YT YouTube2Text

YouTube2Text → Transcripts → CODING TRADING AUTOMATIONS | July 24 | Michele Rossi

CODING TRADING AUTOMATIONS | July 24 | Michele Rossi — Transcript

by FunderPro · 5,360 words · 906 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Here
0:02we are again. Welcome back, beautiful
0:06people. If you can hear me or see me,
0:09leave a comment in the chat. Say hello
0:11and subscribe to the channel. Thank you
0:13so much, Fun Pro, for having me. This is
0:16the second lesson we're going to have
0:19about
0:21creating a strategy
0:24from the idea to the final
0:29functional
0:31strategy and bot to deploy on a prop
0:35firm. So in the last lesson we talked
0:39about everything about on how to create
0:43a strategy. We talked about
0:47technicals let's say I mean just we wrap
0:51around our head around the idea on why
0:55is it a good idea to code the strategy
0:58um and why
1:00is a good idea to use automation build
1:05your own strategy back test the strategy
1:08and then deploy it live on your prop
1:11firm uh today we're gonna
1:14have the first step and we're going to
1:16start to put our hands in the mud and
1:19start creating something. Today we start
1:22from the idea. Um, so we're going to go
1:27from the idea to the first working
1:30version today. So be careful, listen
1:34carefully, take some notes and if you
1:37want you will find this video recorded
1:41on YouTube on the YouTube channel of
1:43funprog.
2:02Trading.
2:18So, welcome to the first chapter of this
2:21week. Today, we're going to build the
2:23first working version of a trading
2:25strategy.
2:27um not just look at the chart and say
2:29this could work. We are going to start
2:32with a market idea. Turn that idea into
2:37precise rules and give those rule to
2:40codeex. We're going to use chat GPT you
2:42can use codeex you can use cloud code
2:46whatever it works best for you and we're
2:49going to use it to generate a pine
2:51script strategy. I know this is not the
2:56institutional way to do stuff but
3:01we're working on proper okay we're not
3:06competing against institutions that's
3:10not our job we are working into a
3:14virtual environment
3:17we're just looking at the real market
3:19we're not actually intervene
3:22in the real market. So, we're going to
3:25go deep on this um statement later on,
3:31but for now, so we're going to code the
3:35strategy with CHBT. We're going to have
3:38a Pine Script strategy. We're going to
3:40add it to Trading View. We're going to
3:43inspect the results and find the first
3:46problems and correct them. So by the end
3:49of this lesson, we will have a real
3:52baseline version of our strategy.
3:55It may be not profitable. It may not be
4:00robust. It will certainly not be ready
4:04to deploy yet, but it will be something
4:08that we can test, that we can measure
4:12and improve. And this is the first
4:15important step because you cannot
4:17properly validate an idea that has never
4:21been defined before. You cannot optimize
4:25rules that keep changing
4:28and you should never deploy a strategy
4:31that exists only as a bug concept in
4:36your head. [snorts] Over the next four
4:39lessons, we will follow the complete
4:42process. Today we move from the idea to
4:45the first functioning version. In the
4:48next lessons we will validate the
4:50strategy using trading view data. We
4:53will collect the complete list of trades
4:55and give that data to codeex. We will
4:58ask Codex to analyze the results and
5:00identify weaknesses, test different
5:03assumptions and help us investigate
5:06whether the strategy has a real edge or
5:08whether the historical results is simply
5:12the product of overfeeding.
5:16Again, we're not using
5:19all the indepth level three data. Of
5:24course, we're using a um an entrylevel
5:29data which is trading view. They have
5:33few tick datas, not even the full
5:35history, but it's more than enough for
5:38us to start
5:40understand the concept on how to code a
5:43strategy,
5:45back test the strategy automatically,
5:47um optimize it, validate it and deploy.
5:51And you are going to be surprised on how
5:56much this could potentially be really
6:00really profitable.
6:02Okay,
6:04we will also look at a Monte Carlo
6:06testing in sample and out of sample
6:09testing and cross validation. Finally,
6:13once we have a version that has survived
6:16the validation process, we will discuss
6:19deployment. At that point, we could
6:22convert the strategy into, if you wish,
6:24MQL5 for MetaTrader 5 or Ninja Trader
6:29with Python or C++ or we could leave the
6:32strategy on Trading View and use an
6:35external bridge to copy the signal to
6:37the account we choose. For this project,
6:40we are going to keep the strategy on
6:41Trading View. But first, we need to
6:43build it correctly.
6:46All right, let's start and let's begin
6:49with the idea. The strategy we are going
6:53to use is the orb is the most discussed
6:58and talked strategy of the last couple
7:01of years, but it's running around
7:07since the 80s, since the 70s,
7:121970s, 1980s.
7:14I don't know. But it's it exists from a
7:17long time. And you might think
7:21okay if it exists
7:24for this very long time maybe we got
7:28alpha decay the strategy is no more
7:30profitable. No that's not the case
7:33because we're not talking about a
7:36precise set of rule. We're talking about
7:40the overall concept which is the opening
7:44range breakout around that we can build
7:48multiple layers of filters not too much
7:51to avoid overfeitting
7:54to find a little small new edge that we
7:59can use if it exists to pass the
8:03challenges and to make payouts. That's
8:05the idea and I've already done that. so
8:09many times. My students have already
8:11done that so many times. So can you.
8:16So the OB stands for opening range
8:20breakout. The name sounds complicated
8:23but the basic idea is very simple. At
8:25the beginning of a trading session, we
8:28are going to
8:31define a range.
8:34Right? [snorts]
8:36For example, we can use the first 15
8:39minutes after the New York market opens.
8:44So, if you're in Italy, as I was for the
8:49last 40 years,
8:52the opening range is going to be between
8:553:30 to 3:45
9:00p.m.
9:04GMT + 2. Okay, this is the New York
9:08Stock Exchange opening first
9:1415 minutes,
9:16right? This is the opening range and
9:19you're going to have such a candle
9:24with a it's
9:28high low opening and close. Let's say
9:32it's like this of HCL.
9:39This is the low. This is the high. This
9:41is the opening. And this is the closing
9:44of the candle. Right at the end of these
9:49of the 15minut period, we have two
9:52important level.
9:55The high of the opening range and the
9:58low of the opening range. After that we
10:01wait. If price
10:05is going to
10:07closes above the opening range high, we
10:11look for a long trade.
10:16If price closes below
10:21the opening range, we're going to short
10:25[snorts] and sell
10:28the price.
10:30That is the basic ORB concept, the orb.
10:35But now we need to ask the question that
10:37should always come before writing code.
10:41Why should this idea work?
10:46Why would the first 15 minutes of a
10:49trading session be important?
10:52The reason is that the market does not
10:54behave exactly the same way throughout
10:57the day. When an important session
11:00opens, activity often increases.
11:05There may have been news during the
11:07night. There may be orders waiting to be
11:11executed. There may be a difference
11:14between the price accepted during the
11:16overnight trading and the price accepted
11:20when the main market opens.
11:23uh buyers and sellers begin to interact
11:26more aggressively.
11:30So the first few minutes can therefore
11:34create a kind of initial balance.
11:38The high range tell us where buyers
11:42stopped the market during that initial
11:45period. The low tells us where sellers
11:48stopped it. If price later breaks
11:52through one of those levels, the
11:55breakout may show that one side is
11:57becoming more aggressive.
12:01A close above the range may suggest that
12:04buyers are willing to continue paying
12:07higher prices.
12:09A close below the range may suggest the
12:12sellers are willing to continue
12:13accepting lower prices. That is
12:18that is the market hypothesis
12:22and I want to be precise about the word
12:24hypothesis. We are not saying that orb
12:28works. We are not saying that every
12:30breakout continues. We are not saying
12:33that the strategy
12:35has an edge simply because the idea
12:38sounds logical. We are saying that the
12:42behavioral
12:45the behavior is specific enough to
12:47investigate.
12:50That is the difference between an idea
12:52and the conclusion. An idea says maybe
12:56the opening range contains useful
12:59information about the next part of the
13:01session.
13:03A conclusion say the historical data
13:06shows that the behavior creates a
13:08repeatable advantage after realistic
13:11costs. We are not at the conclusion yet.
13:14We just started today. We are only
13:16building the system that will allow us
13:19to investigate the idea properly.
13:23Let me show you how this can look on a
13:26real chart. Okay,
13:29let me clean up the chart.
13:34>> [snorts]
13:36>> So for this example, I'm going to use uh
13:40the the a 15minute range. Let me go
13:43back. I I'm going to cherrypicking the
13:46idea. So it's going to be
13:49useful and clear for you.
13:53There it is. Let me go back. Let me find
13:58some good or example. This can be one.
14:01There it goes.
14:04[snorts] Beautiful.
14:06I know it's cherrypicked, but this is
14:08educational, okay? I'm not going to
14:10cherrypick anything else, but I need to
14:13show you how the process work. So, I'm
14:16going to use a 15inut opening range on
14:19the New York session. There it is. Is it
14:243:30? The market open [snorts]
14:27right here.
14:34So this is the first five minute candle.
14:37This one right here. 5 minute 5 minute 5
14:41minute. We have our 15 minute range.
14:48Right.
14:50We're gonna have the orb
14:53opening range high
14:57and we're gonna have we already have
15:01actually our open range low.
15:09The market and the exact ticker are
15:14important. If you use one trading view
15:17symbol and then you use another, we may
15:20not get the same prices, the same spread
15:22or the same session data. So I'm going
15:24to use the exact symbol visible here.
15:28Okay, I'm using the micro MQ micro E
15:33mini NASDAQ 100 index futures of CME.
15:39Right? So you can do the same.
15:42Um
15:44now
15:49the session begin here
15:54for the first 15 minutes. I'm not
15:56interested in taking any trades. One,
16:00two, three, 15 minutes. One, two, three,
16:03five, 10, 15. [snorts] I'm only
16:06measuring the market. During this
16:10period, the range
16:12is being created.
16:18This candle reaches the
16:22the high the highest price of the
16:24opening window. This candle reaches the
16:27lowest price. So at the end of the
16:29opening period, these two levels become
16:33our first range high and range low. From
16:37this point forward, the range should
16:39remain fixed. Now look at what happens
16:42next. Price move above the range high.
16:51There it goes. And we have a closure.
16:54There it is. Let me go back one candle.
16:58There it is. Our opening range high.
17:02Sorry, I already have my
17:05some of my bots installed here.
17:08So, this is the first trigger candle.
17:14That is the type of event we are going
17:16to code.
17:18The important detail is that the entry
17:21is based on the close,
17:24not simply on a week. So if we had a
17:28previous candle that wicked and closed
17:30back inside of the range, that is not an
17:32entry.
17:35This is one of the decisions that makes
17:38the strategy precise. If you we don't
17:42define this, someone else could
17:43implement an intraar entry and get
17:46completely different results. Now let's
17:50use this number. We're going to have our
17:53entry right here.
17:57We're going to have a our stop loss.
18:01Let's figure out we're going to put it
18:03on the opposite range. So on the or low
18:10and we're going to target a one to two
18:14risk to reward.
18:16We're risking one to target two. We're
18:20risking 1% to make it 2%.
18:28And we have to be very careful and
18:31precise on what we are going to say to
18:37define our strategy. So the rules at the
18:41moment are very simple. Wait for the
18:45first 15minut candle. Mark up the high
18:49and the low of the range and wait for a
18:51candle of five minute closing above or
18:55below. We have our candle closing above.
18:58Right at the close, we open a buy
19:04order. We're going to put our stop loss
19:07on the opposite range on the or low and
19:10we're going to target a one to two
19:12risk-to-reward.
19:14These are the rule of the strategy.
19:17Okay. [snorts]
19:20Um, of course, we also have to
19:26uh ask ourself some follow-up question.
19:31Are we going to enter on the same candle
19:33or on the next bar? Are we including
19:35commission? Are we including slippage?
19:37What happens if the breakout occurs five
19:40minutes before the session closes? What
19:42happens if price breaks both sides of
19:45the range? These are not secondary
19:48question. Those are very important
19:50question. We're going to ask these
19:51questions a little bit later. And this
19:54is also why most strategies found on the
19:57internet are not immediately useful.
20:00Often they are presented as a visual
20:03ideas rather than a complete system.
20:06Someone shows us a chart and say by the
20:08breakout. But which breakout? The first
20:12breakout of the day, the breakout after
20:14a pullback, a breakout confirmed by
20:17volume, a breakout that closes above the
20:21level, and what happens after the
20:24entitle, when is the stop-loss, when is
20:27the target, when do we stop trading? If
20:30these questions are unanswered, the
20:33strategy cannot be reproduced. Another
20:36problem is that we usually see only the
20:39best examples like we're doing right
20:41now, but for educational purposes. We
20:44see three winning trades on a shorts on
20:47Tik Tok or Instagram, but we do not see
20:50the losing trades.
20:53[snorts]
20:53We do not know whether the performance
20:56includes commissions and slippage and
21:00perhaps spreads. If you are on CFD
21:02especially,
21:04we do not know if the strategy is using
21:07future information and sometimes the
21:10strategy has been optimized so many
21:12times that it works only because it has
21:14been fitted to one particular historical
21:18period. A real edge is not a magical
21:21indicator
21:22uh is not a win rate is not a beautiful
21:24equity curve in edge is a repeatable
21:28statistical advantage. A strategy can
21:31win 40% of its trades and still be
21:34profitable if its winning trades are
21:38sufficiently larger than its losing
21:41trades. A strategy can win 80% of its
21:44trades and still lose money if the
21:46occasional losses are too large. So we
21:50need to study the complete distribution
21:53of results. We need to look at average
21:56trade, profit factor, draw down, losing
21:59streak, and performance across different
22:01periods. But before any of that, we need
22:04a strategy that is stable enough to
22:07test.
22:09Now, I'm going to write the technical
22:13specification.
22:15I'm doing this before asking Codex to
22:18write the code. We're doing together.
22:20Don't worry. This is important. Codeex
22:23can write pine script very quickly as
22:25well as
22:27uh cloud code or any other AI but it
22:31cannot read my mind. If I give it a v
22:34idea it must make assumptions and if
22:37those assumptions remain hidden I may
22:41end up testing rules that I never
22:43intended to create. So, I'm going to
22:46write the specification
22:48directly here so you can see and you can
22:51try and do the same. I don't like the
22:54red. Let me use blue. No, it's not
22:57visible. Let me use yellow. Let's go
23:00with the yellow.
23:03There it goes. So,
23:07the first part is the market. I will
23:09write we are using
23:13>> [snorts]
23:14>> uh the selected
23:18trading view symbol.
23:22Right? This is the first thing we have
23:24to say codeex. What symbol are we using
23:28at the moment?
23:31Right? MNQ. So we're going to use
23:35MNQ, right?
23:39But of course we can use the same we
23:42want to be able to use the strategy on
23:45different chart and test it on different
23:47scenario. Maybe you want to use it on YM
23:50which is the Dow Jones. You might want
23:52to use it on ES which is the S&P 500. So
23:55we are using
23:59the selected trading view symbol means
24:02that whatever symbol we are using the
24:05strategy whatever chart we are using the
24:07strategy on trading view is going to uh
24:13test the strategy on that symbol. The
24:15ticker
24:17must not
24:19be hard
24:23coded.
24:26This mean the script should work with
24:27the symbol displayed on the chart. This
24:29is simple. The second part is the
24:32session.
24:34The strategy
24:37uses the New York
24:41um
24:44session
24:45and the
24:49America
24:52New
24:54York time.
24:59Sorry, let me go
25:03time zone. Right. You can use the
25:09Italy time zone. You can use the India
25:13time zone. You can use Japan time zone.
25:17But we're going to use the New York
25:18session. We're going to trade in the New
25:20York session. I'm deliberately
25:24specifying the time zone. I do not want
25:27the strategy to use the time zone of my
25:29computer or whatever time zone happens
25:33to be selected on the chart. The third
25:35part is the opening range. This is also
25:38very important for [snorts] Metatrader,
25:41but we're going to talk in another time.
25:43The range begin at 9:30 a.m.
26:01New York time
26:04and ends
26:08[snorts]
26:08at 9:45
26:12a.m.
26:17During
26:20the those
26:2315 minutes,
26:25the strategy
26:28records
26:30the highest
26:33and the lowest
26:37low.
26:39So, the highest high
26:42and the lowest low.
26:46No trade
26:48is allowed
26:52while
26:54the range is forming.
27:00Once the range is complete,
27:06the levels
27:09are fixed.
27:15Okay. Now the fourth part is the entry,
27:20right? We need to tell when to enter
27:22long and when to enter short. So,
27:28a long signal
27:31occurs
27:34only when a confirmed candle
27:41[snorts]
27:42closes
27:47above
27:50the range high.
27:54While for a short
27:57[snorts]
28:01signal, a short signal occur only when a
28:05confirmed candle closes below the range
28:10low. Right? [snorts]
28:14only one entry
28:23um is allowed per
28:27session.
28:32If the first trade, let me dictate. So,
28:34I'm going to go faster. If the first
28:36trade is long,
28:40we do not later open a short trade on
28:43the same day.
28:45Oh, it's not working. Let me write it
28:47down. If the first
28:53trade is long,
28:57[snorts]
28:59we
29:01do not later open a short trade
29:09on the same day.
29:17Same apply
29:20for the short. If the first trade is
29:24short, we do not later open a long
29:27trade [snorts] on the same day just to
29:30be sure
29:32we're hardcoding every property,
29:37every rule, every idea,
29:41everything that can mess up with the
29:43with the logic of the strategy.
29:47So the sixth part is the stop for a long
29:51trade.
29:57Let me see if we can do this right. We
30:00can do this. I can dictate now. Okay.
30:04For a long trade, the stop goes at the
30:07opening range low. It works.
30:12For a short trade, the stop goes at the
30:16opening range high.
30:18Right?
30:20Then let me stop the right. Then the
30:24seventh part is the target.
30:28So what are we going to say
30:31is
30:33we use a configurable risk-to-reward
30:37ratio with a default value of two.
30:41Perfect. Wisher pearl is crazy good.
30:47I need to remember to stop and restart.
30:52Then the target must be calculated from
30:57the actual entry price and the initial
31:00stop distance.
31:02Right?
31:05Then the eighth part is the session
31:07exit.
31:11Any open position is closed at the end
31:14of the session. There is no overnight
31:17holding
31:20this.
31:22Let me go a little bit smaller.
31:29And the ninth part is position size. For
31:33the first version, we will use a
31:35configurable quantity. We are not trying
31:37to solve the entire portfolio risk
31:39management problem here.
31:47There it goes.
31:50[snorts]
31:51Um, so
31:53we are trying to create a clean
31:55baseline.
31:57The final part is the visual behavior.
32:00So we need to tell the script to show
32:03because we want to see we we can we can
32:05do it. So we we can we want to also be
32:08able to see the script must show the
32:11range high and low. It must shade the
32:14opening range area. It must mark
32:17entries. It must show the active stop
32:21and target. That way we can compare what
32:24the code does with what the rule say it
32:28should do.
32:34Come on. Whisper plot, please don't
32:37abandon me. There it goes. So, this is
32:41our complete first specification. It is
32:44simple but is precise. Now, I'm going to
32:47turn that specification into the prompt
32:50I will use with codeex. I want to show
32:53you the exact process because the prompt
32:55is part of the development workflow. I'm
32:58not going to describe the prompt
33:00afterwards. I'm going to write it, sand
33:03it, and use the results. Okay,
33:07so I already have this prepared. I'm
33:10going to copy and paste it inside of
33:15Codeex.
33:17So, let me open up Codeex.
33:26There it is.
33:34There it goes.
33:37Right.
33:39I'm going to use 5.6 salt or whatever
33:43you have. I'm going to paste this whole
33:45code. Let me write it to you. Let me
33:48read it to you guys. Okay. So, it's very
33:53simple and you can copy paste. You are
33:56an experienced Pine script developer and
34:00systematic trading researcher. Create a
34:03trading view strategy in Pine version
34:05six based on the following
34:07specification.
34:09And these are the specification.
34:12Use the trading view symbol selected by
34:14the user. Do not hardcode a ticker. Do
34:16use American New York at the session
34:18time zone. Do not use the computer's
34:20local time zone or the charts time zone.
34:23The trading session begins at 9:30 New
34:25York time. The opening range lasts for
34:2715 minutes and end 9:45 New York time.
34:30During the opening range, record the
34:32highest high. Record the highest low. Do
34:34not open any trade. After the opening
34:36range is complete, enter long only after
34:39a confirmed candle closes above the
34:42opening range high. Enter short only
34:44after a confirmed candle closes below
34:46the opening range low. Allow only one
34:49entry per session. Do not reverse the
34:52position during the same session. For a
34:55long trade, place the stop- loss at the
34:56opening range low. Calculate the profit
34:58target using a configurable
35:00risk-to-reward ratio. For a short trade,
35:03place the stop- loss at opening range
35:05high. Calculate the profit target using
35:08the same configurable risk-to-reward
35:10ratio. Use a default risk-to-reward
35:12ratio of 2.0. Close any open position at
35:16the session end. Do not hold positions
35:18overnight. Create inputs for sessions
35:20serve time, sessions and time, opening
35:22range duration, riskto-reward ratio,
35:24position size, commission, sleepage,
35:28visual requirement. Plot the opening
35:31range high and low, shady opening range
35:33area. Plot the active stop-loss and
35:35target. Mark long and short entries.
35:38Include an optional debug mode.
35:40Technical requirements. Use confirm bar
35:43data. Avoid look ahead bias. Reset all
35:47daily and session variables correctly.
35:50Make the code readable. Add commands
35:53explaining the session logic and trade
35:56logic. Include alerts for long entries,
35:59short entries, stop losses, target, and
36:02sessions close.
36:05Before providing the coin the code,
36:08explain the main assumption and
36:11execution model. then provide the
36:13complete fine script.
36:16Right? Here we go. So this is it.
36:23Is that everything clear?
36:26Do you guys have any questions about it?
36:30Please let me know in the chat.
36:39Beautiful.
36:41Beautiful,
36:52right? Okay, cool. Let's send it to
36:55codeex
36:57and wait.
37:03So while the code is being generated
37:08by codeex notice what we are doing we
37:13are not asking codeex to find a
37:15profitable strategy. We are not asking
37:18it to optimize the parameters. We are
37:20not asking it to add indicators until
37:23the equity curve looks good. We are
37:26asking it to implement a clearly defined
37:31hypothesis.
37:33That distinction is extremely important.
37:37If we ask codeex to make the strategy
37:40profitable, it may change the rules and
37:44filters or optimize parameters before we
37:48even understand the baseline.
37:51So for now, we need a faithful
37:54implementation,
37:55right? We need to wait. It's going to
37:57take
37:59a little bit of time. So, if you have
38:02questions, this is a very good moment to
38:04ask
38:10or we're just going to wait.
38:15[snorts]
38:27assumptions and execution models.
38:31Right here it comes.
38:34Coded codex is writing the code
38:40based on my hypothesis.
38:50It's going to be ready very soon and
38:53we're going to test it in the chart.
39:00If you want the full code,
39:03I mean the prompt to generate the code,
39:08just follow me on Instagram official and
39:11ask or you follow and subscribe to my
39:14YouTube channel and you command by
39:18asking.
39:22that
39:34there it is. It's ready. It didn't take
39:37more than five minutes. Right now
39:41before copying it into trading view I'm
39:46going to ask codeex to explain the most
39:50important parts.
39:53So
39:55I'm going to ask this very prompt.
40:00I'm going to ask this.
40:03Review the script you just created. Do
40:06not rewrite the code yet. Explain how
40:09the session is detected, how the opening
40:11range begins and ends, how the high and
40:14lows are stored, etc., etc., etc., etc.
40:17Identify any possible bug or ambiguity.
40:21Right? This review is very useful
40:27because it forces us to understand the
40:30code before trusting it. So if codeex
40:35says that the range is being calculated
40:37using future bars we know there is a
40:40problem. If it says the that entries are
40:44generated
40:46intraar that does not match our
40:48specification. If it says that the
40:51strategy can reopen a trade after the
40:53first one closes we know the sessions
40:55lock is incomplete. The explanation is
40:59not a substitute for testing, but it
41:02help us know what to look for on the
41:06chart.
41:07Right? Let's see.
41:11As you can see, is automatically fixing,
41:15reviewing possible bugs and ambiguities,
41:18possible duplicate exit alerts, and it's
41:21going to fix the code by itself. So just
41:25be aware that we might have a partially
41:30working code. It might not be completed.
41:34Uh
41:37so if it's not working on the first
41:39trial, you just keep on asking codeex to
41:43fix because trading view is going to
41:44give you what's not working and you got
41:48to go back and forth maybe a few time.
41:50I'm not going to do this right now.
41:53Okay,
41:55now we're asking now write
41:59the final fixed code
42:03ready to copy and paste
42:08into trading view.
42:11So I'm going to fix it later if it's not
42:13working right now. I'm going to just
42:15show you a version that is already done
42:19by me.
42:20It's very similar but with more complex
42:23rules, more advanced rules just to show
42:27you guys. [snorts] Let's see.
43:25Here
43:29he comes.
43:41We're going to create a new strategy
43:48and we're going to copy and paste the
43:51code inside of Trading View.
43:54We're going to save with the name save
43:58script
44:00New York opening range breakout funer
44:03pro.
44:05Save. And we're going to add it to the
44:07shot.
44:09Now, as it's um
44:14compiling, it might find
44:18some errors. Like, as you can see, it
44:20just found an error. We're going to go
44:22back to Codeex, give it the error and
44:26ask it to fix it and give us back the
44:29full code. Now, for the purposes of this
44:31live, we're going to have the full code.
44:33I'm going to show it to you the next
44:34time. I don't want to waste your time
44:37right now. Okay, write the full final
44:42code. Fix it. Let me see if it can make
44:45it within the end. But you're going to
44:48find something that look like this.
44:52This is one of the the orb strategy that
44:55I coded and it's going to detect the
45:00exact range is going to place your
45:02stop-loss, your entry point and your
45:04takerit and most importantly
45:08it's going to give you some
45:12reports such as key stats like total
45:16profit and loss into the last April
45:19July. So it's three month risking 1% per
45:22trade. It makes 3.51
45:25with a 3.28 draw down. Uh 58%
45:31profitable trades. Profit factor 1.2 and
45:34a lot of um a lot of
45:39key metrics.
45:41Okay.
45:43And we're going to talk about all of
45:44these in detail in the next sessions as
45:47well as every trades taken taking in
45:52consideration
45:54the best
45:57the best uh set of
46:02uh historical
46:04um data which is more or less 40 ticks
46:08per bar with high definition.
46:12uh is going to
46:14execute the the script on the bar
46:16clause. We added some I added some
46:19commission 0.525 per contract. I added
46:23two ticks of of slippage. I added one
46:26tick of extra order execution delay with
46:30it's very similar to what happens in
46:32reality as well as inputs area where you
46:35can put all your data. So you're gonna
46:38find yourself into
46:41something that look like this. And you
46:43can go back in time and back test more.
46:45It's going to take more time. After
46:47that, we're going to do uh some
46:50optimization,
46:52some validation, and then deploy it
46:55live. Let me see if it picks up the code
46:57in the meantime. If not, no worries.
46:59We're going to do it offline.
47:02[snorts]
47:04No variables alternative
47:07on character use the block copy button.
47:10Let me let me try again.
47:13Let me try again. But it's just this
47:16first
47:20should work.
47:25Just added few more.
47:29Here it is. Our opening range script is
47:32now ready. As you can see it it okay we
47:37have a problem with impagination.
47:41Okay. This should move this should stay
47:44fixed into the chart but no problem we
47:46can fix that very easily. We have our
47:48session start time. We have our session
47:50end time. We have our opening
47:53uh range duration risk to reward
47:55position sites in contracts and the bug
47:57mode to pigs everything else. We can use
48:01we can choose the sides we can use this
48:03we can use this
48:06for example
48:09commission 0.25 25 per contract
48:13and one tick more execution delay. One
48:17tick, [snorts] one tick of sleepage.
48:20There it goes. Let's see. And you're
48:22going to have your report.
48:25Okay, this this is not looking good, but
48:28we're going to work on that with
48:30optimization. Okay, you can go even for
48:34355 and it's going to do exactly it's
48:38going to tell you exactly what to do and
48:40when to do it.
48:44Okay. Uh so we're going to fix it. No
48:47worries. We're going to fix it and it's
48:49going to look much better something like
48:51this. But you just saw how starting from
48:56an idea and hypothesis we define the orb
49:01using the XA rules.
49:04Right? Then we created a detailed prompt
49:08for codeex that you can use for any
49:10other hypothesis you might have. We used
49:13codeex to generate the pine script. We
49:16copied the script into trading view. We
49:18inspected the trading range, the
49:20entries, the stop-loss, the target and
49:22the session behavior. And we also looked
49:25how to debug a problem without asking
49:29codeex to rewrite everything from
49:32scratch. Just giving back the errors or
49:34what we're seeing right now. I'm going
49:35to say, okay, the range stays stuck in
49:38the chat is not moving with the with the
49:41candles. Please fix it. We can also
49:44provide some screenshot. Okay. And
49:49this is the process that we will
49:53continue throughout this project. The
49:56first version is not supposed to answer
50:00every question. It is supposed to create
50:03a reliable starting point. Now we have a
50:07baseline that can generate a list of
50:10trades. And that list of trade is what
50:13we need for the next lesson because the
50:17next question is no longer can we hold
50:22this idea. The next question is does
50:25this idea still make sense when we
50:28analyze the evidence.
50:31That is where real research begins.
50:36Thank you guys for being here with me
50:37today. Uh you can find the full video
50:41into
50:44Thunder Pro YouTube channel and on which
50:48is my channel. You can find the name
50:50right there. And you're going to find it
50:52also in
51:14Challenge instant futures
51:33trading
51:35quanti.
51:40Thank you so much. Have a great day and
51:42see you soon. Bye.

About this transcript

This page contains the full transcript of CODING TRADING AUTOMATIONS | July 24 | Michele Rossi by FunderPro, generated from the public captions YouTube serves with the video. The transcript has 5,360 words across 906 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.