Curious if anyone has been using or trying to use ChatGPT for technical decision planning?
I have tried a few items and it is interesting. You can’t just ask things like, tell me the best technical indicator or strategy. However, I have asked it to write some simple thinkscript code for simple thinkorswim tasks like flagging ema crosses and calculating position sizing and it does relatively well. Writes in Python, which seems to work in thinkscript with some expression exceptions. But it generates enough backbone that you can often fix it’s errors.
Actually have a small script that I wrote to generate on my trading chart a position size given a set dollar amount, the ATR and actual price to set a stop-loss as well as potential percentage profit targets. I use it as a handy reference to quickly choose position, a flat ATR stop (if there is no other support/resistance level that I prefer) and list a potential profit target to set a OCO order for a short trade. Asked ChatGPT to write one and it did. Had a couple glitches but also streamlined my original code, so merged it.
Did ask it to code for Triangular moving averages, but it didn’t work. But I have it’s suggested format that I’ll try to tweak to work. Thinking about trying to use it to see what it thinks of SAR.
Anyway, the free user level is easy and obviously free. Curious if anyone has tried it.
There are some Youtube videos out there now, I’ve only looked at one that was using it to write some code for Tradeview, which is Pinescript. Interesting. Give it a shot and please drop a note about your experiences.
Haven’t tried ChatGPT, nor will I. I see no reason to aid or abet Gate’s evil empire. Also, there are huge problems with translating anything done with AI to real world investing/trading situations due to the fact that them writing the code don’t understand how markets work. For sure, stuff like neural nets is fun to play with. But how many toys does one need?
Understand and respect your position Arindam. However, I have two reactions. First, in response to your statement of:
I absolutely agree, except that I see a lot of strategies and approaches from humans that I would make the same argument that they don’t understand how markets work. But then, I don’t proclaim to totally understand either. But that doesn’t mean I shouldn’t listen to and consider their perspective. Nor does it mean I need to “drink their Kool Aid” either. I’m very happy to get lot’s of perspectives and toss out what doesn’t seem to fit for me and my perspective. Learning is exploring.
Secondly, AI is here and it will be a part of the market. Period. I have no doubts that the big boys and down to the retail traders will use it. Not sure that is good or bad, but just reality. Understanding it better and having some idea of what it can do seems prudent.
That’s actually a big part of why I posted this, to get perspectives on it, not to sell it nor dump it.
Lakedog
And I don’t know, I kinda like playing with toys, just have to keep in mind they are only toys and need to be put back on the shelve sometimes.
Yeah, AI is here to stay, and it has been a part of the investing/trading scene for a lot of years now. E.g., look up Doyne Farmer and the work he did at the SantaFe Institute. But doesn’t mean that AI and its linear regression models have anything worthwhile to offer anyone involved with the complex, chaotic systems that auction markets are. In fact, just the opposite is the case. The idiots who believe in garbage like MPT blew up LTCM in '98. They blew up markets in '03. Again in '07-'08. And they are trying hard to do it again under the guise of MMT.
If it were just their money they’re putting at risk, I wouldn’t mind. But the Fed and Treasury cartel is bankrupting this country for two reasons. #1. They haven’t a clue how markets really work, nor any idea of what real money is. #2. They have no skin in the game.
What’s going on here is well described by Kuhn in his The Structure of Scientific Revolutions. On one side of the battle field are Chicago School “economists” like Malkiel, Merton, Markowits. On the other side are guys like Taleb, Madelbrot, and Rickards. Obviously, my sympathies are with the latter.
As for playing with AI toys, for the sake of having some fun, not a problem. On one of my old computers, somewhere, is a neural net program that did a surprisingly good job of forecasting future stock prices from past prices fed to it. But would I trust such a program to make big money decisions for me? Not a chance, because when the bet failed, the damage would be more than I’d want to try to recover from.
Point well spoken and taken, Arindam. But I’ll still play with it to help me write thinkscript and VBA.
Lakedog
Tried to think of a good boat building question, but couldn’t come up with one. Although, it thought boat building was “Building a wooden boat can be a rewarding and satisfying project, but it requires careful planning, skill, and attention to detail. Building a wooden boat can be a complex and time-consuming project, and it may be helpful to consult with experienced boat builders, attend workshops or classes, or reference detailed building plans and instructions.” At least it respects you!!
Script editors are whole 'nother matter, as are many “expert systems” that can figure out problems faster and more accurately than us mere humans. But how to invest isn’t one of them, due to flawed assumptions about how markets really work and their inherent nasty surprises.
Actually, building a boat isn’t harder than building a box, and humans have been cobbling together boats for at least 10,000 years. So the technologies are well known.
Chatbots aren’t new. Back in the early 70’s, when I was taking a programming course at Berkeley, the instructor demoed one then running at the Lawrence Science Lab, called Lisa. Yeah, it could/would respond in a seemingly human manner. But it could easily be made to fail the Turing Test. Same-same today, 50 years later. GIGO, and a huge waste of time.
Got it, you don’t care to look at it. Though, Lisa to ChatGPT is a bit like a Model A compared to a Tesla. Regardless, the point behind the thread was to explore AI since it IS a part of stocks going forward. You don’t have to participate. Probably not a problem, thread pretty much squashed.
Gees, you give up easily. If you really think that ChatGPT is useful, and not yet another unnecessary evil, then make the case why that’s so with specific examples others can test for themselves.
What do would-be investors most need? Not more money. But how to make better use of the money and the opportunities they already have. In short, they need sound advice. Can ChatGPT provide that?
It’s not a contest or debate to “give up” on or win. As I stated in the original post, I just got exposed to it’s use and wanted folks to be aware that they can try to apply it to whatever they do with stocks. I stated that you couldn’t just say, “tell me what to trade” or what indicator. However, it does work easily with coding. It’s been very interesting for me in building Thinkscript and Excel VBA. But that is obviously only the surface. So it was an open request for input from others who have tried it and how. Not a debate on it’s form, validity or social implications.
There are multiple Youtube videos that show folks building strategies with ChatGPT and fine-tuning them with it. Mainly in Tradeview, but applicable to others. They are a bit boring as much of it’s coding, but the principles and concepts are interesting. Here’s a couple, NONE of which I am endorsing. I present them as an example of how others are approaching it’s use. None of which I am doing.
As a side note, Wall Street Journal today had an article on JPMorgan and that they are restricting employees from using ChatGPT. No real explanation, may be more of a move to block AI generated emails, etc. Other nonfinancial companies are doing it also.
As I stated in the original post, for me, at this time it’s usefulness is in generating thinkscript and VBA. I have absolutely no intention of using it to make trades or giving me advice. But I am sure there are some smart people in these forums who have tried using it for other things. That’s what I am asking and interested in hearing about.
But then, I don’t proclaim to totally understand either. But that doesn’t mean I shouldn’t listen to and consider their perspective. Nor does it mean I need to “drink their Kool Aid” either. I’m very happy to get lot’s of perspectives and toss out what doesn’t seem to fit for me and my perspective. Learning is exploring.
