Using AI For Trading | Podcast Episode #435

0:07
Hey, I’m Ryan Mallory and this is my Swing Trading the Stock Market podcast. I’m here to teach you how to trade in a complex, ever changing world of finance. Learn what it means to trade profitably and consistently, managing risk, avoiding the pitfalls of trading, and most importantly, to let those winners run wild.

0:25
You can succeed at the stock market and I’m ready to show you how. Hey everybody, this is Ryan Mallory with shareplanner.com’s Swing Trading the Stock Market.

0:35
In today’s episode, we are going to be talking about AI and algo trading. So does AI have a place? Artificial intelligence, does it have a place? And our trading strategy, does it have a place in our day-to-day trading?

0:49
And that’s what we’re going to talk about today from a guy who asks to be called Loki and he has a reason for wanting to be called Loki.

0:57
And you’ll hear about that at the end of the e-mail. But Loki has got some questions as he’s nearing retirement.

1:03
Loki writes, Hey Ryan, I’m a long time listener if playing your backlist counts and appreciate the straight up unglitzy advice delivered with a plum. A plum. A plum. I had to actually look that one up. What I didn’t do is look at how you pronunciate that. I I’ve heard the word before, but I think that’s how you say it. A plum he goes on the right. I’m a believer in if it’s too good to be true, it’ll cost you and so your information keeps things real.

1:28
For me, I have a reasonable long term portfolio of ETFs that I dollar cost average into and another small portfolio for active trading and can lose entirely without losing sleep.

1:39
My background is IT of over 30 years and I’m within 15 years of state retirement age. I started very early, I’m from the UK, that’s the United Kingdom but looking to move somewhere with less rain and want to use trading to supplement the retirement fund.

1:52
The job interests me and I believe I’ll get to this eventually and understanding it enough to make good decisions.

2:03
I have coded indicators and strategies on TradingView and have found that simple indicators based on short term, that is less than one week durations are manageable. That is looking at volume, price action, historical sentiment, and known oldies such as gaps do give a simple limited indicator. Advancing past this though becomes very complicated very quickly.

2:24
So then I thought, what about AI? After Googling these platforms, they look expensive and I don’t think my trading experience would understand how to best drive them. Now I’m not new to AI, it’s a great tech, but for me it incurs a lot of money and time.

2:40
And as a retail trader, would I use it enough to warrant the expense? Do I trade often enough where I need millisecond intervention?

2:48
Another part of me says you have to have the experience in order to recognize the good decisions. So how would I know if AI is making good decisions for me as a result?

2:57
Therefore, an AI shortcut would still leave me out of my depth none the wiser. What is your take on algorithms for use in trading and AI’s use in trading?

3:07
If you decide to use this for the podcast, please give me a shout out to my mascot, Loki, a gastrointestinally challenged cat.

3:15
Regards, Loki.

3:19
This is actually some new territory for us on this podcast. I haven’t really dove into the AI and how that relates to trading. Now for this podcast episode, I went on to ChatGPT and I played around with it some. I haven’t really done much work with it in the past, but I do have some strong opinions on the AI and my thoughts on how it would relate to trading. So I put in a simple command prompt.

3:45
I said to ChatGPT, and the reason why I use ChatGPT, it’s the most well known AI platform out there. Now, I think in his e-mail, he’s also talking about trading platforms that are run on AI. But here I just wanted to take a generic, most popular ChatGPT and just see what kind of information it gave us. So I said to it, I said I’m a retired person, about 55 years old.

4:10
That’s not me, but I just wanted to put that into the equation since he’s looking to get retired here. So I said OK. I’m a retired person at 55 years old. I want to earn an income in the stock market, swing trading. I need at least a one to two risk reward ratio on my trade with a success rate of at least 50%.

4:27
I need you to provide me with a trading system. I don’t care what kind of trading system it is, but it just needs to meet all the above criteria about the risk and reward. Plus it cannot have a drawdown of more than 10% and you must back test it for the last 15 years to ensure that is a profitable trading system.

4:40
I don’t want to have to make more than three trades a day.

4:46
The reason why I put that last part in there, I don’t want them to give me something that’s like 10,000 trades a day. I’m like, OK, that’s definitely not feasible because there’s—you could do back testing and so forth.

4:56
And one of the things that I found with back testing is you can put in these like really specific requirements and you may not have a trade for like a week or two. But when those requirements are met, you might have like 1 or 200 trade setups that are all like lighting up like a Christmas tree on the same exact day.

5:13
Well, you’re not going to be able to have 200 trade setups and it’s assuming that you’re putting 10% on each one of those and it just turns into a mess.

5:19
So there does need to be kind of like a max expectation for how much trading you’re going to be doing in a given day when you’re looking to back test the strategy or asking AI in this situation for a strategy.

5:29
So what the AI came up with, they say key parameters for the system: one to two risk reward, success rate at least 50%, maximum drawdown less than 10%, number of trades no more than three per day. And then it gives me components for the system.

5:44
And then it tells me about, oh, you should know about the 50, the 14 and the MACD. And then it gives me a buy signal.

5:52
It says the price crosses above the 50-day SMA, RSI is above 30, indicating that the asset is not oversold.

5:59
And then it tells me the MACD line crosses above the signal line. Your sell signal is the price crosses below the 50-day moving average.

6:06
RSI is below 70, indicating that the asset is not overbought. MACD crosses below the signal line.

6:13
That’s your sell signal and then you also use a 1% stop loss and you take profit above 2%. So I feel like the exit criteria was just essentially adding percentage to what I said.

6:24
The risk reward was not realizing that I meant that if I’m using a 3% stop loss, I want to be able to make, you know, 6% off of the trade or 4%, I want to make 8%. It just didn’t pick up on that and it tells me that it’s going to back test the whole strategy using Python.

6:41
But I don’t, I don’t really know much about Python, but OK, let’s assume Python can do whatever it says it can do, and then it can’t actually access any financial data. So it tells me it can’t do that.

6:52
So it’ll do a synthetic data to simulate the trading environment, apply the strategy to that data. Then it tells me it can’t do that either.

7:00
Then it tells me that it can do the synthetic back test. It tells me that the total return using its strategy is approximately 104%, which I don’t know if it’s basing that off of the last 15 years, what I told it to back test. And then its maximum drawdown is 224%, which that means I just blew up my capital twice.

7:20
So I don’t think that’s a strategy that’s actually going to work. And I could, I could tell you a strategy that’s—I hate to say basic because I do think there’s some basic strategies that you could benefit from.

7:30
But this is like the price crossing above the 50-day moving average, RSI above the 30, MACD crossing above the signal line. You’re essentially saying buy when the stock’s green and sell when the stock’s red.

7:46
I mean it, it’s really like that juvenile of a trading approach. So using AI, I would chalk this up to garbage.

7:55
One thing that’s not garbage though, is swingtradingthestockmarket.com.

8:00
Go to that website and it’ll take you to all of SharePlanner’s offerings. And one of the ones that I want to highlight for this show is just being able to have access to all of my stock market research each and every day. That’s going to give you my daily watch list.

8:14
That’s going to give you the list of bullish and bear stocks that I’m following each and every week. I also, like I said, I provide a daily watch list each and every day and then on top of that, I do a review of that watch list towards the end of the day. So you’re getting the watch list.

8:29
We’re reviewing the watch list as well to see what went on and what was good, what was bad and what we’re indifferent towards. Plus you’re going to get big tech updates, plus you’re going to get stock market updates as well.

8:35
So it’s a really, really good value worth checking out. Again, SwingTradeInTheStockMarket.com. And in the process you’re, you’re supporting the show. So that was an attempt by me.

8:51
Now, maybe if I did it 15 or 20 more times and got a little bit better with the command prompts, maybe there’s something to the AI. Granted, I did a pretty elementary command prompt as well.

8:58
I didn’t give it a lot of details, but I wanted to kind of give it the liberty to figure out something that maybe I hadn’t even thought—maybe would have thrown in like a Fibonacci retracement or some kind of analysis where maybe it lumps in P/E ratios there or fundamental analysis with technical analysis.

9:18
I wanted to have the room to be creative.

9:23
And it really was just—it felt like it was being very generic, very plain and something that you would see as an example of what a trading system could be.

9:33
So that leads me to my next couple of ideas. When we’re trying to use AI to create a trading system, let’s say in the future, AI gets to a point to where it can.

9:40
And AI right now is probably in the infancy stage of like what we saw the internet being at in like the late ’90s. And so we’re likely to see a lot of crazy advancements in the next 5 to 10 years in AI.

9:56
And with that, it’ll change the financial industry to some regards. But here’s the thing—let’s say somebody comes out with like the ChatGPT of trading and it’s extremely advanced.

10:05
This thing can spit out trading systems with no problem.

10:10
So if it’s spitting out trading systems and they’re profitable trading systems, what’s that going to do?

10:15
It’s going to attract more traders to the financial investing/trading ChatGPT of that time. And then when it does that, more and more people are going to be using it.

10:26
And then more and more people are going to be trading off of these signals that they think that might be unique to them.

10:30
And it may very well be a very unique trading system that it’s not giving out to anybody else, but it may just be variations of that trading system.

10:38
Or if they’re completely unique, it’s going to be very difficult when you consider that all indicators are really derivatives.

10:46
Indicators and oscillators are derivatives of price and volume.

10:52
And so it’s like that expression: there’s only so many ways to skin a cat.

11:02
In this particular example, how many different profitable trading systems is the market going to be able to objectively give you?

11:08
I mean, in theory, you could say there’s probably an infinite number of ways to make money off of the stock market, but if everybody starts working off of the same kinds of signals, one—at some point—it becomes very unprofitable to trade off of AI.

11:19
Because the more crowded anything gets, the less opportunity there is for profits.

11:26
That’s why you have market tops. That’s why you have market bottoms. Because at market tops is when everybody’s in a stock, like what we just saw with NVIDIA. NVIDIA shoots up to 140.

11:34
Everybody and their mom owns it, and there’s nobody else to really keep buying the stock. So what does it do?

11:38
It plummets down below 100 in a matter of weeks. That’s because everybody’s in it.

11:43
When does it bottom? When there’s nobody left to keep selling it—when the selling has been exhausted.

11:49
And so the same goes for AI: people will get into these systems, into this artificial intelligence trading at some point in the future.

11:57
But you get enough people into it and if one person’s saying they’re making money, what are they going to do?

12:01
They’re going to go tell somebody. He’s like, “Hey, you really got to check out this system.”

12:04
It’s working really well for me. So then they go tell their buddies about it.

12:08
And it’s not scalable. It’s not scalable on a mass retail side.

12:13
And that’s why you have like newsletters out there where people are promoting penny stocks.

12:20
It’s like, “Oh, follow me. I’ll make you a ton of money off of penny stocks. You’ll become a millionaire tomorrow.” That’s what people will say. But what is it really?

12:26
It’s really just a system for people to be able to get out of their own positions by pumping that stock to you because those things are so illiquid.

12:33
They’ll put a stock idea out there of a penny stock and everybody piles into it and by the time everybody’s piled into it, the stock starts to drop and then everybody’s piling out of it and it’s causing the stock to create a snowball effect to the downside as well.

12:43
So the mass effect of retail jumping into AI-based trading systems in the future or down the road will essentially make it useless.

12:52
Does that make sense?

12:56
Now I know some people say, “Oh, but it’ll keep evolving, it’ll keep getting better.” And that may be true, but people will still be using it.

13:02
So if it’s just constantly putting out new systems of profitability for people to trade in and that’s that, you got millions upon millions—maybe you even have a billion people—trading off of AI trading systems.

13:12
Are they really going to be profitable at that point? No, I don’t think so. At least it’s also why most people are not profitable in the stock market.

13:23
I would say at best 15 to 20% of people trading in the stock market are profitable. And the way that the 15 to 20% can be profitable in the stock market is for the 80% to really lose, because if everybody was winning, you wouldn’t have any losers.

13:35
You have to have losers in the stock market. And these banks, they need a lot of losers in order for them to be able to liquidate their winning positions.

13:43
I know it’s not possible, but think with me for a second—if Las Vegas could come up with a way to where both the house and the gambler that comes in their doors could be profitable, would they do that?

13:58
Thinking off the cuff, I would probably say no. Maybe at first they would, but at some point they’re going to try to figure out how to get your gains that you might be getting.

14:05
Even though they’re making money too, they’re going to figure out a way to get your gains into their pockets as well.

14:10
That’s how society works. It’s a greed-driven environment. And so the same goes for Wall Street.

14:18
If retail starts becoming really profitable trading these systems, Wall Street’s just not going to sit back and be like, “Oh, that’s great. I’m glad we can all share in the profit-making.” No, they’re going to start trading in such a way to be able to take advantage of the fact that you’re reliant on the AI, and they’re going to take that capital from you because they do not want that.

14:39
Because if you’re making money, that’s money that’s not going into their pockets. Just keep that in mind.

14:44
Feel like I’m kind of getting deep on this podcast episode. Holy cow.

14:48
But in essence, we’re talking about whether or not AI can create a profitable system for the masses and will it become so advanced that it’s able to do so.

14:56
But the likelihood is that all these different systems are just going to be variations of each other.

15:01
And if they’re profitable, so many people will get into it that eventually it won’t be then.

15:10
There’ll be people creating AI systems to outthink the other AI systems. And then when those AI systems are outthinking, those people will be creating more AI systems to outthink the AI systems that are outthinking the AI systems.

15:19
And you’ve got this ripple effect. It’s going to get wilder and wilder.

15:24
So in the end, subjective trading becomes more desirable. Human-based trading that AI can’t duplicate becomes more desirable because then that becomes the edge, because AI can’t do that.

15:42
No matter what they try to say or do, AI cannot be human.

15:53
Now also in this e-mail that Loki writes, he’s asking about, you know, algo trading. And a lot of these algos and the algorithms that these traders are using, they’re usually, you know, scalping pennies off of the bid and the ask price, and they’re making trades in milliseconds just like he alludes to.

16:04
At that point, are you really retired? I mean, if the trading system can work on its own, that’s great, but then you’re checking the trading system, making sure it’s not fouling up.

16:13
I mean, I remember—what was it, like 2010—where you had the fat finger crash. And you always hear about these algos going wild and certain events creating these huge spikes in either direction in the market.

16:23
You don’t want to be one of those people that have an algorithm that’s running, that’s making you money, that all of a sudden fouls up and all of a sudden, you know, you’re buying calls when you should be buying puts, and puts when you should be buying calls, or all of a sudden somebody figured out your system and they’re going to go ahead and take advantage of it.

16:37
I mean that kind of stuff is out there and that stuff happens.

16:42
But I do think that there is a place for code. I’m not saying that there isn’t. And I think, you know, Thinkorswim, you know, you can write your own codes in that. TradingView, you can build your own codes there as well.

16:51
But like I always talk about, you know, trading needs to be based around your lifestyle and around your trading beliefs.

16:57
And if you can incorporate that into your trading system and into your coding, I think you’ll be much better off.

17:03
But sometimes the systems that you create take more work than just being able to do it on your own.

17:11
For instance, one of the popular things out there right now is this ElevenLabs. And it’s pretty spectacular what it can do. It can clone your voice.

17:14
Now, in theory, I could clone my voice and use it on this podcast. And then, you know, I have to write out the script word for word what I want to say.

17:23
And I don’t write out a script word for word when I’m doing these things. I jot down notes and everything on the emails themselves, but I would have to write it down word for word.

17:31
And then if I write it down word for word, then I have to have my cloned voice do the podcast for me.

17:37
And then I have to listen to the podcast voice that was cloned to sound like my voice to make sure that the variation in the voice matches how I would say things.

17:43
If it’s just talking in monotone like this, it’s not going to sound right.

17:52
You guys would be like, that sounds like AI.

17:59
So in the end, it actually just becomes much easier for me, despite the fact that there’s advanced technology out there that can talk without me having to talk.

17:59
It becomes much easier for me just to talk and to just record my voice saying what I want to do.

18:07
It becomes a much simpler approach. And likewise with trading, you can come up with these fancy trading systems and so forth that can trade for you.

18:14
But then you’re having to check the system to make sure that it’s working right and check the parts of the system that’s supposed to be checking the system to make sure that it’s working right.

18:19
And making sure that in the future there’s not going to be some crazy drawdown, even though in the past there wasn’t any drawdowns.

18:27
And then when you start to get into a drawdown that goes beyond what you saw in your backtesting, you’re having to question to yourself, OK, does the system still work?

18:33
You know, has something changed? Has something broke?

18:39
So when you’re trying to do these codings and these trading systems to make life easier for you, oftentimes they can be much harder for you.

18:45
Much like a boat. For instance, here on the Space Coast, everybody’s got a boat, I think except for me. I don’t have a boat because I don’t want to deal with the maintenance that comes with a boat.

18:59
Boats are a lot of work. And what we oftentimes associate boating with is relaxation and island living and tropical paradise.

19:10
But in the end, you’re scrubbing that boat, making sure it’s not getting rusted. You’re replacing parts in the motor, you’re replacing parts on the boat.

19:18
Things are breaking down. You got to store it. It becomes a pain.

19:20
So much so you’re wondering, why did I buy a boat in the first place?

19:26
In my case, I have friends with boats or you join a boat club and they handle all that maintenance for you.

19:31
I think that actually sounds like a great deal too. But in the end, sometimes having a trading system that is based on these very intricate algorithms or, you know, AI-based trading, it can create a lot more problems for you, and it can suck up a lot more time than you would normally expect.

19:51
If you enjoy this podcast episode, I would encourage you to leave me a 5-star review on whatever platform you’re listening to me on.

19:57
Those I really, really do appreciate. And check out SwingTradeInTheStockMarket.com.

20:03
So that you can get all my stock market research each and every day.

20:09
One of the things that people don’t do is write the show enough.

20:09
You know, send me your questions. This was a great question.

20:11
I really enjoy doing this podcast episode, but send me your questions. Tell me your stories.

20:16
Tell me the problems that you face as a trader and let’s make an episode out of it.

20:22
I would love to hear from you guys. ryan@shareplanner.com. I’m the only person that reads these emails.

20:27
Thank you guys, and God bless.

20:27
Thanks for listening to my podcast, Swing Trading the Stock Market. I’d like to encourage you to join me in the SharePlanner Trading Block where I navigate the stock market each day with traders from around the world.

20:34
With your membership, you will get a seven day trial and access to my trading room including alerts via text, e-mail and WhatsApp.

20:42
So go ahead, sign up by going to shareplanner.com/tradingblock.

20:48
That’s www.shareplanner.com/tradingblock and follow me on SharePlanner’s Twitter, Instagram and Facebook where I provide unique market and trading information every day.

21:00
If you have any questions, please feel free to e-mail me at ryan@shareplanner.com.

21:00
All the best to you and I look forward to trading with you soon.