Line Your Own Pockets

Michael: 00:00

Alright, everyone. Welcome back to another episode of Line Your Own Pockets. Me and Dave have been agreeing a little bit too much, we noticed on recent podcasts. So we grabbed a topic that I think probably I started this one that we're gonna disagree with, I'm sure, a little bit on. So get your popcorn out.

Michael: 00:19

We'll see. So, Dave, what are we chatting about today?

Dave: 00:23

So this is a common thing I see traders ask for. And they're specifically talking about Amber broker usually. So I think RealTest has this feature where And you can tell me, because I don't really use this feature. I'm interested to hear why you think it's so important. I mean, we've talked about this a little bit.

Dave: 00:46

So it's basically a way to compare, to basically portfolio level back testing. You can test a bunch of different strategies and then look at them in aggregate. You can do that in real test, apparently makes it really easy. I'm a broker. There's not a great way to do that.

Dave: 01:06

So some people that go from real test and a broker are a little bit lost because they can't do this. I'm really interested about this because I've never really done that. Or maybe I do it in a certain way, but I don't realize. So I'm interested to hear your take on this because maybe I'm missing something.

Michael: 01:28

Yeah. So first, RealTest not only I would say does it really well. That's the reason for it. That's it's of built from the ground up. And the I got turned on to real test actually from a gentleman again, we should have on Nick Raj, who left Amnibroker after using it for twenty something years solely for this feature.

Michael: 01:50

Right, to run these multiple strategies and multiple portfolios because it's kind of built from the ground up. That's the whole idea. It's built from the ground up so you can run these these multi strategies now. You know, not just to kind of cop out right away. I think the difference we're gonna find is the swing trading or long term investing versus versus day trading thing.

Michael: 02:12

But this to me, I believe personally, like really, really strongly that the the whole holy grail as as if there is one in trading that you could find is multi strategies and more importantly, uncorrelated strategies in in trading. And I think that's a lot of what people are looking for because they're finding a lot of success being able to build two strategies and compare two strategies and understand that they trigger at different times. So there's no buying power conflictions. And then also that they're not just, you know, reinventing the wheel. Right?

Michael: 02:50

They don't have two strategies that have drawdowns that happen at the same time and have profits that happen at the same time and all of that. So to create a smooth equity recur curve, the more multiple strategies. And kind of where this comes from too is is Ray Dalio of of Bridgewater fame, one of the the largest hedge funds in the world. He's been talking about this for twenty or so years is that's the the kind of key to their success is as many uncorrelated strategies or systems on top of each other as they can get. So that's the RealTest side of things, and and that's really what what people are looking for, I think.

Dave: 03:23

Okay. So a couple of things just stick out at me there. Nick Rag, if you moved from I'm a broker to real test, it tells me you're swing trading only. You're not day trading.

Michael: 03:36

And he does some does some day trading.

Dave: 03:40

But not really. He doesn't backtest with it, or he doesn't backtest well with it. That's my first point. My second point is, yeah, we agree on all that stuff, like all the things that we're trying to do on correlated strategies, all that stuff, of course. But my question is, how does this tool help you do that in a way that you can't live without?

Dave: 04:01

That seems, I don't get that part.

Michael: 04:05

Well, the difference I think is when you're testing an Amibroker, you're testing kind of one strategy at a time. And then maybe after the fact, you're trying to create a spreadsheet to merge them. The part of the the issue there is things like buying power. Right? I think that's the first thing that you could touch on, where in real test, because it's it's looking at all of the strategies together, It knows that you've taken maybe all your buying power up in strategy A, and then you can't trigger on strategy B, where if you're if you're building those things separately, you have that problem.

Michael: 04:38

Then also, it's it's a correlational aspect. You know, when you output from from real tests, it gives you here are the correlations of your different strategies when it comes to returns and also when it comes to drawdown. So you have a giant correlation met or matrix on top of it. And then also charting your equity curve. We always talk about how important the equity curve is.

Michael: 05:00

It charts a combined equity curve and says this is the, you know, the strategy combined as opposed to looking at it in its individual parts. So, you know, I'm sure all of this stuff can be done because it's it's done in real test. Yeah. The question is, can it be done that quickly and and simply? Because every time I'm looking to build a new strategy, I am obviously concerned that the strategy is is a good strategy.

Michael: 05:24

Right? You always want that. But I would say I am more concerned with how it fits in the basket of my day trading, investing, swing trading, all of these different time frames, all of these different strategies here.

Dave: 05:37

Okay. Yeah, I think we have fundamentally different starting points for coming up with strategy, I think. So I think that Well, let me back up and ask you how you use the tool in real test. So can you say, Strategy A, I want to trade with this certain size, And then with strategy B, I want to trade with this other size. And you operate those independently.

Dave: 06:11

I assume you can do that. But then, so what I don't get is what's the action you take? Like what's the actual concrete decision you make as a result of this that you build into one of the strategies? Like, what's the output that you say, okay, this is great. This is the improvement I'm gonna make.

Dave: 06:32

What is it?

Michael: 06:34

Well, it will have to do with, you know, so say I build just hypothetically, I build an additional strategy and I plug that into the overall returns. And although the strategy itself is fine, it is very correlated to another strategy that I have. You know, it triggers roughly at the same time. It has roughly the same win rate or loss rate or whatever it is, and it's not adding any sort of strat diversification value to what I have already. Well, in that case, if it's doing the exact same thing on mass as another strategy, I'd just be better off trading that other strategy larger.

Michael: 07:11

Right? If if I'm adding two strategies together and it's not smoothing out the equity curve at all, well, then there's not really much point to do it because all I've done is I've created, you know, more beta and and not really more alpha. So the in a perfect world and the and the easiest thing to look at is a simple trend following strategy for you. Let's just say like an opening range breakout or a multi day high breakout or something like that. And then a mean reversion strategy.

Michael: 07:37

Right? As soon as you combine two strategies together, even if those two are are suboptimal, the the strength really comes together because they're they're making up for each other's weak points. On a on a day that the market may not be trending, right, one strategy is gonna suffer, but then the other strategy really makes up for it. So that's more or less what we're looking for is to say, okay, A, do I have the buying power to do both? Are they are they utilizing buying power at the exact same moment?

Michael: 08:06

And then B, are they diversified enough that it's even worth my time implementing another strategy into my kind of overall trading my systems.

Dave: 08:17

So what if you look at this tool and you find that there's no correlation or there's, like, very little correlation between the strategies, but you don't have enough buying power to trade them like you want. What do you do then?

Michael: 08:30

Well, then that that's the easy part. Right? You just go into the strategy. Maybe you lower the buying power, utilize in one, and you raise the buying power, in the other. You can split your buying power between them, assuming that they have that correlation.

Michael: 08:45

But obviously, if they don't have that correlation, there's no sense to, right? There's no reason to trade the second strategy. It's only when you find that they're uncorrelated in nature. And that's, again, it's the speed is which I can take another strategy and I can test it versus my other one. Otherwise, the only thing I can really rank by is like pure raw performance.

Michael: 09:06

And, you know, by testing, sometimes you get a lot of strategies that are pretty close, right? They're close enough. Then the question is, how do I understand which strategy to run? And is there advantage of splitting it into two?

Dave: 09:23

Okay. Very interesting. Yeah. I think I've got a lot of different thoughts here. The first one is the correlation you would find across strategies like this seems kind of spurious.

Dave: 09:44

It seems to me like I'm thinking about my strategies, and if I found this correlation that I was concerned about, it's just hard for me to believe that something fundamental about these two strategies are correlated in a way that wasn't obvious to me before that I need to take some action on. And that seeing the correlation laid up like that would make me do something different. So the way I think about it is strategy A doesn't care what strategy B is doing. There's no input from strategy B. Strategy A doesn't care about strategy B.

Dave: 10:33

So they're independent things. And if there's any correlation you see, then it seems to me that that would be I wouldn't even create a strategy to begin with if I knew it was going to end up being so correlated with one I already had. Maybe that's how I think about it.

Michael: 10:55

Well, and it may also be, you've got to get your head out of your day trader bias. In in swing trading and in long term investing and things like this, there may just be more aspects, right? Like you're trading a lot of, call them, in play stocks or or stocks that are are unaffected by what the market potentially is doing. When you're swing trading, you're holding things, different market regimes will obviously affect different strategies. So if the market goes nowhere for two years, your trend following strategy is going to suffer.

Michael: 11:32

Right? Unless you have some sort of magic predictive power to say, I think the strategy or, you know, the market's gonna go sideways for two years. You have to if that if you just had a breakout strategy, then your or a trend following strategy, then your performance is gonna suck for that two year period. But if in that time, you also have a mean reversion strategy right, you don't have the crystal ball to say this is the moment in which I should be mean reverting versus whatever, you essentially just trade both as opposed to trying to predict what the market is going to do a year or six months out and and apply different strategies accordingly. You're basically saying, okay, well, I'm gonna try both.

Michael: 12:12

And in periods of of volatile sideways markets, strategy A is gonna really make me money. And then in trend trending markets, strategy B is gonna make me money. So again, I think a lot of this will come down to the styles of trading and the types of trading. Whereas your stuff is just gonna be, right, whether or not a company is opening a FARC coin treasury that day or, right, invented or failed a drug trial or whatever makes these crazy penny stocks do what they do. Right?

Dave: 12:43

Yeah, I guess Yeah, maybe it's going to boil down to the strategies fundamentally because, first of all, if a strategy is going to be in a drawdown for, I don't know, like a couple months, I'm not trading it. I've got other strategies that don't do that. I think that yeah. So Maybe it maybe it does boil down to swing traders are gonna find this a lot more valuable. That that that could very well be the case.

Dave: 13:22

So the other thing so like, I've got

Michael: 13:23

lots Okay. Let's let's look at this. Let's say, for example, and and I'm sure you see this a lot with your your kind of, you know, lower lower cap or penny stock or whatever you want to call it, where there's periods of time where there's just many, many, many trades. And then there's periods of time where there's nothing. Right?

Michael: 13:40

And or even just assume that you're someone who's not really adverse like you are to holding stocks overnight, you think that not having any buying power utilized overnight, you're missing out on opportunities, which, right, I'll I'll go to the grave to say that that you certainly are. You would want to make sure that those moments are are maybe an uncorrelated strategy as well. So maybe you have a strategy that just doesn't it's not that it's bad for a period of time. There's just no action. Right?

Michael: 14:09

So you you're you're shorting stocks that gap up huge. And for for a week or two at a time, there's nothing really there. If you implement another strategy as opposed to manually testing that it's working during those periods of time, And if you had just a system that you could just plug the two together and say, oh, this one actually does pretty well. It does it does nonzero amount of of returns while this other one's doing nothing. Well, that great.

Michael: 14:38

That that's enough to to potentially take a trade off of. So, you know, don't think of it as potentially making up for drawdowns, although that's what a lot of people do. They're just maybe inactive periods of time, and I just wanna see if this strategy plugs into that inactive period of time pretty well in the long run.

Dave: 14:53

Yeah. Okay. So how much time do you spend looking at this?

Michael: 15:03

Well, the beauty of with RealTest, I don't have to spend much time looking at it. Right? I spend most of my time because you're building them all from the same window, right? You're building them all from the same strategy. I'm spending most of my time building the strategy.

Michael: 15:17

And then just quickly, I'll just turn on the other strategies, run the back tests, make sure that, you know, they complement each other in some way. Or at the, you know, the very least what you're looking out for is you're looking out for something that's, again, the buying power and that it's not just like perfectly correlated. Like, I've just built a a second strategy that's pretty much exactly like the first one. Like, one example right now is I'm building some crypto strategies. Right?

Michael: 15:40

I'm working on some crypto strategies for for my website and my product. And the crypto market is very correlated because it's all kind of sentiment driven. Either people love crypto and it's ripping or or or hate it or whatever. And what I found was because of the limited data, I wanna build multiple strategies. I had built two strategies while I was doing this that were both great on paper.

Michael: 16:02

They triggered pretty much the same time. They drew down at pretty much the same time. They right? And it's so, you know, I spend most of my time just building the strategy, and then I I check and I go, okay. Well, if if both of them are very, very correlated, I'm just gonna trade the one that's better and get rid of the other one.

Dave: 16:18

Choose one.

Michael: 16:19

Okay.

Dave: 16:19

Yeah. Alright. So yeah, it's hard for me to think about this world, but it just strikes me lot

Michael: 16:33

of us.

Dave: 16:34

It strikes me as the first thing I think of when I think about this is a term called premature optimization. So there's saying made by a famous computer scientist, Premature optimization is the root of all evil. And for developers and programmers, it means a very specific thing, which is you don't optimize something that might happen in the future, or you can imagine might happen in the future. You end up wasting a lot of time for something that may not even turn out to be a problem at all. And this happens a lot.

Dave: 17:23

In fact, I released the Strategy Cruncher, the payment site for that, I went through a lot of rigmarole because I thought I needed to have the shopping cart on the site. I couldn't just send people to Stripe and make the payment because I thought that there was this very rare situation that was gonna occur potentially that I wanted to account for. So I went through days of making the shopping cart thing work on the site rather than just sending a link to Stripe. Well, it turns out that the situation that I accounted for here has not happened at all. So I did all this work for something that I haven't run into yet.

Dave: 18:11

Like there's been a whole bunch of people that use it and there's no This situation that I spent a lot of time worrying about hasn't happened. So I wasted a lot of time. So it's a very common problem. And I think it strikes me as this sort of thing is premature optimization. So I always step back and think the strategy is the most important thing.

Dave: 18:37

If you get a good strategy and it's got a good equity curve, all those things that you're talking about that you get with this tool, you could push that down the line and fix them as they happen. So it strikes me as something not to focus on before, like, I can't imagine an input I would get from this to my strategy that where I would make any sort of change to it because it just seems like premature optimization. It seems like you can wait. You're optimizing for a problem that may not even happen.

Michael: 19:21

Yeah, I apologize for the dog for a second, but I just it's one of those things. So let's let's do the polar opposite and extreme. Right? Say you have two strategies that are okay, and they don't trigger super frequently. And just like super extreme example, you got strategy A that only triggers on Mondays and strategy B that only triggers on Tuesdays.

Michael: 19:46

You would definitely want to run those. And, you know, that data would be pretty easy to find. But when you're dealing with, you know, limited buying power and especially holding overnight and holding for long periods of time, it may be a lot harder to say, right, I'm backtesting for twenty five years, and, you know, this strategy has a couple months where it really, really shines, maybe around earning season. And this strategy has a period of time where it does really, really well outside of that. To be able just to test that simply, the question is why not?

Michael: 20:20

Right? So if there's a program that's built, which means there can be other programs that are built, then why wouldn't you just test those together? Right? And that's the thing, I think with your thinking this is like a hard task that's going to create a lot of time and effort. And what real test does is you just click one button and they're just tested together, right?

Michael: 20:44

So I think that's where you're getting a lot of these things where, you know, maybe somebody wants to build, again, a really infrequent I don't know, go back to your day trading example. It's just a strategy that triggers somewhat through the day when there's a volume spike. But they find that they're really maxed out in buying power for most of the day. And they just wanna see, is it worthwhile for me to incorporate this other infrequently triggered system even though I'm kind of they're developing their systems to use as much buying power as they can during the day, which they should. You know, the the question I think you're always getting is because there's other programs that you just click one button and it just does that test for you.

Michael: 21:23

Then you go, okay, is there a way to do that in Amni broker where I could just run it together? So as opposed to looking at this as like extremely kind of over optimization, there's obviously, as we know, and as you've already agreed to, there's obviously a lot of benefit of running uncorrelated systems together. And if other softwares can do that super, super easily, why wouldn't we look for a way in additional softwares to to do that as well? Right? Like, so the way I look at it is that I don't have 10 different backtests that I have to keep track of.

Michael: 21:55

And when I'm running orders for trades in RealTest, I don't have 10 different orders that I have to run. It all comes from one page, and I hit orders, and it goes, here's all your orders for tomorrow across your your 10 different strategies at one time. Correct. And I test them all at one time so that the amount of time and energy it just saves to say, you know what, everything's doing okay. Because you know, all of these different strategies are running in parallel, even if one strategy is having a poor time where everything else is doing okay.

Michael: 22:27

Right? So it's not about that this is like, you know, again, I believe it's very important and you agree for the uncorrelated strategies. It's just like, if I can do it over there really easily, right? Why can't we build something for the day trading thing in Ambiena broker that can also do it really easily? Because there's no world where it doesn't make you a better trader to have the ability to test strategies against each other quickly.

Dave: 22:50

Yeah. Yeah. I just I think it boils down to in the swing trading world, a lot of the strategies simply by definition are gonna be pretty correlated or at least have a lot higher correlation than the default day trading strategy. Maybe that's where it comes from because I know that from the ground up, my day trading strategies, there's very little correlation between because they're coming from different ideas. They're not relying on the overall drift of the market to work.

Dave: 23:29

Fundamentally unusual things that are happening that aren't correlated with the overall market. By definition, these are going to be different things. My guess is that when you have strategies where the range of correlation across them is tighter, which I think would have to be true in the swing trading world, maybe this tool is more valuable.

Michael: 23:52

Well, again, it's also it's all I I kind of agree with you, but, right, it also depends. I think you're looking at this from your world, which is like low flow, you know, a crazy mover trading. What if you're somebody who is trading large caps? They're gonna have some sort of inherent intraday correlation to the market, and you just say, well, don't trade large caps. But, like, what you know, there's lots of reasons that people don't want to trade these these kind of fast movers.

Michael: 24:22

Right? There's you can think of a million examples of of why. So even intraday, depending on the securities that you're choosing, there may be, you know, some correlational aspects that that occur there as well. Right?

Dave: 24:38

Well, just a small portion of my trading strategies do low float stuff. So I do all sorts of things intraday, just to be clear. But the more I think about it, the more I think it does come down to the swing trading thing. If you're holding overnight, then by definition, you are relying on the overall drift of the market to help the strategy. So I mean, that's sort of by definition sometimes.

Michael: 25:10

Right? I I have some and this is this is another beauty of the correlational test. I can tell you not only the correlation of all my different strategies, but their correlation to the market itself. And I have some strategies that are inversely or negative correlated. And again, it just comes down to a why not, you know, as part of your testing.

Michael: 25:30

And if there's places out there that do it incredibly easy, I would say, why not? Why not make that part of, you know, if I can click one single button and say, okay, this one's very correlated to the S and P 500, which is okay. And then, you know, maybe it just does better in the long run. But this strategy here, this is inversely correlated to the S and P 500, and it's inversely correlated to these other strategies as well. That becomes a really good idea to split my buying power between.

Michael: 25:57

If they were both very correlated to the S P 500 or using the drift of the S and P 500 for their profit, then you look that as well. That's just another way to compare two strategies. You create one strategy that's just buy and hold, and then you compare your strategies back to that to make sure there is some sort of some sort of edge there and some sort of at least low, if not negative correlation to those, right?

Dave: 26:24

Yeah. So So I think about trying to make my process as smooth as possible. And when you trade a bunch of strategies, that's going to be the important thing. We've talked about before. Beginning traders think they're starving for ideas, but as you get more experience, all of a sudden you turn the spigot on and you've got a whole bunch of ideas.

Dave: 26:49

The problem is you've got too many ideas.

Michael: 26:51

Right.

Dave: 26:51

So you really need to think about your process and getting that streamlined. So to me, to introduce this other potential constraint or decision point for every strategy where I would have to test them against everything, then that's just another potential bottleneck in my process that I would really have to justify, am I getting value from that?

Michael: 27:17

See, I

Dave: 27:18

like something that I'm getting where that that would improve my overall performance. And I I from where I'm sitting, I don't see that. I mean, the other thing

Michael: 27:28

I like it too. Before you get before you get off that point, because I'm looking at it completely the other way around. I'm looking I'm I'm approaching it as the same way as I've got tons of ideas. A very quick and dirty thing I can do to nullify whether or not I should spend time looking at that idea is whether or not it's just it's just going to increase the beta of my other ones. Is it just I if I have a strategy that's a nice so I've got two strategies that I'm I'm I'm playing with right now.

Michael: 27:54

One is just a simple swing trading breakout strategy. Right? It's how a lot of these guys, Quillamari, and then a lot of these guys turned like 10,000 into millions, a lot of these guys do a simple strategy like that. And then I'm pairing that now with a earnings gap strategy. And the idea is right, you're looking for something that gapped up on earnings and that held strongly, and it has this kind of natural drift over a number of days.

Michael: 28:18

If I were to take those two base strategies, and because I can do it in five minutes, like why would I not just take those, pop them together, if they're just really, really correlated in nature, then I can just disregard. Right? Because now I I'm just saying instead of taking this strategy, testing it all the way into fruition, getting it into some sort of paper trade or getting it into a live trade and go, shit. This is just correlated to my breakout strategy. If, right, with you can in real test, take two lines of code.

Michael: 28:51

Right? Copy the thing over, paste it over there, hit a button, and say, okay, these things are correlated or they're not, and then go through and and start doing my testing from there. So I'm looking at the same thing. I look at it as a very fast disqualifier. Right?

Michael: 29:04

I don't want 20 revision to the mean strategies on the market because they're all gonna trigger at the same time. I'm not gonna have enough buying power for all of them. There's there's no sense. So the ability to look to see is this correlated to what I'm doing right now and just use that as a you know, it's it's it's the same thing I'm doing. There's no sense looking at it.

Michael: 29:21

The fact that I can do that in five minutes, I think, is amazing, and I don't know why it wouldn't.

Dave: 29:27

Yeah. Well, if you're if you're swing trading, yeah, I'm starting to see how that would be important to you. I mean, because you've got

Michael: 29:35

to day trading. Right? So say, just use an opening range breakout strategy. Right? If you're gonna incorporate something else that's going to trigger at the same time, well, you only have so much buying power.

Michael: 29:44

Right? Even if you're leveraged to to the tits. Right? You only got so much buying power. So it'd be a very quick and dirty thing to do, which, again, you can do in other softwares.

Michael: 29:54

You just put the two of them together and say, yeah, these trigger around the same time. I'm not interested in in doing this new strategy. I've got these other list of 30 things that I could work on. Now this one's going to get moved to the bottom of the list because I know it's already correlated with the strategy that I have that's working just perfectly fine for me right now.

Dave: 30:12

Yeah. I guess so I guess, again, we're coming down to the difference between swing trading and day trading. And with swing trading, by definition, you're gonna have at most half the buying power that you have intraday. And on top of that, if you've got multiple swing trading strategies where you're holding multiple days, then your buying power dries up so quickly. So yeah, this tool is going to be great.

Dave: 30:43

But intraday, you're not holding overnight, you're holding for a portion of the day. You've got way more buying power, you can cycle your account over and over and over during the day. It's a way more efficient use of buying power. But also the strategy space that you can come up with is by definition not correlated. It would be weird to come up with two intraday trading strategies that were correlated.

Dave: 31:15

I think that's fundamentally what it comes down to. It would be strange if two of my strategies were correlated in some way where I made some decision about one or the other. It just wouldn't make sense to me.

Michael: 31:31

Well, so so buying dips and shorting stocks that go up. Right? Those are those are two things that I think could have correlation to be negative correlation, but they could have that in there. But even though I don't I don't really accept the buying power argument because you're just starting let's assume you have a

Dave: 31:50

one simple math, Michael. Go

Michael: 31:51

on. I'm just $1 less than in okay. Let's start with a $100,000 account. Right? And let's say the swing trader doubles that to a $200,000 account, and the day trader quadruples it because four times to a $400,000 account.

Michael: 32:05

Doesn't matter. Everyone's still limited on buying power. The question is, if you implement all of that 400,000 and you have another strategy that could trigger in, you need to figure out whether you're going to have buying power and on which days you're going to have buying power to see if they overlap. So, you know, you have strategy A that is a gap reversal strategy, and you have strategy B that is a revision to the means strategy, and you want to test both of those. You wanna find out is it worthwhile taking your 400,000 and splitting it into two looking for those particular strategies?

Michael: 32:42

Or do you want to combine them and just do one of them? Right? And does it make sense to split those? Do you get that correlational benefit? And if so, is it worthwhile?

Michael: 32:53

Like actually, like what I think I'm not understanding from you is that, let's say, Amnibroker had something like RealTest in which you could do it in a minute. Why not? I guess is what I'm I'm trying to find out. Like, why and this is the world where you're probably getting people from. You're getting people from worlds where they can test the correlation in their strategies.

Michael: 33:16

Like or or just something like, I wanna trade futures versus I have this one thing that triggers very rarely in futures, and maybe it's triggering at the same time in stocks. I just wanna make sure that these things aren't the case. And in real test, again, it's one button. You hit it. You say, yep.

Michael: 33:32

You you add this much value by combining the strategies. Can I just why not? Like, what's what's the harm?

Dave: 33:40

It's not that there's harm, but it's just, okay, why would you like, what is the what is the decision you would make from looking at the data to that would improve things?

Michael: 33:52

To And to go

Dave: 33:52

on Maybe it's maybe in swing trading, that's definitely the case. But Well, no. It's even in

Michael: 33:57

day trading. Right? Let's go to the $400,000 buying power. The the decision is simple. Do I utilize all 400 buying thousand dollar buying power in strategy strategy a, or do I split between strategy a and strategy b and how much?

Michael: 34:10

So with real tests, you can just run an optimization. Let's say you have three strategies and saying, okay, I'm going to allocate either just a dollar amount to make it simple, or you could do it on on a risk based metric. How much risk should I allocate to each strategy? Because there may be some strategies that trigger frequently that you want to lower risk on and strategies that trigger infrequently that you wanna raise risk on. Well, that becomes an easy, again, one button optimization is just run through these different scenarios where I allocate risk a little bit differently to these different scenarios and, you know, go eat a sandwich and come back and you have a 100 different ways you could allocate risk properly.

Michael: 34:48

And that creates just a super interesting secondary question where, yeah, maybe this particular strategy I'm allocating too much risk to, and this particular strategy I'm allocating too little risk to, instead of having to monitor that in real time, and just click one button, I can do it over, you know, fifty years of data and and back

Dave: 35:07

to So to me, like, I do sort of that by but I just look at my intraday buying power. And when I run out, I go back and analyze that and see, okay, what can I adjust to make that happen? Or to, what strategy, what trade took up the most buying power? How can I make sure that doesn't happen again? What was the actual harm here?

Dave: 35:34

Is this something I should do something about? How likely is this to happen again? Was this some aberration?

Michael: 35:40

And that

Dave: 35:41

way more work

Michael: 35:42

than just clicking

Dave: 35:43

No, big button and because you can't. With the way, like if you said, David, you have to do it the way that you're describing, you couldn't even do it because like how do you know what size you're gonna be trading strategy A at compared to strategy C? I mean, that's gonna Because that comes over time and you get confidence over time. So all these strategies have different buying powers and sizing associated with them, And those are independent from each other. I couldn't even do a test now to predict what it's gonna look like in a year.

Dave: 36:22

That wouldn't even make sense. My approach is, all right, let's go live with these strategies. This one's good enough. I'm not gonna overanalyze it. I wanna get this thing live, start seeing what it does and wait until the bad thing happens, which is something you run out of buying power and then it's okay, what happened?

Dave: 36:44

Let's figure out how to adjust and do that. And that just seems like way more quick than what you're describing. Like you'd have to be able to predict the future a year ahead of time and it just seems impossible.

Michael: 36:55

No. Well, it's the same thing you're always doing with backtest. You're not predicting the future a year. You're saying what happened in the past. You're saying, you know, do these strategies trigger roughly around the same time in the past and being able to hit one button and get all twenty five years worth of data and say, oh, yeah, you know, I should allocate or maybe even dynamically allocate across strategies based off different things that you're seeing in the actual trade setup itself or in the strategy itself.

Michael: 37:24

You know, you say it yourself, backtesting is a superpower. If I could very quickly and easily, that's the thing. Maybe I should sit you down and show you how easy this is to do in real test. If I can just sit there and do it, then, you know, why wouldn't I? That'd crazy not

Dave: 37:38

to. Yes. So I do have a tool that does sort of this. Like, it it can combine equity curves. And the other thing I like about it is it can use different backtesting software as sources for it.

Dave: 37:51

So it can do it across backtesting software. But I'm not doing it It's sort of like a nice to have, or it's not something I regularly look at. So you're wanting to get into day trading. How are you gonna need to do this when you use a different tool other than real testing? And how are you gonna do it?

Dave: 38:13

Like, are you gonna be stuck in the mud there when you have to compare? Like, what do you how are you planning on that? What are you thinking about there?

Michael: 38:19

I'm I'm just planning on on trying to hack it. So, you know, I'm gonna try to avoid it as much as possible in the beginning. And I think that's that's fairly easy to do. Or, you if I come up with an opening range gap strategy, and then maybe I come up with like a closing and balance strategy, like you're going to just know that those are going to trigger at the same time, and they're not going to be correlated, and that's fine. But, you know, let's say I have I developed strategies that hold all day, you know, like, we were both reading Max Maximoyen.

Michael: 38:48

I'm gonna butcher his name. He's doing this strategy. He's doing this test on his substack, which I think is really cool. Just shorting a stock that's up 70% or more and holding until the end of day and just tracking those results. And I I think that's an interesting thing, but he's holding them all day.

Michael: 39:04

So the the question then quickly becomes, you know, is that buying power better utilized somewhere else? So they're right off the bat, I don't see it being a problem, right? Because again, you could do an opening range break strategy, you could probably do something intraday that looks for like unusual volume, you could do some sort of closing, you know, fade strategy. Those are the three that I'm thinking. And those would be fine because I know the buying power would be done.

Michael: 39:30

And for viewers, my plan is, you know, I do my swing trading and I just have more additional intraday buying power. So the idea is that during intraday, you use that buying power that's given to you to do day trading strategies and you go back to a more manageable buying power overnight. So doing that to recycle them, but I know there will come a point where I've got two strategies that I could potentially be running during the day, and I'm going to have to ask that question. So I'm going to wanna see if a strategy works. And, you know, right now, like you mentioned, with real tests, I'm doing some day trading.

Michael: 40:06

I don't think it's like super serious, but it has been profitable for me. And I've done that by combining four strategies and being able to go back twenty five years and seeing that over time, these strategies seem to have a different correlation where each of these strategies, because I'm only using daily bars, I probably wouldn't run on their own. It's just the fact that I can combine the four to do it. So yeah, I could see skirting around it to get started, but I always look at it as an inevitability where I'm gonna have to make a choice. And I would just rather use the power of back testing and be able to go back and get that data than have to just wait for things to occur in real time.

Michael: 40:49

It just seems slow, and it just seems a little little unnecessary for me if if there's tools out there that can do it.

Dave: 40:56

Yeah. So so I mentioned this on my mailing list the other day. Somebody's asking, how can I do this in Avidbroker? And my suggestion was, Hey, this is a great project to improve your usage of ChatGPT. And it would be very simple to come up with a Python script to read these and come up with view your different strategies in a really easy way.

Dave: 41:30

And then you wouldn't be tied to a particular tool. You could do this for real tests. You can combine them and use different sources for it. So I wrote that and it wasn't two days before, I think his name is Matthew, wrote back and said, Hey, I did this project. Here's a GitHub where you can share it with the group.

Dave: 41:49

So by the time listeners are listening to this, I will have already sent the message to the list with this guy's GitHub, with a tool that does exactly what we're talking about across all tools. So it was really cool. I love getting feedback like this from readers. It's just really fantastic. And so, yeah, I think it's a good tool to be able to do.

Dave: 42:14

And I do have some tools in my toolset where I do that and I can, like I said, do it across different back testing tools. So yeah, it's not that I think it's worthless. It's just something that's not quite as important in my routine as yours, I think.

Michael: 42:30

Well, and I would look into that too, right? You know, and this is why I love when we have kind of debates like this because neither of us are ideologues, right? Neither of us are unmovable. But like we always talk about, more data is always going to be a good thing. Like, so if you were to take the time, which would for you would be a lot because a lot of strategies and be able to run them together.

Michael: 42:53

And especially for the listeners and the viewers, you might find holes. You might just look at it and and, you know, grab a so one thing that, again, RealTest does does buying power graph. You know, this is for swing, but maybe intraday, you're noticing that there's a huge lull during some time of day where you're just utilizing no buying power. And just like I always make the argument, I'm gonna bring it back to this Dave, where I think you are leaving a lot of money on the table, not doing any swing trading, even if it's, you know, something that you're buying on the close and selling on the next open, I think not coming up with a strategy is is actually silly for that. Even if you only did 10% of your buying power, if it made a non zero amount of money throughout the year, and you could automate it.

Michael: 43:42

I just I just don't see why not. But same with intraday, you might be able to go through and just say, yeah, there's, you know, during lunch, I don't have any strategies. So as opposed to trying to build the next cool gap strategy, maybe I just look at something that does a non zero return during lunch because I know I'm not trading anything there. And I can just, I can combine these strategies together. And, you know, if you only have one or two strategies, that would be obvious.

Michael: 44:10

But, in backtest over time, I think there's just little holes that you might be able to find to say, yeah, I should be working here instead of here and helping you take that long list of ideas you have and make a little more sense of them.

Dave: 44:27

Yeah, so I do that. I keep track of buying power over time through the tool that I trade with, the TradeClimb. It keeps track of that so I can see. So I'm doing that in real time and out of sample. Yeah, I think we agree on that.

Dave: 44:47

We could probably talk a whole lot of We could probably do several episodes on why I don't swing trade and maybe we'll add that as a topic at some point. Feel pretty strongly about it, but

Michael: 44:59

Well, let's do a couple that we agree on first and then we'll do one that because I that's like a it's like a personal goal for me by by before we end the podcast, which I I hope doesn't happen for for many, many years. It's just to get this one this one strategy that, you know, and start small, like 10% of your buying power. It buys a bunch of stuff, maybe shorts a bunch of stuff on the close, it opens, you know, on the next open. Because like we talked about, eventually, there's not gonna be I guess that's the way that you could win is eventually, there's not gonna be day trading and swing trading when they're twenty four hour markets.

Dave: 45:32

You can

Michael: 45:32

just ride that out, and then you can just call everything day trading at the end of the day.

Dave: 45:38

Yeah. Well, yeah, we've got that bet. We'll see how that plays out. But, yeah, it's just interesting. Yeah.

Dave: 45:43

A little more a little more understanding of of the tool and and why you you know, why it's so important to you. So, I get it.

Michael: 45:52

And then, yeah, and the one last thing I wanted to talk about is, again, multiple timeframes across things, right? So say one worry that I have getting more into day trading is I have a couple revision to the mean strategies that I do for swing. And these are these are ones that I wouldn't classify as they care about the market. They're looking for an extreme condition that you're trying to take advantage of for a very short period of time, which for me is a couple days, like a two week. I could see a world in which a revision to the mean strategy that I do from a day trading point of view is going to overlap with this, but you're just gonna be getting out at the end of the day.

Michael: 46:34

Right? You know, again, think of something gaps down because of really crappy earnings and then tanks the next day. I'm actually looking down at some some strategies that have taken that today. And then I worry, and this one thing that I'm looking at is to say, okay, well, if I end up doing that from a swing trading point of view, and then I end up doing the same thing with the same symbols at the same time from a day trading point of view, I don't know, even though if one strategy is profitable, the other one will likely also be profitable as well. Am I just leveraging up in a single idea and a single strategy?

Michael: 47:08

So that's another thing is just to say, you know, across strategies, right? I have some that I hold for a month, have some that hold for weeks or some that just exit based off some kind of trailing stop and making sure that those timeframes are not just leveraging up on something, again, assuming some sort of catastrophic event where, you know, I lose all my money overnight. That kind of stuff is super important, I think, well to across timeframes. So when I do day trade, I'm going to try to figure a way to take what I'm doing from RealTest and take what I'm doing from Amnubroker and hopefully with maybe this gentleman you were talking about script, combine those and say, what's the output across these different timeframes?

Dave: 47:50

Yeah. Yeah. I think that's a sensed two or three other episodes for you know, topics for other episodes and just what you said right there. So we'll add this list.

Michael: 48:02

Write them down. And yeah. I again, we we talked about before we got started that we love these kind of we want to always sprinkle in one I think where we disagree after just agreeing too much. But let us know, right? We read the comments, read the emails, we read all that.

Michael: 48:21

Know, is it something you do in your particular trading and how important is it to you to do multi strategy testing? Are you more like me where you think it's huge or you're like Dave where it's just kind of an afterthought And and, you know, you've got some different tools for it there. I'll be interested to read.

Dave: 48:40

Yeah. I'm I'm very curious to understand my blind spots if I have any. So feel free to call me out in the comments if you think I've got one.

Michael: 48:50

Awesome. Well, as always, I'm Michael Noss.

Dave: 48:53

And I'm Dave Mabe. Join us next week on Line Your Own Pockets.