Check out this video. It's a brave new world.

Transcription:

Hey, Dan Murphy here, and today I want to talk to you about the ultimate pattern recognition system. I have to tell you that stay on for this entire video. Actually, I want to link you to something that's not any of that sales stuff or anything like that. It’s all about learning. That’s why I really want this whole thing to be about. No entertaining facts anything like that in today’s video. I'm just going to get right to it.

The reason why I want to talk about this is really about being an innovator. Innovation. What does it take to be an innovator? What does it take to do something that no one else has really done or even thought of or had the resources to do. My take on the whole subject is that you just have to do it.

For example, there is this story told be before that was talking about you're in this pottery class and the teacher says 100% of the final grade is going to be based on your final project. Then they tell the other group of people in the class that actually you're going to be judged - 90% of your entire grade is based on the weight of the pottery that you fire. Which group do you think would have the best grade? Who would have the nicest stuff at the end of class is what I should say?

You would think that it was the person that focused on just the one project, just making it really nice. No. The group of people that had the best looking pottery were those that actually just did a lot of stuff in order to have a lot of weight of their clay and everything thrown in the kiln. They would have had to actually made stuff right? Just doing things, just getting your hands dirty non-stop is what is going to make you an innovator.

I'll tell you a story about one of my failures here. That would be this. This was my automatic pattern recognition or pattern matching trading system that was a high frequency trading. Interesting concept. I'll tell you what. I'm going to tell you in a couple of minutes what I learned from it. It was a huge ah-ha for me for sure.

You see one of the most interesting things is that when you innovate and you have a certain platform, and then from that platform you can jump into new ideas and so forth. Let me just give you the brief overview of what this was all about.

Basically, it was trading the e-mini S&P 500 since it's the most liquid market that I can get my hands on. My account's not exactly small these days. So I wanted to have lots of liquidly and this is one of the hardest markets to trade. What I decided to do was this crazy brute force method of pattern recognition.

You know I don't give a crap about what they say in the books because a lot of it's going to be things that they’ve done by hand. One of the things I like to tell my clients and so forth is that a lot of people are fooled by randomness, meaning that a lot of people see these patterns.

But you see patterns in clouds. Look up at the sky and see some nice white fluffy clouds. Some of you are getting a vision of Bob Ross. You might remember him being a painter. Look up in the sky, look at the clouds and you will see images. You’ll see faces, you'll see buildings, you'll see all kinds of things, animals looking at the clouds just because our mind makes us want to match patterns, want to see patterns.

So what I did I said whatever the pattern of the market is right now; we could just simplify things by looking at a.... I'll tell you this really quick as well. If you haven't seen volume bars - it's basically forming volume. Let’s say you're using 3200 contracts on the e-mini S&P 500, forms one bar.

That's a whole lot better than looking at the mess that you get in something you get in something like minute charts. I'll show you minute charts here. Basically that same data over that same day. It looks great when you zoom out a little bit.

If you can see on the volume right here, there's times where the market is not doing anything. It’s mostly about in the middle of the day you won't see a lot of action. A lot of the action will come around the open and then around 2:00pm Eastern you'll start seeing a lot of new trends develop. In the middle here you just get a bunch of blah, right?

So one way to counteract that is to form your bars out of volume. Another way people do it is by tick as well. Either way is pretty good. They all have their different pluses and minuses I have to say. But as far as the system goes of pattern matching what I did is I found some discoveries before and I found patterns visually and then I accidently found a pattern and then I traded that for a long time. Very interesting pattern, by the way.

So what this did is it looked at the recent moves in the market and then used heuristics and say, you know what let's check all the past data that we have, all of it, and see how many times this pattern gets picked up - however many moves it has here. Let's see how many times a pattern has existed in the past and how predictive of the future it is. So a lot of heavy thinking went into this.

This was actually a three-month project that I did every single day for sometimes 16 hours. I mean just an absolutely huge one-man show on that one. I didn't have any way of really outsourcing any of that stuff. Very difficult overall.

Also, I wanted to make sure that I wasn't just looking at short time frames. It would also find large time frames as well and try to find patterns in that data. I'll tell you what. This thing was a failure because for one thing I don't think I had enough data. In trade space they only give you six months and I of course worked on getting some more data for that and it just never really came to fruition. I found other projects that I wanted to tackle.

But what this thing did is it was too curve fit, but there's some things that kept popping up over and over. Basically, it would find the same pattern and it's that pattern that I can share with you right now. It's very simple. It’s this. I call it the lightning bolt for no other reason. Basically gets down to this.

The market becomes much more predictable when it's not chopping around like this, but when it’s actually moving big. This move over here, if I want to use the yellow line. Whenever you had extended moves, the market is not as random as it usually is like over here. It becomes much more predictable for buying dips and then selling into rallies just looking for small little profits.

You'll see that a lot of times after these big moves that you can find a good spot to buy the dip and then find a good place to place a limit or and get out of that. So you can go after base hit after base hit after base hit.

That was definitely my biggest failure. The most time I've ever put into any project before or since then because it was like 15,000 buy-ins of code. It was really crazy. Especially because it had to be efficient to do all that pattern recognition, to look back in time and calculate the odds for success based on the past history. Which again, it was too curve fit because I didn't have enough data really to go in there. One pattern would show up about a 100 times over the past six months, and that was not enough data whatsoever.

That's one thing I can really tell you is test as much data as possible. For the e-mini S&P 500 where we can go back to 1997 on that one and have a good idea what the market looked like and so forth. But again that was the really interesting thing I could take away from that was you have that pattern, just look for extended moves, in a late wave they could call it an impulse move. After you see one, the market is much more predictable because it’s not being random at that time.

What I wanted to show you was something really cool and I want to introduce you to something that I've been reading about a lot. It’s basically about the singularity. Some of you might have heard about that before. Just the pace of our technology and the way that I was describing how from a previous platform I came up with the idea for this and then I've come up putting in all this work that I was able to extend this work further.

The only way you can really make these innovations is just by doing. The more you get your hands dirty, the more experience you’re going to get as a trader. So hopefully I've helped you out in determining that direction. Also just stay on for this video - really cool stuff talking about the direction of technology. I find it to be a very serious subject. So here you go.