Visit our channel for more free Pinescript Strategies. Timecodes 0:00 - Intro 1:45 - Create algorithmic trading strategy chatgpt 2:35 - Backtesting moving average strategy 3:30 - Generating Code for backtesting in chatgpt 4:27 - Download stock data code in python 5:25 - Python code for yfinance chatgpt 5:13 - Forward Testing 7:00 - Optimization 9:10 - Course Strategy results via Optimization 11:24 - Machine Learning Optimization Code in Python in Chatgpt 12:35 - Parameter Optimization 13:45 - Moving Average parameter Optimization 15:35 - API for brokers python 16:45 - Trading strategy in cloud chatgpt 17:30 - Pinescript code for moving average 18:40 - Trying the trading strategy generated by Chatgpt 21:00 - Asking chatgpt to create a 3000% strategy 22:00 - Creating a momentum long only strategy 23:00 - Backtesting th estrategy generated by chatgpt 23:40 - Tweaking strategy to get 20097% returns 28:15 - Finding intrinsic value of company using chatgpt Dont take the strategy as financial advice. The whole video is just about how to get information and codes regarding financial trading and algorithmic trading from chatgpt. Further research has to be done before going live in any strategy. Thanks much. Let me know if you guy shave any doubts or clarification. Happy to help
@@just_another_smoker8687 Not really. The strategy is backtested from inception of the stock. But multiple trades are done. However the results doesn't mean anything like i said in the video. We need to do other quant analysis like portfolio backtest, optimization, forward testing and monte carlo before we even think of going live. The whole idea of this video is to help traders use chatgpt to get information and codes to backtest efficient strategies.
Alot of companies going to be taking hits! Also you need to consider that majority of the population are not smart enough to use to ChatGPT to its full capabilities. But, for those who do know how to use ChatGPT, it's going to total domination in their field
Something to keep in mind when using CGPT is it's most effective when you converse and collaborate with it. Within the thread, it remembers everything you and it have already said, so you can refer back to those things to ask for refinements and set expectations for future answers. What most people don't get is that everything you might say to another person, or in the case of the video audience, you can say to CGPT to help it provide better answers. Instead of saying to the audience, "it didn't say why it used something," or "it didn't use the right thing," say that to CGPT: "Why did you use that?" or "rewrite this using x strategy or technology." An example of the expectations for future answers, you can tell it: "for all future code that uses pinescript in this thread, include the version comment at the top, set to version 4, and use 4 spaces for each indentation." What's revolutionary about this technology is that if it gives you something that isn't what you are looking for, you can describe to it what is wrong with it or what would be more right about it, and it will understand that with natural language. The more you converse and refine the results, the better they will be. Great content and thanks for the video! Hope this comment is helpful to whoever reads it.
Wow this was really insightful! I think a major misconception with this tech is that it is like a Q & A forum with the AI, not an open dialogue with the AI
I don't use it for Trading, but I do use it for coding and absolutely conversing with the AI is amazing. Initially it often provides wrong or bad code but with a bit of back and forth dialog it can produce some fantastic results.
@@jadebreed3333 Yes, true that once you disconnect it forgets. That sucks, but during a coding session it does pay off to converse ideas as it will alter it's outcome better.
Investing in the stock--market and crypto is the best option to make a passive income. Virtually all the markets are crazy, most people pay more attention to the shiniest position on the graph, I'm keeping a diversified portfolio. !!!
@@Stephenfanos37 That's impressive!!! I could use the expertise of this advisor, my portfolio has been stagnant..... How can I get started with PATRICIA DRITA JOSEFINA ??
1. You should look into prompt engineering to help you modify the responses you get from the AI based on how you word your prompts. It helped me immensely and i believe it will increase the quality of the results you get. 2. Try to fact check some of the information because the model can sound factual about non-factual subjects sometimes(OpenAI indicates this in their documentation). Great content.
@@quantprogram Yeah, you can ask to explain CGPT's answers in detail, and you don't need to refrain your question every time. It'll follow and "remember" the conversation down to 1st question.
@@Qban220Thanks Jack. Will attempt it that way as well. I did notice this sometimes i ask the same question but its in a different chat, the wordings are quite different. I wont be surprised if chatgpt is improving on its information as more people asks questions
I tried making a simple trend indicator and telling cgpt all the errors i was getting, he kept modifying the script till it worked. This thing is amazing. 🤖 edit: i used "he" instead of "it" lol 😁
The easiest way to succeed in any field is to work on it consistantly, and consistancy is the key to success. I experienced my self. By the way, thanks for the uploads , a fellow creator.
Yes I trade one stock and have the chart up all day everyday. Ii think my subconsious has picked things up because I see pattern recognition am making money almost everyday.
I am blown away by the versatility of AI . It's amazing to see how this technology can analyze market trends and make informed decisions in real-time. I can't wait to see what other innovative ways AI will be used in the future. Great job on the video!
It's kind of an unusual strategy that it used for generating 20000% profit. If the current price is higher than the price 50 days ago, it's a buy. This algo could be modified to track momentum over the past few time periods to see if momentum is increasing. The takeaway I have from this is that ChatGPT can help someone start learning something new and as they apply the knowledge, they can learn further on their own and still use ChatGPT to further their learning as they refine their requests and start new conversations.
This thing has walked me through learning python and Keras, it also walked me through changing an engine manifold with specific information about the vehicle. Don't expect it to do brain surgery and understanding how it works helps you to engineer the prompts better.
In the future it will help with brain surgery and much more. It’s still in its infancy, think of it like a 2 month old baby that can teach you quantum physics.
I tried using chatgpt for simple web dev. If you ask for something straightforward it works, but the more you ask for modifications and adding different technologies, it creates bugs because I don't think it really knows how to combine them. So you can use as a template but still need to review code and verify it makes sense.
It’s best to generate code in chunks and not let it generate an entire app or file. The bigger the prompt gets, the faster it’ll forget context due to the restriction of tokens (8000 I think)
One major oversight was made that changed everything. With ChatGPT you’re having an active conversation. Once you ran a test based on the 50 day sma everything else was based upon that…. You need to clear the chat to try a new strategy, otherwise you’re still talking about the same thing, ie the 50 day ma. Without clearing the chat you’re always working off of the beginning of the conversation. There are various other ways to improve upon the output as well, but I think that’s enough free Alpha for today. 🤙🏼
@@quantprogram By now I’m guessing you’re a lot more impressed with the program, it’s crazy. The next model will be much easier to teach new things to 🤯
Absolutely mate. I understand the learning and development advantage the user has. But I'm quite wondering how its going to be in the personal space. Maybe we will reach a point that we don't know if the person we are talking to is AI or human. It also amuses me that its free. As if they are trying to collect as much as data from us as possible so maybe in future they can mimic us? Reminds me of movies like Ex machina or Her
Part of my fund is creating ways for investors to have a personal connections and opportunities to make choices on their own. Nobody is ready for a black box / chat bot investment world. So, I agree with the personal experience for people.
@QuantProgram I guess it’s free because it’s still learning and now learning very rapidly as more than 1m users are taking it for a spin. Win-win because it gets better the more it learns.
Excellent material although as someone reminded me back testing does not indicate the strategy would work today as the observable data does not factor inputs of real time trading which would change the outcome on the system.
3:08 - you could communicate with it almost as a human, so you can refer the previous parts of the conversation and ask it to give it in code. Example: " Now show/give it in code." or "Show me [insert a small portion of a previous prompt, as a reference] in code", etc. The second example can be used and its variations for prompts and answers that are earlier in the conversation and are not the immediate last one. Hope this helps. P.S. You can also pre-define and tell it to give you way more detail information and even long lists of things, but you should add clarification that "you are limited in the length of your answers so after each answer I will say "Then what?" and you will continue from where you stopped until you answer completely" or something in along these lines, letting it know you know it is limited in its answers and that you will use a specific command to let it know to continue.
I appreciate your approach to teaching.. To my understanding this just proves how much we need an edge as investors because playing the market like everyone else just isn’t good enough, we just need to hold onto our hopes and wait to see how things turn out because market movements are almost always unpredictable. In my portfolio, I'm noticing more red than green.
I tell you how to do it and save your time. ADX moves up above 40 level (14:5 settings) and stochastic (9/5/5 setting) goes up above 50-level (buy) goes down below 50-level (sell). Works on any time frame. Thank me later. No bots needed.
@@jaybee6347 Like said BUY when ADX indicator slopes up above 40 level + MACD slopes up above the zero level + Stochastic slopes up over its 50 mid-line level. Just cannot go wrong with this and the opposite applies for SHORT when the ADX slopes strongly upwards but all other indicators slope downward. Take profit 1:5 to 2 and SL goes below the previous price closing low/high. Trading is a lot about waiting for the correct set-up and when it happens you go in heavy. Avoid mediocre set-ups and just buy/sell the premium ones. I only trade intraday with 30-minute time frame, this way when I go in on a trade it runs longer that what it would be do a 1M or 5M time frame.
@@jaybee6347 Cool, let me know how you find it. As per writing this mornign here in Europe, the US30, US100 and US500 were dipping with these exact values and ADX was up-sloping, Again I was able to hold short on all the indecies for good 45-minutes before a short correctiong is been seeing and I am out of the positions until I see the indicators aligning again to a mutual direction. Let me know how your back test goes with intraday values and time frames ether its from 1 minute, 5, 15 or higher. Happy to hear the results.
I don't see any edge being provided. Writing Pine script is very easy for anyone who knows how to program. ChatGPT just saves you the time and money of hiring a developer or learning to code? I have 2 years of Pine experience could easily code anything you did in this video in half an hour or less
Oh boy! I tweaked it further and saw a return of 5,623,187.13% !!!! for the same AMZN daily chart and time period (Sept'1997 to Jan'2023!! thank for being the torch bearer. I am still pinching myself and can't stop playing with it, even tried the short side as well. I know it is difficult to believe, but if 20,000% plus result was unheard of, I know what folks will talk about this number.... itching to implement it live with $1000 only to start!! ....will come back in few years from today. Never ever commented before, but this result was to much to ignore and not share. Thanks a ton!!!!!!!!!!!!!
Please keep in mind that's amazon during a massive bull market. Everyone's a genius in a bull market. Even bear markets can make a lot of money, what about when markets range? If something is very simple and "too good to be true" it more often than not is
@@heeroyuy298 Yes, after 40yrs of playing with data/databases and 23yrs of trading, I finally found the holy grail. I am not using any of the indicators in this or any other tutorials/youtube/chatGPT etc., but created my own indicator which gives even more astonishing results live on ThinkOrSwim (TdAmeritrade platform) with real money. The moment I make couple of millions with it, I will launch a trading portal for public use with a server farm in the cloud to take care of any load. At the current rate of returns, expect to launch it around July'2023 !! It could be a gamechanger for many, as it is turning out for.
i tried chatgpt the same day it was available for metatrader code :) you need to specify of what you want exactly, lets say after a code created , you can tell him add to this code moving average and it does what you tell him, sometimes the code is too long, and you have to reenter last 2 rows of code or full script he wrote to continue
Sweet baby jesus, its the indentation.. Bro. I've been trying to get chatGPT to write me some scripts but not knowing pinescript all that well, it errored out for no apparent reason so in the end I just abandoned the whole 'write me a script' idea. I have some sick ideas. Thank you!!
Happy it was helpful. Watch the "Tradingview Pinescript version 5 Full tutorial" video on our channel if you wish to master the coding bit in pinescript. Thanks for watching and for the comment mate. Appreciate it
Those of us in the trading world know you could build such devices with a super computer and still lose your shirt. Unfortunately some people are too lazy to "THINK", and look for every conceivable machinery to make trading easier.
I believe that this recession has taught individuals the value of having several sources of income. Unfortunately, having a job does not guarantee security; rather, the real security comes from diversifying your investments. When it comes to speculating in the financial market, understanding institutional investment is essential. Whether it's real estate or cryptocurrency, investing in knowledge will yield the highest returns. Since I started trading with Mr. Thierry Veilleux, whose skill sets are extraordinary, I have seen enormous profits on my investment. Working hard is good, but working wisely is preferable. Trade with a professional! It really is simple.
i was a very good algo trader, but left this field bcoz i was not able to make enough profit. now looks like i need to revisit my skills. and start again...
Great content. Subbed. More, please. I'm interested in learning a language for trader station and want to work out the best coding to learn in order to notice when chatGPT gets it wrong or has missed something. Looking forward to seeing how a newbie can spot and handle the bugs in the chat code.
Big question, where does it get information. I don't think even Google AI could have been smart without support and consent of original content creators. You can't self learn by yourself. Literally it has many flaws especially regarding copyright laws.
ChatGPT is the best tool so far for research, do homework, create stuff by order and to provide any info that is needed. AI is the best thing after Blockchain technology in my opinion.
Blockchain is still looking for a problem to solve whereas ChatGPT3 is a serious step towards KI and will definitely be very useful. The technology behind blockchain is essentially cryptography and dates back to the 70ties, so really old stuff.
You're not talking to the bot ......... you're trying to "query" stuff the same way you do with Google. That's not the way this is designed to work. It's called "Chat" ... it's a language model you need to "talk" with it in human language not just through keywords at it like we do with Google
What we really need is a system that accepts ideas, then does backtesting (multiple periods) and Monte Carlo simulations in bulk volume, without all the manual work. This would "solve the market." Later, a system to generate testable ideas... and then what is left to automate?
This already exists tbh, I found it not too long ago made by a solo developer haha. Just trying to figure out my Python code now for applications but I can steer you to it as well bro
Thank you for the video. It's very interesting. However, referring to the actual Trading strategy that produced 20 thousand % need to be backtested with what would happen if you just buy at the begining of the period and just sit and look how your investment growth. I tried it in TV and got pretty similar results, a profit of 19.620. However in my list of transactions, the first one is a buy at 0,12 cents. Taking into account that the current price in your example is around 86 per share you've got 72k% return just watching TV (television not Trading View) . So the massive return is for the Amazon stock.
Interesting but you can get nearly infinite returns on past data. The rest test is applying this strategy for a few weeks to see what the results are. I wish you would have included the results of that in the video!
Absolutely and i wont recommend at all to use this strategy, i didn't recommend it in the video either . I assume you dint watch the video fully. I have discussed in the video about optimization and forward testing. No strategy is bullet proof until you go through portfolio backtest with position sizing, optimization, forward testing and Monte Carlo Simulation. This strategy has not been through that. The whole idea of the video is how we can use chatgpt to help us with finding information and codes for creating strategies. Nothing else. We have to do our bit by doing more research into the strategy by using these quant methods.
I noticed that my backtesting of a week back is more reliable than backtesting all time Coukd this mean we should regularly backtest strategy settings weekly?
This has so much potential it's crazy. I feel if this is perfected and becomes wide spread stock exchanges would ban the use of automated bot trading. How'd they go about is hard to say but personally I feel it is a possibility. Amazing video btw!
Thanks for watching mate. I'm more worried about how this can affect our life personally. Its free that makes me super suspicious. Is it collecting our data so much to use it against us?
@@cellc6191 Well my theory is by us seeking basic knowledge, they get to know what we want to know. If they get to know what we want to know they can control or limit that information or manipulate the information. Manipulation can go deep, just like people find beauty standards these days in social media. It can also destroy jobs. As in the video i tried to get some codes, even though its error prone and not perfect it leads to less people being hired in certain departments of coding industry. Or marketing industry for creating good images. I mean this i just what i think. but i dunno what the future holds. Could be worse or could be not that bad
Here's an interesting idea, try to teach it a few times before actually asking for a strategy. Like priming it before shooting it. Feed it data to base off of
Hi Christer, yes you are correct, it works much better that way. I created the video when chatgpt just came out and wasn't aware of the different approaches in prompts to get better results. There was very few info available online when it got released either. But now since chatgpt is taken off there are better info about ideal prompts including the advice you gave to get better results. Thanks for the comment,
At around 3:00 I tried the exact statement, but instead of Python, I put in PHP. It ended up giving me the code in PHP. I have my own web host, so I can try it out on my web server.
Another flaw I noticed in your conversation, is that it's not a conversation. You aren't building on things. You point out a flaw, but you don't talk to ChatGPT about it. It can actually remember what it was talking about... It feels like you're using it like DALLE2/Stable Diffusion. "I know, let's throw python at the end of the sentence!"
20.000% revenue pretty decent? Am I missing something? Which period did it backtest on? Seems quite unbelievable especially considering it's such simple strategy.
I think AI trading is going to be big. And AI companies therefore insanely rich if they use their AI for trading. Not sure the advancements will come from llms though.
It won't IMO, maybe for a short period. If everyone is going to use AI for trading, no one will succeed. It will just make market behave even weirder. To succeed in trading (as opposed to long-term investing) you have to do something differently, have an edge (or have insider info) to screw your fellow traders up. There is no magic method in trading and no concrete rules for success. AI generated text content will spam already spammy internet to further reduce the amount of quality content to learn from. Similarly, AI based trading will make markets unstable and always changing, making it difficult to train a new generation from. I am not very happy from all of these aI developments... but we'll see
Mahatma Gandhi said: “A small body of determined spirits fired by an unquenchable faith in their mission can alter the course of history.” It seems like the scale for being literate just got tipped, so now using AI is going to be the norm, just like using google once tipped the scale and now most literate people know how to use it, let see how much we have to pay for it. (soon) lol
Thanks for this great content! I just subscribed as I'm new to quantitative trading field. I'd like to ask : to which extent an average quantitative trader can make profit ? I mean, I'm not new to trading, I know the 90% of traders lose money rule etc. But I'm asking myself, is a decent return (for example 15% a year) reachable to anyone willing to put the efforts in ? Also, I was asking myself, is there any other languages useful to learn apart from python and pine script, the latter being similar to the former ? + Last question but not least : Have you been able to trade with your own quantitative trading strategy ? If so, what were your results if you're willing to share them ?
I think personally from my experience, I always look to beat the market in the long run. By market i mean SPY buy and hold CAGR returns. That roughly comes around 10%. So my goal is to beat that, not year on year but on average on the long term. My portfolio of trading strategy is not purely quantitative strategies alone. It has got value based investing and also a few day trading strategies. This year since the market has been down I was able to purchase lots of good companies at bargain prices. From the quant side, the mean reverting strategies has also performed well. However from a day trading perspective it has been a huge struggle. The advantage of quantitative trading is that you know what your'e gonna get. The forward testing and Monte Carlo Simulation helps you prepare for the worst case scenario regardless and from my experience it has gone accordingly to the expected results. The cons of quantitative strategy is that you need at-least 5 strategies at the same time and also the belief to stick to it. I remember earlier this year the strategy that i have explained in 09:10 didn't work that well, there was a few consecutive losses and you feel like you want to shut it down. The temptation exist but this results was expected and eventually it started making spectacular returns after couple of months. I also dont use leverage in neither my quant based or value based strategies which think have benefited me a lot unlike the other traders. The other con of quant based trading is there is lots of room for improvements. Like now I'm learning lots of machine learning models to predict volatility contraction and expansion so i can apply in option prices to benefit the theta decay in far OTM options. So there is a constant need and urge to learn.
Thats exactly what i said in the video. I assume you didnt watch it fully. The whole idea of chat gpt is to get knowledge and information especially coding knowledge, not to predict anything
@@quantprogram also… people think they can sell or use the data produced for commercial purposes… This is Napster only for data. I see many people spending the next 5 years in court losing everything.
@@quantprogram even worse… now that I’ve tested it and seen what other people think they can do with it… this is a lawsuit waiting to happen. This gathers copy written material and presents it in its own chat form, without reference to the originating author. Then people are suing this data as if it was their own. Which is not only a federal criminal fraud, but also a civil liability action. Weird thing… chatGPT doesn’t understand this.
Do you have videos that show how to incorporate human input into the tradingview bot? For example, you want the bot to follow a certain protocol according to trendlines or levels you draw on the chart, or other manually inserted variables.
Exactly! Specific protocol tailoered. For me, I want it to hedge from the get-go but not like most of those trash-bots ppl make. More like a real trader w/ slower solid action for bigger returns. Maybe if A.I could watch tons of the traders style for each specific strategy so it could machine-learn (mimic) it to a T
very interesting man! in the parameters of the strategy settings the trading commissions should also be added, it was enough for me to add a 1% commission to go totally negative. how can it be solved? Thanks in advance
Yeah, I think you missed that entire part about curve fitting and how it renders any strategy worthless. Additionally, once you use the word "Fit", you no longer have anything to do with AI, except that it uses the language model to write non-AI code for you (which, if you knew the language well enough, you could probably write as quickly yourself).
Its pretty clear you didn't watch the video fully, i definitely did not miss that, talked about curvefiiting at 06:00 and 07:00 and also how to solve the issue of curvefiiting. Next time before you comment put an effort in watching the whole thing or else the comment will look super silly
Bro don't just type in words. Ask how to do what your looking for or what is this used for. A another thing once it's answers you, you can ask a follow up question about the initial question and it will elaborate more
@@quantprogram great. I actually did see that. But see me, I move fast and in the beginning it....... No worries bro were all here tryna help each other. Thanks for your input and view on the subject. This is really a good way to go about it. I'm working on some things using this chat system as well 👍
Great vid! Can GPT do any sort of basic stock analysis for value investing to determine entry points on certain stocks for long term? Will it make u an excel for these etc? Or do you have to ask it to write code for tradeview in order to find this info. I am not particularly interested in daily quant trading but it is important regardless. And easing my workflow for stock picking would be awesome
// Define the strategy strategy("Long-Only Momentum Strategy") // Set the lookback period lookback = 50 // Calculate the momentum momentum = close - close[lookback] // Set the entry rule entry = momentum > 0 // Enter a long position when the entry rule is triggered if entry strategy.entry("Long",strategy.long) // Exit the long position when the exit rule is triggered if close
Not saying that you are this way but keeping it to yourself when there's such a massive market or markets this technology can be used for and keeping it all to yourself would just be being too greedy and my luck Karma would come very me and teach me a huge lesson for doing it. Plus if I had it and kept it from people I imagine the old thing called jealousy would find its way into my area and someone get that way and know about it and go tell everyone that will listen I got it and end up getting whacked for being greedy with it so before that I would Def be selling the crap out of it.
Hey thank you very much for your videos, it’s great content. One fast question, should I invest time in learning python or with Pinescript I can arrange a strategy and then make it work in Metatrader? Thank you
I think if you're series about your quant goal focussing on python or amibroker would be the best way to go. If you're just backtesting and not performing any additional quant resarch tools(which is absolutely necessary) then pinescript would be the way to go
Good video, but yeah work on your prompt engineering 👍 You are treating it more like a search engine than a conversing with a "person" that knows everything. Think of it that way
@@dancegalop If you are a non techie with no coding knowledge chagtgpt will help you a lot and this video should guide you. However if you are full time coder, i doubt chatgpt can help further except in understanding information or developments that coder is not aware off
@QuantProgram OK! nice to find your channel; I have wasted much time and money this year by following the major crypto influencers who are always pushed to the top of search results, while serious people like you are hard to find. Also, refreshing to see NO spam comments :)
This is quite interesting thank you for sharing!! I was trying emulate your process described but keep running into errors in Pinescript for "undeclared identifier" on the lookback, the momentum, and the entry terms. I'm not a coder --- can you do a vid on how to declare these identifiers? Or what you did not to have to further enhance the code you input to make it work? I have all other items set same from what I can see.
Hi mate, Please go through our video in ou channel "Trading view pinescript version5 full tutorial". Video is designed for absolute beginners with no experience in coding. Should get you up tp speed
@@quantprogram hey so I got it to work but the strategy is really only not that great considered how much 🤔 AMAZON went up during the time of your chart --- years and years no? Ie I'd you didn't pick Amazon which went up alot in prior years, or July used this strategy the last 5yrs... it really would have sucked performance-wise.
@@Idontdohandles Absolutely and i wont recommend at all to use this strategy, i didn't recommend it in in the video either . I assume you dint watch the video fully. I have discussed in the video about optimization and forward testing. No strategy is bullet proof until you go through portfolio backtest with position sizing, optimization, forward testing and Monte Carlo Simulation. This strategy has not been through that. The whole idea of the video is how we can use chatgpt to help us with finding information and codes for creating strategies. Nothing else. We have to do our bit by doing more research into the strategy by using these quant methods.
@@quantprogram sure thing I think you did reasonably well categorize what you were seeking to accomplish with the vid. The words I would have been looking for was something like "while it looks like this performed pretty well 20000% its also a good possibility you might have beat that return just buying and holding AMAZON over the same time period". Clear, concise, also would have outlined the time period tested. Just my .02.. . I greatly appreciate 🙏your efforts and knowledge here!!! Thank you!!!!
@@Idontdohandles Thanks much for the comment mate. Appreciate it. But with regards to buy and hold , there ar lots of preconceived notions. What most people don't understand is that buy and hold of a specific stock is not a good measure of comparison for performance of a strategy. The reasons are few things 1. 1000s of stocks listed around the world each year in each exchange. nobody can buy and hold each one of them. 2. Nobody can analyze all these stocks. Nobody knows what the next apple or amazon is. 3. Majority of these stocks maybe even more than 90% go bust or get delisted or get taken over. 100s each year in each exchange. 4. Because most of them were penny stocks including apple and amazon, their buy and hold returns since inception are substantially large. 5. Just because buy and hold returns of the stock since inception is 70000% for the past x years doesn't mean it will do the same next x years. We can start buying and holding now but don't expect that returns, maybe it can even go bust. Its easy to say "i could have bought amazon or apple back then" but at that point when it got listed most of people were not even aware of those companies(apple was listed in 1980). just like we aren't aware of the many penny stocks listed last month. This is why buy and hold comparison of specific stock to any quant strategy or any strategy is not ideal. Even warren buffet performance don't beat buy and hold of the stocks he owns neither does Li Lu or Peter lynch. The ideal measure is our portfolio performance relative to buy and hold of market, ie SPY over the specific time frame. This is the same goal that hedge funds and institutional investors all have, to beat the spy average annual cagr of 10%
Chat GPT is not a magic lottery ticket for trading. Its a programmers tool to easily create a script. If you don't know how to think like a programmer the Chat GPT script will not compile.🤔
i want to see how AI will act once we are limit Down all assets ahaha, that will be a fun watch .AT least this is one career i know won't be disrupted any time soon by AI
How is the $249-$299 course different from your free CZcams courses? And which algo gave you the 20097% in return since you have 5 in Alpha and 10 in How is the $249-$299 course different from this? And which algo gave you the 20097% in return since you have 5 in Alpha and 10 in Prometheus?
@@shahabnisar2817 How were you able to automate this? Ive seen pine connector, but all I see is sites saying you cannot automate tradingview pine code. What did you use?
Fantastic video! I get that you have a huge channel and are likely a busy individual but figure I should shoot my shot. Any chance you would be open to helping out with getting an active trading bot up and running using the LunaAlgo indicator?
Update I’m past that and I can’t get the images to show up on. Like the sending arrow and chat columns. Even though I completed that part without error and the browser loaded.
We are already in the big crash, Inflation is a catastrophe. This CPI report is a colossal failure. To bring the housing market to a halt, the FED will have to pull all the stops. The unfortunate issue is that other markets are being decimated. If you want to stay green, you have to rely on a lot of diversification. Currently up 14% and being careful. Still a better deal than leaving it in a savings or checking account yielding 0-1 percent interest.
It didn't give yout code, because you didn't ask it two write a program. Keep in mind, Chat GPT is a language model, not a search engine. You can't prompt it like a search engine. You must speak to it like you are speaking to a human programmer.
ChatGPT is not going to change trading you will never be able to just ask ChatGPT to guarantee you returns. If AI will ever be utilized to help someone trade a stealth startup will create their own AI for their use only and leverage AI to make trading strategies to benefit them. Quants are already experimenting with AI aided tradiing. And it's not click click copy paste and print money.
Visit our channel for more free Pinescript Strategies.
Timecodes
0:00 - Intro
1:45 - Create algorithmic trading strategy chatgpt
2:35 - Backtesting moving average strategy
3:30 - Generating Code for backtesting in chatgpt
4:27 - Download stock data code in python
5:25 - Python code for yfinance chatgpt
5:13 - Forward Testing
7:00 - Optimization
9:10 - Course Strategy results via Optimization
11:24 - Machine Learning Optimization Code in Python in Chatgpt
12:35 - Parameter Optimization
13:45 - Moving Average parameter Optimization
15:35 - API for brokers python
16:45 - Trading strategy in cloud chatgpt
17:30 - Pinescript code for moving average
18:40 - Trying the trading strategy generated by Chatgpt
21:00 - Asking chatgpt to create a 3000% strategy
22:00 - Creating a momentum long only strategy
23:00 - Backtesting th estrategy generated by chatgpt
23:40 - Tweaking strategy to get 20097% returns
28:15 - Finding intrinsic value of company using chatgpt
Dont take the strategy as financial advice. The whole video is just about how to get information and codes regarding financial trading and algorithmic trading from chatgpt. Further research has to be done before going live in any strategy. Thanks much. Let me know if you guy shave any doubts or clarification. Happy to help
@@scenefilm4230 He used Python. I do not get what you intent
But what you showcased means enterning on day 1 of history right? thats why it has such profit or am i wrong?
@@just_another_smoker8687 Not really. The strategy is backtested from inception of the stock. But multiple trades are done. However the results doesn't mean anything like i said in the video. We need to do other quant analysis like portfolio backtest, optimization, forward testing and monte carlo before we even think of going live.
The whole idea of this video is to help traders use chatgpt to get information and codes to backtest efficient strategies.
I’ve tested it with law cases… it gives general information and it isn’t quite right. It is OK… this is knowledge, not intelligence
where is the sript?
I don't think people realise in how many areas this is being used and how game changing it will be.
Alot of companies going to be taking hits!
Also you need to consider that majority of the population are not smart enough to use to ChatGPT to its full capabilities.
But, for those who do know how to use ChatGPT, it's going to total domination in their field
Something to keep in mind when using CGPT is it's most effective when you converse and collaborate with it. Within the thread, it remembers everything you and it have already said, so you can refer back to those things to ask for refinements and set expectations for future answers. What most people don't get is that everything you might say to another person, or in the case of the video audience, you can say to CGPT to help it provide better answers. Instead of saying to the audience, "it didn't say why it used something," or "it didn't use the right thing," say that to CGPT: "Why did you use that?" or "rewrite this using x strategy or technology."
An example of the expectations for future answers, you can tell it: "for all future code that uses pinescript in this thread, include the version comment at the top, set to version 4, and use 4 spaces for each indentation."
What's revolutionary about this technology is that if it gives you something that isn't what you are looking for, you can describe to it what is wrong with it or what would be more right about it, and it will understand that with natural language. The more you converse and refine the results, the better they will be.
Great content and thanks for the video! Hope this comment is helpful to whoever reads it.
Wow this was really insightful! I think a major misconception with this tech is that it is like a Q & A forum with the AI, not an open dialogue with the AI
Thats correct mate. Thanks for the comment. Appreciate it
I don't use it for Trading, but I do use it for coding and absolutely conversing with the AI is amazing. Initially it often provides wrong or bad code but with a bit of back and forth dialog it can produce some fantastic results.
@@jadebreed3333 Yes, true that once you disconnect it forgets. That sucks, but during a coding session it does pay off to converse ideas as it will alter it's outcome better.
@@surfingsub5854 couldn't agree more. It's incredible.
Investing in the stock--market and crypto is the best option to make a passive income. Virtually all the markets are crazy, most people pay more attention to the shiniest position on the graph, I'm keeping a diversified portfolio. !!!
Investing in stock is a good idea, a good trading system would put you through many days of success.
@@Stephenfanos37 That's impressive!!! I could use the expertise of this advisor, my portfolio has been stagnant..... How can I get started with PATRICIA DRITA JOSEFINA ??
1. You should look into prompt engineering to help you modify the responses you get from the AI based on how you word your prompts. It helped me immensely and i believe it will increase the quality of the results you get. 2. Try to fact check some of the information because the model can sound factual about non-factual subjects sometimes(OpenAI indicates this in their documentation). Great content.
I did not know this. Thats quality information mate. Thanks much. Will look into it
@@quantprogram Yeah, you can ask to explain CGPT's answers in detail, and you don't need to refrain your question every time. It'll follow and "remember" the conversation down to 1st question.
@@Qban220Thanks Jack. Will attempt it that way as well. I did notice this sometimes i ask the same question but its in a different chat, the wordings are quite different. I wont be surprised if chatgpt is improving on its information as more people asks questions
I tried making a simple trend indicator and telling cgpt all the errors i was getting, he kept modifying the script till it worked.
This thing is amazing. 🤖
edit: i used "he" instead of "it" lol 😁
Vino, where's a good resource to learn about prompt engineeirng? what's the best you'vve come to so far
The easiest way to succeed in any field is to work on it consistantly, and consistancy is the key to success. I experienced my self. By the way, thanks for the uploads , a fellow creator.
Well said! Thanks much for watching
Yes I trade one stock and have the chart up all day everyday. Ii think my subconsious has picked things up because I see pattern recognition am making money almost everyday.
@@beefybitcoin8085 this is right. The subconscious mind is very powerful.
I am blown away by the versatility of AI . It's amazing to see how this technology can analyze market trends and make informed decisions in real-time. I can't wait to see what other innovative ways AI will be used in the future. Great job on the video!
Including your existence??
@@VEILSIDEFTO01 😂😂😂😂😂😂bold statement
@@VEILSIDEFTO01that’s not going to happen. The worst that’ll happen is that we’ll merge with AI.
It's kind of an unusual strategy that it used for generating 20000% profit. If the current price is higher than the price 50 days ago, it's a buy. This algo could be modified to track momentum over the past few time periods to see if momentum is increasing. The takeaway I have from this is that ChatGPT can help someone start learning something new and as they apply the knowledge, they can learn further on their own and still use ChatGPT to further their learning as they refine their requests and start new conversations.
This thing has walked me through learning python and Keras, it also walked me through changing an engine manifold with specific information about the vehicle. Don't expect it to do brain surgery and understanding how it works helps you to engineer the prompts better.
In the future it will help with brain surgery and much more. It’s still in its infancy, think of it like a 2 month old baby that can teach you quantum physics.
@@femiairboy94 I was referring to this particular model, newer models with more perimeters I'm sure will have more functionality.
I tried using chatgpt for simple web dev. If you ask for something straightforward it works, but the more you ask for modifications and adding different technologies, it creates bugs because I don't think it really knows how to combine them. So you can use as a template but still need to review code and verify it makes sense.
Exactly. I think its more beneficial for newbies but still an individual should have good coding background to fix all this erros
It’s best to generate code in chunks and not let it generate an entire app or file. The bigger the prompt gets, the faster it’ll forget context due to the restriction of tokens (8000 I think)
One major oversight was made that changed everything. With ChatGPT you’re having an active conversation. Once you ran a test based on the 50 day sma everything else was based upon that…. You need to clear the chat to try a new strategy, otherwise you’re still talking about the same thing, ie the 50 day ma. Without clearing the chat you’re always working off of the beginning of the conversation. There are various other ways to improve upon the output as well, but I think that’s enough free Alpha for today. 🤙🏼
Thanks for the information mate. I didn't know that. I also realized giving more specific details can also change the results
@@quantprogram By now I’m guessing you’re a lot more impressed with the program, it’s crazy. The next model will be much easier to teach new things to 🤯
Absolutely mate. I understand the learning and development advantage the user has. But I'm quite wondering how its going to be in the personal space. Maybe we will reach a point that we don't know if the person we are talking to is AI or human. It also amuses me that its free. As if they are trying to collect as much as data from us as possible so maybe in future they can mimic us? Reminds me of movies like Ex machina or Her
Part of my fund is creating ways for investors to have a personal connections and opportunities to make choices on their own. Nobody is ready for a black box / chat bot investment world. So, I agree with the personal experience for people.
@QuantProgram I guess it’s free because it’s still learning and now learning very rapidly as more than 1m users are taking it for a spin. Win-win because it gets better the more it learns.
I was thinking of asking chat gpt to write me an algo a couple of days ago and today I find this video !!!
Excellent material although as someone reminded me back testing does not indicate the strategy would work today as the observable data does not factor inputs of real time trading which would change the outcome on the system.
The problem with optimization is that you are overfitting the strategy. Marcos del Prado makes a good insight in this toppic
3:08 - you could communicate with it almost as a human, so you can refer the previous parts of the conversation and ask it to give it in code. Example: " Now show/give it in code." or "Show me [insert a small portion of a previous prompt, as a reference] in code", etc. The second example can be used and its variations for prompts and answers that are earlier in the conversation and are not the immediate last one.
Hope this helps.
P.S. You can also pre-define and tell it to give you way more detail information and even long lists of things, but you should add clarification that "you are limited in the length of your answers so after each answer I will say "Then what?" and you will continue from where you stopped until you answer completely" or something in along these lines, letting it know you know it is limited in its answers and that you will use a specific command to let it know to continue.
Good comment mate. Appreciate it
@N Gthis is what i dob
I appreciate your approach to teaching.. To my understanding this just proves how much we need an edge as investors because playing the market like everyone else just isn’t good enough, we just need to hold onto our hopes and wait to see how things turn out because market movements are almost always unpredictable. In my portfolio, I'm noticing more red than green.
This is really cool. Idk where to start learning this stuff though the financial stuff alone is going over my head.
I tell you how to do it and save your time. ADX moves up above 40 level (14:5 settings) and stochastic (9/5/5 setting) goes up above 50-level (buy) goes down below 50-level (sell).
Works on any time frame. Thank me later. No bots needed.
Do you mind sharing your entry and exit rules please?
@@jaybee6347 Like said BUY when ADX indicator slopes up above 40 level + MACD slopes up above the zero level + Stochastic slopes up over its 50 mid-line level.
Just cannot go wrong with this and the opposite applies for SHORT when the ADX slopes strongly upwards but all other indicators slope downward.
Take profit 1:5 to 2 and SL goes below the previous price closing low/high.
Trading is a lot about waiting for the correct set-up and when it happens you go in heavy. Avoid mediocre set-ups and just buy/sell the premium ones.
I only trade intraday with 30-minute time frame, this way when I go in on a trade it runs longer that what it would be do a 1M or 5M time frame.
@@FullFinnoy as simple as that? Brilliant, thanks for the response, I'll back test!
@@jaybee6347 Cool, let me know how you find it. As per writing this mornign here in Europe, the US30, US100 and US500 were dipping with these exact values and ADX was up-sloping, Again I was able to hold short on all the indecies for good 45-minutes before a short correctiong is been seeing and I am out of the positions until I see the indicators aligning again to a mutual direction.
Let me know how your back test goes with intraday values and time frames ether its from 1 minute, 5, 15 or higher. Happy to hear the results.
@@FullFinnoy just one more thing, sorry! What MACD settings are you using? Thanks
I don't see any edge being provided. Writing Pine script is very easy for anyone who knows how to program. ChatGPT just saves you the time and money of hiring a developer or learning to code? I have 2 years of Pine experience could easily code anything you did in this video in half an hour or less
Oh boy! I tweaked it further and saw a return of 5,623,187.13% !!!! for the same AMZN daily chart and time period (Sept'1997 to Jan'2023!! thank for being the torch bearer. I am still pinching myself and can't stop playing with it, even tried the short side as well. I know it is difficult to believe, but if 20,000% plus result was unheard of, I know what folks will talk about this number.... itching to implement it live with $1000 only to start!! ....will come back in few years from today. Never ever commented before, but this result was to much to ignore and not share. Thanks a ton!!!!!!!!!!!!!
Please keep in mind that's amazon during a massive bull market. Everyone's a genius in a bull market. Even bear markets can make a lot of money, what about when markets range? If something is very simple and "too good to be true" it more often than not is
Sorry, but it's trivial to find a strategy that works during backtesting. And do you really think making a trading bot is this easy?
Please send me a download for the script please.
@@heeroyuy298 Yes, after 40yrs of playing with data/databases and 23yrs of trading, I finally found the holy grail. I am not using any of the indicators in this or any other tutorials/youtube/chatGPT etc., but created my own indicator which gives even more astonishing results live on ThinkOrSwim (TdAmeritrade platform) with real money. The moment I make couple of millions with it, I will launch a trading portal for public use with a server farm in the cloud to take care of any load. At the current rate of returns, expect to launch it around July'2023 !! It could be a gamechanger for many, as it is turning out for.
@@obelarcapital I get astonishing results with BTCUSD, COIN, TSLA etc. In fact any symbol with ATR(average True Range)>5
iv been using chat gpt to build an online business from scratch. shocking how much of the work it can acrually do for you
So glad I found this channel. Finance focused programming has immediate use-case but is so perplexing to me.
Do you have a background in finance?
@@lapatria100 hi Rob, sorry I’m late, I do not! Zero background in finance unfortunately
It’s remarkable how you are misspelling words and it knows anyways what you intended.
I think you should have asked for BOS and Choch code. Because that's the thing which gives more profit rather than any other strategy.
Try with SMART MONEY CONCEPTS BY (LUXALGO) SCRIPT on Trading view
@@micarranco yes i did
TOTALY SHOCKED!!! ty for content, now i'm forever your sub.
i tried chatgpt the same day it was available for metatrader code :) you need to specify of what you want exactly, lets say after a code created , you can tell him add to this code moving average and it does what you tell him, sometimes the code is too long, and you have to reenter last 2 rows of code or full script he wrote to continue
Sweet baby jesus, its the indentation..
Bro. I've been trying to get chatGPT to write me some scripts but not knowing pinescript all that well, it errored out for no apparent reason so in the end I just abandoned the whole 'write me a script' idea.
I have some sick ideas.
Thank you!!
Happy it was helpful. Watch the "Tradingview Pinescript version 5 Full tutorial" video on our channel if you wish to master the coding bit in pinescript. Thanks for watching and for the comment mate. Appreciate it
Those of us in the trading world know you could build such devices with a super computer and still lose your shirt. Unfortunately some people are too lazy to "THINK", and look for every conceivable machinery to make trading easier.
You got that right brother. Unfortunately people new to the space think it's like printing money. They'll learn.
I retyped the code and wasn't able to reproduce your results -- nowhere near 20097% return. more like 1.2% return
Same here. I got 3.6%.
I believe that this recession has taught individuals the value of having several sources of income. Unfortunately, having a job does not guarantee security; rather, the real security comes from diversifying your investments. When it comes to speculating in the financial market, understanding institutional investment is essential. Whether it's real estate or cryptocurrency, investing in knowledge will yield the highest returns. Since I started trading with Mr. Thierry Veilleux, whose skill sets are extraordinary, I have seen enormous profits on my investment. Working hard is good, but working wisely is preferable. Trade with a professional! It really is simple.
Oh my days! I'm stunned. Thrilling to meet other students of Veilleux Thierry, what a great mentor
I count myself as a successful trader because of the amazing trading strategy of Mr. Thierry
Now I understand what it takes to win in trading, Thierry is good >
Very good video. It’s interesting to see the variety of things you can use AI for. Keep up the good work!
Thanks, will do!
i was a very good algo trader, but left this field bcoz i was not able to make enough profit.
now looks like i need to revisit my skills. and start again...
Great content. Subbed.
More, please. I'm interested in learning a language for trader station and want to work out the best coding to learn in order to notice when chatGPT gets it wrong or has missed something. Looking forward to seeing how a newbie can spot and handle the bugs in the chat code.
Big question, where does it get information. I don't think even Google AI could have been smart without support and consent of original content creators. You can't self learn by yourself. Literally it has many flaws especially regarding copyright laws.
ChatGPT is the best tool so far for research, do homework, create stuff by order and to provide any info that is needed. AI is the best thing after Blockchain technology in my opinion.
Blockchain is still looking for a problem to solve whereas ChatGPT3 is a serious step towards KI and will definitely be very useful. The technology behind blockchain is essentially cryptography and dates back to the 70ties, so really old stuff.
You're not talking to the bot ......... you're trying to "query" stuff the same way you do with Google. That's not the way this is designed to work. It's called "Chat" ... it's a language model you need to "talk" with it in human language not just through keywords at it like we do with Google
Where is chat gpt pulling its information from to even have these conversations?
What we really need is a system that accepts ideas, then does backtesting (multiple periods) and Monte Carlo simulations in bulk volume, without all the manual work. This would "solve the market." Later, a system to generate testable ideas... and then what is left to automate?
assuming you can simulate the market is the first mistake
This already exists tbh, I found it not too long ago made by a solo developer haha. Just trying to figure out my Python code now for applications but I can steer you to it as well bro
@@StereotypicalShea what’s it called?
@@StereotypicalShea steer please?
@@michaelholmes2846 BuildAlpha, just remember me if you find anything nutty lol
Thank you for the video. It's very interesting. However, referring to the actual Trading strategy that produced 20 thousand % need to be backtested with what would happen if you just buy at the begining of the period and just sit and look how your investment growth. I tried it in TV and got pretty similar results, a profit of 19.620. However in my list of transactions, the first one is a buy at 0,12 cents. Taking into account that the current price in your example is around 86 per share you've got 72k% return just watching TV (television not Trading View) . So the massive return is for the Amazon stock.
See the more you have basic conversation and ask for you you want it's works straight forward
i copy your code and its not showing up and i dont have any foult code
Need answers? Read my tag
Interesting but you can get nearly infinite returns on past data. The rest test is applying this strategy for a few weeks to see what the results are. I wish you would have included the results of that in the video!
Absolutely and i wont recommend at all to use this strategy, i didn't recommend it in the video either . I assume you dint watch the video fully. I have discussed in the video about optimization and forward testing. No strategy is bullet proof until you go through portfolio backtest with position sizing, optimization, forward testing and Monte Carlo Simulation. This strategy has not been through that. The whole idea of the video is how we can use chatgpt to help us with finding information and codes for creating strategies. Nothing else. We have to do our bit by doing more research into the strategy by using these quant methods.
I noticed that my backtesting of a week back is more reliable than backtesting all time
Coukd this mean we should regularly backtest strategy settings weekly?
Only problem is chatgpt is only as smart as 2021...
it might be supersticious but i think if you say please and thankyou the AI feels appreciated and loved and gives you a better answer
This has so much potential it's crazy. I feel if this is perfected and becomes wide spread stock exchanges would ban the use of automated bot trading. How'd they go about is hard to say but personally I feel it is a possibility. Amazing video btw!
Thanks for watching mate. I'm more worried about how this can affect our life personally. Its free that makes me super suspicious. Is it collecting our data so much to use it against us?
@@quantprogramow can it be used against us when we're just using it for basic knowledge
@@cellc6191 Well my theory is by us seeking basic knowledge, they get to know what we want to know. If they get to know what we want to know they can control or limit that information or manipulate the information. Manipulation can go deep, just like people find beauty standards these days in social media. It can also destroy jobs. As in the video i tried to get some codes, even though its error prone and not perfect it leads to less people being hired in certain departments of coding industry. Or marketing industry for creating good images. I mean this i just what i think. but i dunno what the future holds. Could be worse or could be not that bad
@@quantprogram well said
@@cellc6191 Thanks mate
Sorry to burst your bubble this has been going on for a while know it’s been used left right inside in out
i wanna try this first in local market instead of directly going for the international market
Here's an interesting idea, try to teach it a few times before actually asking for a strategy. Like priming it before shooting it. Feed it data to base off of
Hi Christer, yes you are correct, it works much better that way. I created the video when chatgpt just came out and wasn't aware of the different approaches in prompts to get better results. There was very few info available online when it got released either. But now since chatgpt is taken off there are better info about ideal prompts including the advice you gave to get better results.
Thanks for the comment,
Very easy strategy, need more conditions
At around 3:00 I tried the exact statement, but instead of Python, I put in PHP. It ended up giving me the code in PHP. I have my own web host, so I can try it out on my web server.
@LivingAndroid You can re-do the statement like he said, but use PHP instead of Python the way the guy in the video said.
Another flaw I noticed in your conversation, is that it's not a conversation. You aren't building on things. You point out a flaw, but you don't talk to ChatGPT about it. It can actually remember what it was talking about... It feels like you're using it like DALLE2/Stable Diffusion. "I know, let's throw python at the end of the sentence!"
Really reminds me of the computer on star trek. Shatner just passed on.
This is the beginning of the end of trading.
You can ask it to write it in pinescript v5 and you can ask it about the errors you get.
20.000% revenue pretty decent? Am I missing something? Which period did it backtest on? Seems quite unbelievable especially considering it's such simple strategy.
Daily candles maybe the who the chart history of amazon
Do you have the copy of the code?
can i have a copy of that code please?
ChatGPT isn't a search engine, it's an ai, talk to it instead of using keywords
I think AI trading is going to be big. And AI companies therefore insanely rich if they use their AI for trading. Not sure the advancements will come from llms though.
It won't IMO, maybe for a short period. If everyone is going to use AI for trading, no one will succeed. It will just make market behave even weirder. To succeed in trading (as opposed to long-term investing) you have to do something differently, have an edge (or have insider info) to screw your fellow traders up. There is no magic method in trading and no concrete rules for success. AI generated text content will spam already spammy internet to further reduce the amount of quality content to learn from. Similarly, AI based trading will make markets unstable and always changing, making it difficult to train a new generation from. I am not very happy from all of these aI developments... but we'll see
Isnt that what aladdin is for?
Mahatma Gandhi said: “A small body of determined spirits fired by an unquenchable faith in their mission can alter the course of history.” It seems like the scale for being literate just got tipped, so now using AI is going to be the norm, just like using google once tipped the scale and now most literate people know how to use it, let see how much we have to pay for it. (soon) lol
I was wondering how they gon charge us, eventually…
can you share basic video which show how to take actual trade. i thing you are using tradingview but how to earn actual money
Thanks for this great content! I just subscribed as I'm new to quantitative trading field.
I'd like to ask : to which extent an average quantitative trader can make profit ? I mean, I'm not new to trading, I know the 90% of traders lose money rule etc.
But I'm asking myself, is a decent return (for example 15% a year) reachable to anyone willing to put the efforts in ? Also, I was asking myself, is there any other languages useful to learn apart from python and pine script, the latter being similar to the former ?
+ Last question but not least : Have you been able to trade with your own quantitative trading strategy ? If so, what were your results if you're willing to share them ?
I think personally from my experience, I always look to beat the market in the long run. By market i mean SPY buy and hold CAGR returns. That roughly comes around 10%. So my goal is to beat that, not year on year but on average on the long term. My portfolio of trading strategy is not purely quantitative strategies alone. It has got value based investing and also a few day trading strategies. This year since the market has been down I was able to purchase lots of good companies at bargain prices. From the quant side, the mean reverting strategies has also performed well. However from a day trading perspective it has been a huge struggle. The advantage of quantitative trading is that you know what your'e gonna get. The forward testing and Monte Carlo Simulation helps you prepare for the worst case scenario regardless and from my experience it has gone accordingly to the expected results. The cons of quantitative strategy is that you need at-least 5 strategies at the same time and also the belief to stick to it. I remember earlier this year the strategy that i have explained in 09:10 didn't work that well, there was a few consecutive losses and you feel like you want to shut it down. The temptation exist but this results was expected and eventually it started making spectacular returns after couple of months. I also dont use leverage in neither my quant based or value based strategies which think have benefited me a lot unlike the other traders. The other con of quant based trading is there is lots of room for improvements. Like now I'm learning lots of machine learning models to predict volatility contraction and expansion so i can apply in option prices to benefit the theta decay in far OTM options. So there is a constant need and urge to learn.
@@quantprogram Thanks a lot for your time writing this very helpful answer. Looking forward to learn a lot thanks to your videos !
@@aurelienn.9789 Thanks mate
@@aurelienn.9789 you can achieve 15 solely on S and P 500
@@nickoleo7817 how is your S&P 500 trade going this year? 🙃
Great idea. Why shouldn't I use an open source tool to try and compete with google and the wallstreet quants?
What could possibly go wrong?
It can not reason. It can only provide knowledge. It is not active AI. So it can not be used to trade, unless you are trading in the past.
Thats exactly what i said in the video. I assume you didnt watch it fully. The whole idea of chat gpt is to get knowledge and information especially coding knowledge, not to predict anything
@@quantprogram also… people think they can sell or use the data produced for commercial purposes…
This is Napster only for data. I see many people spending the next 5 years in court losing everything.
@@thatguy7085 Thats an interesting point
@@quantprogram even worse… now that I’ve tested it and seen what other people think they can do with it… this is a lawsuit waiting to happen.
This gathers copy written material and presents it in its own chat form, without reference to the originating author. Then people are suing this data as if it was their own. Which is not only a federal criminal fraud, but also a civil liability action.
Weird thing… chatGPT doesn’t understand this.
Do you have videos that show how to incorporate human input into the tradingview bot? For example, you want the bot to follow a certain protocol according to trendlines or levels you draw on the chart, or other manually inserted variables.
Exactly! Specific protocol tailoered. For me, I want it to hedge from the get-go but not like most of those trash-bots ppl make. More like a real trader w/ slower solid action for bigger returns. Maybe if A.I could watch tons of the traders style for each specific strategy so it could machine-learn (mimic) it to a T
very interesting man!
in the parameters of the strategy settings the trading commissions should also be added, it was enough for me to add a 1% commission to go totally negative.
how can it be solved? Thanks in advance
It only does what you ask it,
Not what you want.
Like code
Because it is code
Yeah, I think you missed that entire part about curve fitting and how it renders any strategy worthless. Additionally, once you use the word "Fit", you no longer have anything to do with AI, except that it uses the language model to write non-AI code for you (which, if you knew the language well enough, you could probably write as quickly yourself).
Its pretty clear you didn't watch the video fully, i definitely did not miss that, talked about curvefiiting at 06:00 and 07:00 and also how to solve the issue of curvefiiting. Next time before you comment put an effort in watching the whole thing or else the comment will look super silly
Bro don't just type in words. Ask how to do what your looking for or what is this used for. A another thing once it's answers you, you can ask a follow up question about the initial question and it will elaborate more
i assume you didn't watch the whole video. i did all the things you said
@@quantprogram great. I actually did see that. But see me, I move fast and in the beginning it....... No worries bro were all here tryna help each other. Thanks for your input and view on the subject. This is really a good way to go about it. I'm working on some things using this chat system as well 👍
@@immanueljenkins5036 Thanks much mate
@@quantprogram no thank you 👍 maybe we can collaborate on things in the near future with what we discovered
Great vid! Can GPT do any sort of basic stock analysis for value investing to determine entry points on certain stocks for long term? Will it make u an excel for these etc?
Or do you have to ask it to write code for tradeview in order to find this info. I am not particularly interested in daily quant trading but it is important regardless.
And easing my workflow for stock picking would be awesome
Do you think you could release this strategy as a indicatior on tradingview that gives you alarms what you should do
For how much time does tradingview backtests this strategy that shows 20000 percent returns ?
After you ask a question about coding, if it doesn't output the code, just simply say "show me the python code for that"
So... you can plug this in and just let the code run, and that's it?
// Define the strategy
strategy("Long-Only Momentum Strategy")
// Set the lookback period
lookback = 50
// Calculate the momentum
momentum = close - close[lookback]
// Set the entry rule
entry = momentum > 0
// Enter a long position when the entry rule is triggered
if entry
strategy.entry("Long",strategy.long)
// Exit the long position when the exit rule is triggered
if close
Something that i never understand, if I had the best tools to make money trading, why should I bother to share it with people?
Some strategies work best when shared with max people. Some work best when not shared.
Self fulfilling strategies are shared the most.
Not saying that you are this way but keeping it to yourself when there's such a massive market or markets this technology can be used for and keeping it all to yourself would just be being too greedy and my luck Karma would come very me and teach me a huge lesson for doing it. Plus if I had it and kept it from people I imagine the old thing called jealousy would find its way into my area and someone get that way and know about it and go tell everyone that will listen I got it and end up getting whacked for being greedy with it so before that I would Def be selling the crap out of it.
Does 400% ROI is good return?
Code??
Hey thank you very much for your videos, it’s great content.
One fast question, should I invest time in learning python or with Pinescript I can arrange a strategy and then make it work in Metatrader? Thank you
I think if you're series about your quant goal focussing on python or amibroker would be the best way to go.
If you're just backtesting and not performing any additional quant resarch tools(which is absolutely necessary) then pinescript would be the way to go
Good video, but yeah work on your prompt engineering 👍
You are treating it more like a search engine than a conversing with a "person" that knows everything. Think of it that way
After watchin 21:20 I laughed hard because if it was going to work, this video was never shared with us.
Lol
@@quantprogram HANG ON, so, it does NOT work? I am skimming the comments to find out if it's worth studying the video (as a nontechie)...
@@dancegalop If you are a non techie with no coding knowledge chagtgpt will help you a lot and this video should guide you. However if you are full time coder, i doubt chatgpt can help further except in understanding information or developments that coder is not aware off
@QuantProgram OK! nice to find your channel; I have wasted much time and money this year by following the major crypto influencers who are always pushed to the top of search results, while serious people like you are hard to find. Also, refreshing to see NO spam comments :)
This is quite interesting thank you for sharing!! I was trying emulate your process described but keep running into errors in Pinescript for "undeclared identifier" on the lookback, the momentum, and the entry terms. I'm not a coder --- can you do a vid on how to declare these identifiers? Or what you did not to have to further enhance the code you input to make it work? I have all other items set same from what I can see.
Hi mate, Please go through our video in ou channel "Trading view pinescript version5 full tutorial". Video is designed for absolute beginners with no experience in coding. Should get you up tp speed
@@quantprogram hey so I got it to work but the strategy is really only not that great considered how much 🤔 AMAZON went up during the time of your chart --- years and years no? Ie I'd you didn't pick Amazon which went up alot in prior years, or July used this strategy the last 5yrs... it really would have sucked performance-wise.
@@Idontdohandles Absolutely and i wont recommend at all to use this strategy, i didn't recommend it in in the video either . I assume you dint watch the video fully. I have discussed in the video about optimization and forward testing. No strategy is bullet proof until you go through portfolio backtest with position sizing, optimization, forward testing and Monte Carlo Simulation. This strategy has not been through that. The whole idea of the video is how we can use chatgpt to help us with finding information and codes for creating strategies. Nothing else. We have to do our bit by doing more research into the strategy by using these quant methods.
@@quantprogram sure thing I think you did reasonably well categorize what you were seeking to accomplish with the vid. The words I would have been looking for was something like "while it looks like this performed pretty well 20000% its also a good possibility you might have beat that return just buying and holding AMAZON over the same time period". Clear, concise, also would have outlined the time period tested. Just my .02.. . I greatly appreciate 🙏your efforts and knowledge here!!! Thank you!!!!
@@Idontdohandles Thanks much for the comment mate. Appreciate it.
But with regards to buy and hold , there ar lots of preconceived notions. What most people don't understand is that buy and hold of a specific stock is not a good measure of comparison for performance of a strategy. The reasons are few things
1. 1000s of stocks listed around the world each year in each exchange. nobody can buy and hold each one of them.
2. Nobody can analyze all these stocks. Nobody knows what the next apple or amazon is.
3. Majority of these stocks maybe even more than 90% go bust or get delisted or get taken over. 100s each year in each exchange.
4. Because most of them were penny stocks including apple and amazon, their buy and hold returns since inception are substantially large.
5. Just because buy and hold returns of the stock since inception is 70000% for the past x years doesn't mean it will do the same next x years. We can start buying and holding now but don't expect that returns, maybe it can even go bust.
Its easy to say "i could have bought amazon or apple back then" but at that point when it got listed most of people were not even aware of those companies(apple was listed in 1980). just like we aren't aware of the many penny stocks listed last month.
This is why buy and hold comparison of specific stock to any quant strategy or any strategy is not ideal. Even warren buffet performance don't beat buy and hold of the stocks he owns neither does Li Lu or Peter lynch.
The ideal measure is our portfolio performance relative to buy and hold of market, ie SPY over the specific time frame. This is the same goal that hedge funds and institutional investors all have, to beat the spy average annual cagr of 10%
Chat GPT is not a magic lottery ticket for trading. Its a programmers tool to easily create a script. If you don't know how to think like a programmer the Chat GPT script will not compile.🤔
i want to see how AI will act once we are limit Down all assets ahaha, that will be a fun watch .AT least this is one career i know won't be disrupted any time soon by AI
How is the $249-$299 course different from your free CZcams courses? And which algo gave you the 20097% in return since you have 5 in Alpha and 10 in How is the $249-$299 course different from this? And which algo gave you the 20097% in return since you have 5 in Alpha and 10 in Prometheus?
the trouble with this IMO is people will follow trends even more extremely plus the bot only has access to the information you give to it.
Great video. How can you automate these strategies on the stock market? Is there something like 3commas but for stocks?
search pine connector i did it
@@shahabnisar2817 How were you able to automate this? Ive seen pine connector, but all I see is sites saying you cannot automate tradingview pine code. What did you use?
What are you trading on? Like what are we looking at when you pull up your bots and charts
This seems like a way to get free money essentially, with only a little outside input. Why doesn't everybody do this and make millions?
Fantastic video! I get that you have a huge channel and are likely a busy individual but figure I should shoot my shot. Any chance you would be open to helping out with getting an active trading bot up and running using the LunaAlgo indicator?
You need to give it an action. For example: write some python code.
New subscriber here. This looks interesting. I will dive in to it right now.
I’m stuck at nose -v it keeps saying it’s not recognized, but I’ve already downloaded it and updated tools.
Update I’m past that and I can’t get the images to show up on. Like the sending arrow and chat columns. Even though I completed that part without error and the browser loaded.
If ChatGPT is not connected to the internet. Then what is it connected to ?
would love to have you on are Podcast...
We are already in the big crash, Inflation is a catastrophe. This CPI report is a colossal failure. To bring the housing market to a halt, the FED will have to pull all the stops. The unfortunate issue is that other markets are being decimated. If you want to stay green, you have to rely on a lot of diversification. Currently up 14% and being careful. Still a better deal than leaving it in a savings or checking account yielding 0-1 percent interest.
It doesn't know PINE script version 5.
hey, sir! Can you make a tutorial on how to connect this or any other strategy to an exchange to automatically execute trades ?
It didn't give yout code, because you didn't ask it two write a program. Keep in mind, Chat GPT is a language model, not a search engine. You can't prompt it like a search engine. You must speak to it like you are speaking to a human programmer.
I Assume you dint watch the whole video it gave me code multiple times
Hi.. I lost 1000 dollars in market.. Can I recover with your guidance...
ChatGPT is not going to change trading you will never be able to just ask ChatGPT to guarantee you returns. If AI will ever be utilized to help someone trade a stealth startup will create their own AI for their use only and leverage AI to make trading strategies to benefit them. Quants are already experimenting with AI aided tradiing. And it's not click click copy paste and print money.