Okay, so I built this system by myself over the weekend. Did I show this already? I don't think I've shown this. Right. The West has fallen. Have I shown that already?
Well, the system works now, which is huge. Okay. So, so basically I wrote this impossibly large program. So the dark blue line is Nvidia. Okay. And so let me, let me see if I could do it like this. Hold on. This image viewer is not so great. Hold on. Oh, this is Discord. That's why. All right. So let's just... I don't know what to do here. Hold on one second. There we go. All right. The dark blue line is, uh, Nvidia. Okay. And Nvidia, this is like, um, March 18th, 2020 for Nvidia. Okay. At 2:02 p.m. Got it.
And, um, so you guys are asking lots of questions while I'm trying to show you something and that's pissing me off. So shut your mouths. Thank you. Okay. So, it's March 18th, 2020, 2:02 p.m. and you're looking at the last 30 minutes of Nvidia. Okay. These, do you see these little lines around the Nvidia line? These are... Oh my god. What is this? Ban the sky strategic fractal. I'm in the middle of demonstrating something and you're bothering me with nonsense.
Um, okay. Back to this thing. So, these little lines around the, um, Nvidia... The cam is intentionally in the way. I, I did it to lock it in a certain place. Okay. So, the little lines around Nvidia are something, uh, where I basically... I, I actually don't want to, uh, talk about how I did it, I guess, but I searched through the entire history of the entire stock market. And, um, these are the closest fits to that exact chart pattern. And you can see they match the shape pretty well. So, in the history of all of investing, Halliburton in 2000 had the exact same 30 minutes. HXL... I don't even know what these stocks are, by the way. I know ADI is Analog Devices. And these are the exact same, almost the exact same chart pattern. So, for any stock using my new system, you can type in the chart and the ticker and you get the closest matching stock chart. Okay?
So it's, it's, you know, I'm not going to go into the technical explanations of it. You guys, if, if you're interested at all, you can do it, do that yourself. But the basics are, are that it's the closest matching possible matching. It's over any amount of time you want, intervals. So you could do this over one second, five minutes, five years, whatever you want. You want to say, 'Show me all the stocks that are the same, the same price, uh, the same shape of price.' And you could do this with volume. You could do this with candles. There's, there's any self-similarity or similarity. Okay. So, why do this? One question is, would it predict anything at all? And that's kind of the, the bigger question. Um, and that's why I moved my camera. So, the orange line is what actually happened to Nvidia. So, this is like the way to actually do technical analysis, quote unquote, is called statistical arbitrage. Um, so these are...
This helps you dispense with technical analysis. This helps you prove that TA is... because why? Well, look what actually happened for the next 40 minutes. These stocks had perfect correlation, right? Up until this time point, which I chose arbitrarily. Then right after that time point, even within 10 minutes, they were all divergent. They're all different. 10 minutes, and by 20 minutes, they were even further apart. I know, Epic, you're watching, you're actually watching history being made, and you realize it, but a lot of other people don't. Um, so what's the predictive value? Well, you could see that statistically. You could say, 'Show me within, um, Euclidean distance of x, what is the average expected return?' And it should be zero, right? So when you look at a stock that is, quote unquote, breaking out, you match that pattern with history and say, 'Well, history says there's actually no return here.'
And I think that's going to be something that's really helpful. The searching throughout the market's history is very technically difficult. It requires terabytes of RAM. But this will show you live. Imagine this same chart but live as the chart is being drawn. It shows you the future possibilities based on the past. I wouldn't call it fractal. And, you know, you basically can see like, okay, in 1983 IBM did the same thing, in 1995, you know, Microsoft did the same thing, in 2003 Exxon did that. And, um, you know, that's the kind of, uh, system.
I think TA stopped working in the late '80s, maybe early '90s. So we'll be putting this on Godel. It basically requires a supercomputer or a data center to run. But, um, I think that, you know, using this kind of, um, system can really be, um, powerful because one, it can dispel some myths in your head, but two, if there is real signal you can at least see it, whereas instead of guessing about is there signal, you know, you can see it. And you can just add this to the normal chart, uh, you know, chart library, right? So you can just add it to the chart library. And in fact, like, um, pretty simple.
So imagine you have this. This is what I was thinking for the interface. Okay. So instead of... this is obviously pretty basic, but just getting it to spit this out took me about two weeks. Um, so you could have this be like, I don't know, this. You click this and maybe another box pops out where it says something like n = 100, mean return 0.0001 or whatever, and standard deviation 0.05. So you can see there's no signal there for the top 10,000 matches, and maybe you do the top thousand matches. Same thing, etc. And what's cool is this is real time. So as the next candle draws, all of these will change. This might go up, this might go down, this might go up. These will all change. You know, this will become some other stock that matches it better, etc. And you could adjust your window size. If you want 30 candles, that's fine. If you want 40 candles, etc. And, you know, I think that...
What's important to have in life is data. You know, if you are guessing or hoping to pattern match from your history, your mental history... Yeah, you could do any time interval you want. One second, it wouldn't make much sense, but you could do that. Um, there should be no predictive power theoretically, but there's also... there, there's, there's some mystery to it, too.
I don't, I don't know that there's going to be no predictive power ever, but we're going to keep making tools that'll help you do that. So, one really good example is this is one tool I'm working on right now. Yeah, we could do Kogan integration, of course. So one, one tool I'm doing right now is... there are 390 minutes in the stock market. Okay. So, which minute is the most profitable? That should be a one-click question. Now, I hope you ask better questions than that, right? But it doesn't matter.
Let's say you wanted to know the answer to that question. A lot of my job isn't to help you make money, really. I cannot make money for you. It's to help you make money by giving you software. That's my, that's my day job these days, in case you guys didn't know. And I can give you the news. We have double the news Bloomberg has per ticker. We did an analysis of how much news we have. I have to get my boy Zach to tell me how much. Um, yeah. What's the return by minute of the 390 minutes of the stock market? The U.S. stock market is open for 390 minutes. Which one is the most profitable? Yeah, it should be a one-click button. You should be able to ask what day is the best day to be long. Again, is there predictiveness? Uh, I don't know. No, quantum computers are not useful for this.
You should be able to ask the question: when the moving average 20, you know, EMA crosses 50 EMA, what's the return? You could increase or decrease any N. I mean, you could compare N longitudinally. There's lots of, you know, things to do. It's hard to say, Kenan. I mean, nobody's really sure. Some funds feel one way, other funds feel a different way. But anyway, I'm going to be giving you a lot of these tools in the near future. Yeah, I think it's important, Epic. I mean, I think if, if you like, one example would be for boundaries. So if you think, let's say you're looking at an option. Okay, here's a good example.
Okay, let's say you're looking at an option and these are days instead of... pretend these are days instead of minutes. These are minutes, by the way. But pretend these are days instead of minutes and you're looking at an option and you're here right now and you say, 'Man, I don't know if my option's going to make it, if I'm going to be in the money.' My in-the-money price, let's say you have a call option and you say my in-the-money price... And remember we don't have, in this example, we know the future of what happened at Nvidia. But for the example's sake, we actually have no clue what happened, right? This is why this would be useful, is that we don't know what, what, what was going to happen. The, the orange line doesn't exist. And you say, 'I'm long these calls and I don't think I'm going to make it by expiry.' You know, will, will this be a success by expiry? And you load up all the similar patterns and you see, let's say your call options are struck here.
You say, well, the best of the last five or 50 or 500 observations didn't quite reach that. So, my likelihood of making this is pretty low statistically. And maybe that's useful information. Maybe it's worthless, but at least you can ask the question. And that's the point, is my job isn't to answer the question for you. That's your job. My job is to give you the software to help you answer the question.
But yeah, I think... yeah, you can look at correlations. Easy to calculate this. This is more of a slightly different technique. But anyway, that's something I'm working on amongst many other things. But let's go back to STA.
