25:40overnight
1:25:44risk premium.
1:25:50So the holders of the position overnight
1:25:54needs to be rewarded by the fact that
1:25:57they are holding that position
1:25:58overnight. Yes.
1:25:59>> Where there is lower liquidity, the
1:26:02market is closed. So you need to be
1:26:04rewarded by that. What is weird however
1:26:08is that it's 90% of the returns which is
1:26:12that's why like there is some academics
1:26:14that are saying is a little bit too
1:26:15high. Um even like pract um even like
1:26:19professionals they're saying it's not
1:26:21possible that all of that is explained
1:26:24>> within that.
1:26:25>> Yeah within like overnight risk premium.
1:26:28And [snorts] the second one on the D is
1:26:30related to overnight liquidity.
1:26:35So
1:26:37overnight
1:26:43liquidity during the night there is
1:26:46informations that get released as we
1:26:48have seen before with the opening range
1:26:50breakout thing that uh new informations
1:26:54are out there is news there might be
1:26:56earnings and given the fact that the
1:26:59liquidity is not as strong as uh during
1:27:03regular trading hours, the price
1:27:06reaction might be a little bit more
1:27:08aggressive
1:27:11than it would have been during regular
1:27:14trading hours and then there is a
1:27:16reversal
1:27:18during regular trading hours. Exactly.
1:27:20So like this one like once you see it
1:27:23across like multiple days it might
1:27:26explain that overnight we have an
1:27:29overreaction and then we have like a
1:27:31movement towards the fair market value
1:27:35of the underlying uh trading instrument
1:27:38>> and this one like I feel like it's a
1:27:42very strong why
1:27:44>> and what is nice is that you can use
1:27:48like knowing is knowing that 90% of the
1:27:52returns of the index are coming
1:27:54overnight given a low liquidity.
1:27:58>> This is just a starting point like this
1:28:00one is like one of those things that
1:28:03where I want to conclude with is that
1:28:06you could take this idea and you start
1:28:09like saying okay we have seen the
1:28:11opening range breakout how it is working
1:28:13but if you're telling me that 90% of the
1:28:16returns
1:28:17are happening overnight. Mhm.
1:28:19>> What about trying to develop an opening
1:28:21range breakout that only takes place
1:28:24overnight because that's where like you
1:28:27have the biggest directional moves if
1:28:29this research is right. And this is like
1:28:32the full process of like having an idea
1:28:36having an idea triggered by research and
1:28:39then like start framing
1:28:42>> your trading strategy around it.
1:28:44>> And that's like how you go from idea
1:28:50to rules
1:28:53and the next steps would be to encode it
1:28:58and test your idea until you don't
1:29:00finally arrive to in terms of encoding
1:29:04obviously it's can't whiteboard that uh
1:29:06but what does that look like so if
1:29:08you've got the ideas like the four ideas
1:29:10we've gone through uh you have the rules
1:29:12around it so the rules I guess would be
1:29:14the criteria no
1:29:16>> and then once you have those two things.
1:29:18What is just a general idea? Maybe it's
1:29:21something we can do in the future, but
1:29:22general idea, what does that look like
1:29:24in terms of the next step when when
1:29:27inputting into a language model?
1:29:29>> So with the we need to understand that
1:29:32the language model is a tool, right? You
1:29:35we cannot pretend that is a
1:29:38senior developer that is working for
1:29:41Palanteer.
1:29:43I wish it was like that, but it's not
1:29:44there yet. they say he's going to be
1:29:46there in six months. So fingers crossed.
1:29:48>> But um the the idea is like um to follow
1:29:52an iterative process.
1:29:54>> So let's say that our entry is supposed
1:29:57to get us long at 4 p.m. when
1:29:59[clears throat] markets close. So like
1:30:01first you develop the first part. So
1:30:04like okay, I want a strategy that goes
1:30:07long at 400 p.m. Eastern time. Then you
1:30:11check if it's working fine. If it's
1:30:13working fine, then you at the next
1:30:15condition. Okay. Now, encode an exit on
1:30:18the top of what you just developed,
1:30:20which exit the position and 930M.
1:30:23>> Mhm.
1:30:23>> Then you apply to your data and you make
1:30:25sure that the logic is working fine. If
1:30:27the logic is working correctly, then you
1:30:29start working on the position sizing
1:30:31like linking the position to the
1:30:34volatility of the instrument. Have a
1:30:37larger position when the volatility is
1:30:41low. have a smaller position when the
1:30:43volatility is high and so on. So
1:30:45following all these steps with an
1:30:48iterative process until you don't have
1:30:50all the rules that you defined
1:30:53>> yes
1:30:54>> compiled so into the piece of code and
1:30:57that you know that is working the way it
1:31:00is supposed to work. So that one is like
1:31:02the process on a very high level and the
1:31:05good thing is like again there is so
1:31:07many softwares um in first place these
1:31:11multi charts that where the programming
1:31:13language is literally called easy
1:31:15language because it's so like intuitive
1:31:17to understand like for a human on how it
1:31:20needs to be coded
1:31:22>> okay
1:31:23>> you can use uh pine script so like
1:31:26threading view that is getting quite
1:31:28popular
1:31:30to follow exactly the same process of
1:31:32like you can even feed to the machine
1:31:36like uh the manuals on how to code with
1:31:39pine script
1:31:40>> and then like the machine
1:31:41>> build knowledge pool. Yeah,
1:31:43>> exactly. To build the knowledge pool
1:31:44like right now we really are in this
1:31:46space where retail traders can have this
1:31:48massive step to really start trading
1:31:51like an institutional trader would. And
1:31:53it's not that one thing one point I want
1:31:56to make that all of you needs to start
1:32:00automating your training strategies but
1:32:02if you follow the process even if you
1:32:05learn how to encode basic version of
1:32:07your strategy
1:32:08>> to make test it properly and then like
1:32:10understanding if what you're doing
1:32:12actually makes sense if you might end up
1:32:14with an uprising equity line then you
1:32:16might decide to only automate a portion
1:32:19of it or only to use it as confirmation
1:32:22and then still trade manually. So it's
1:32:24like it's really just an additional
1:32:27knowledge base
1:32:28>> validating.
1:32:28>> Exactly. Tool box that
1:32:32you can add to your arsenal.
1:32:34>> This is like one thing that I really
1:32:37thought that retail traders are missing
1:32:40and that now finally like um
1:32:43>> they can start piecing it together. in
1:32:45terms of uh as an institution and
1:32:48generally what retail can do if they
1:32:50choose to is you could have these four
1:32:52ideas but you you could have them all
1:32:54four running simultaneously
1:32:56>> 100%. Like that one you got me there.
1:32:58That one is the end game
1:33:00>> because the end game is not just to run
1:33:03one single trading strategy. The goal is
1:33:05to run uncorrelated trading strategies
1:33:08and that's what allows institutional
1:33:11traders to make money every month.
1:33:12Because you might have a strategy that
1:33:15works well in trending market, right?
1:33:18But if you are in a mean reverting
1:33:19market, the strategy will perform
1:33:21poorly.
1:33:22>> But if you have another strategy which
1:33:24is taking advantage of mean reverting
1:33:27conditions, then the other strategy
1:33:29Exactly. So like that's what
1:33:31uncorrelated means that you might have a
1:33:34set of strategy that works well in
1:33:37different market conditions and overall
1:33:40they give you that beautiful job is to
1:33:43really refine and learn over time how to
1:33:46manage them so that you limit draw downs
1:33:48on one maximize profits on the other and
1:33:50vice
1:33:50>> versa. Exactly like uh the exact process
1:33:53that I follow is like once a month I
1:33:56review which are the strategies that are
1:33:58running live. Are you able to give an
1:34:01insight in terms of you know when you
1:34:03were in your career how many systems
1:34:06would be operational at once or or you
1:34:08would be managing?
1:34:08>> Yeah. So like what I suggest as a retail
1:34:12trader to have at least three four
1:34:16strategies trading at the same time
1:34:17because in a case you are leveraging the
1:34:20power of your machine to have this
1:34:22trading strategies trading automatically
1:34:25for you. Yes. which is a massive
1:34:26advantage like if you try to follow four
1:34:29strategies at the same time it's very
1:34:31hard to do it manually and that's why
1:34:33like it's important like to eventually
1:34:35trade everything uh as automated trading
1:34:41>> while I was working for the bank I had
1:34:43something like 400 strategies running
1:34:46simultaneously but that way it's like
1:34:48>> you get to that point especially like
1:34:50once you are
1:34:53>> monitoring the performance across
1:34:55different instruments or you are
1:34:57monitoring the trading activity across
1:35:00different instruments. To put a context
1:35:02like the day I left the bank the total
1:35:05volume that was going through me was
1:35:07approximately 15 billion euro on a
1:35:10yearly base. So like you need to have
1:35:14multiple systems running at the same
1:35:17time. Um right now like personally I use
1:35:20a maximum of between 25 and 100 systems
1:35:25really depending on the type of market
1:35:27conditions that we are in.
1:35:29>> So would you do you find yourself
1:35:31consistently developing tweaking
1:35:33>> 100% like that one is the job the job is
1:35:36freely
1:35:37>> so changes from manually trading and
1:35:39having to like maybe do analysis to more
1:35:42an analyzing history analyzing edge you
1:35:46refining performance. Yeah, because
1:35:47obviously like trading strategy stops
1:35:49working. So you really need to have this
1:35:52organism that starts from research,
1:35:55>> refining, validating,
1:35:58>> deploying and back.
1:36:00>> Something we'll go deeper in tomorrow's
1:36:02words of wisdom for sure because I think
1:36:03that's the right place for it. But in
1:36:05terms of when we think about that and as
1:36:07we probably move into an era where more
1:36:11retail traders start to implement such
1:36:13systems, do you feel like manual trading
1:36:16will always have a place? uh or do you
1:36:18feel like as maybe more and more
1:36:20automation takes place that it's almost
1:36:24going to be necessary to have a
1:36:26portfolio of edges that you manage and
1:36:29uh the manual side becomes your research
1:36:32the manual side becomes your refinement
1:36:34your man your manual side becomes your
1:36:36review of your automated strategies or
1:36:38do you think the retail side in terms of
1:36:40manual discretionary trading will always
1:36:42be there
1:36:44>> I mean I think it will always be there
1:36:46but like the distribution will change
1:36:48because like once you understand that
1:36:51your computer can trade on your behalf
1:36:54and you can run and if you run multiple
1:36:58systems and with multiple again just
1:37:00three or four simultaneously can be
1:37:03enough to make your profits on a more
1:37:06consistent base like people will be like
1:37:09okay so you're telling me that I can
1:37:11have four strategies running live while
1:37:15I'm at work
1:37:17without the need of me sitting in front
1:37:19of the monitor to wait for a specific
1:37:22entry condition. People will move
1:37:24towards that direction because if
1:37:26freedom time give them more confidence,
1:37:28they have no emotions interfering with
1:37:31their trading activity.
1:37:34>> So it's going to be some sort of natural
1:37:36step
1:37:38>> moving towards that uh that direction.
1:37:40>> Definitely. I think obviously this first
1:37:42step is education like this, you know,
1:37:44being able to hear about the process,
1:37:46how to go about it, how to think. Uh
1:37:48excited to do a full master class over
1:37:50on Char Academy with you as well. Super
1:37:52fun.
1:37:53>> Um but yeah, no, this is this has been
1:37:55incredible. Is there is there more to go
1:37:56over or is
1:37:57>> No, I think we're done.
1:37:59>> It's incredible. I know something
1:38:00different for the audience there. As I
1:38:02said, we're going to sit down and do a
1:38:03Words of Wisdom, dive into your your
1:38:05background and more of that experience
1:38:07that you've had at the institutions and
1:38:09then where you're heading moving forward
1:38:10as well. So, if you're interested in
1:38:12that, that will be coming out very very
1:38:14soon. Keep an eye out. Uh, but for now,
1:38:16of course, links for Matteo will be in
1:38:18the description below. Make sure you
1:38:19check them out. Drop a like. Again, this
1:38:21is knowledge and and wisdom really that
1:38:24he doesn't have to share, but he's doing
1:38:25so because he wants to bring the new era
1:38:28of information and I I guess getting
1:38:30ahead, right? getting ahead of what's to
1:38:32come as you just said. Uh so make sure
1:38:34you drop a comment with your biggest
1:38:35takeaway from this episode. Any
1:38:36questions you have, throw them in the
1:38:38chat. I'll tell you why. Because when we
1:38:40do that master class, we can use some of
1:38:42those questions to really give you uh
1:38:44the answers that you're looking for,
1:38:46right? It'll be a perfect opportunity.
1:38:48And uh what other episodes are on screen
1:38:50right now? And until next time everyone,
1:38:52take

About this transcript

This page contains the full transcript of 4 Institutional Scalping Strategies Market Makers Keep SECRET (Automate Prop Firm Trading) by Chart Fanatics, generated from the public captions YouTube serves with the video. The transcript has 14,343 words across 2,275 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.