How To Generate/Brainstorm Trading Strategy Ideas

On a post I made a few weeks ago, several people asked specifically in the comments about how to come up with ideas for trading strategies. I didn’t see anyone make a post on this topic, so I figured I’d do it myself to share my own thoughts and give an opportunity for experienced folks to share theirs.

My general thoughts:

Instead of “ideas for trading strategies”, I think a more useful framing is “how do we come up with hypotheses for trading strategies?” . A rough hypothesis that can be tested/refined about a potential opportunity in the market. Some sort of vague “I wonder if” statement. “I wonder if there’s a spike in the bid-ask spread on a stock before the volatility increases. maybe I could purchase options to capitalize on that”. “I wonder if this crypto has long, persistent trends I could capture with some kind of moving averages and then trade on it” “I wonder if I can use ___ indicator to tell me when I need to switch from RSI mean version trading to MA-based momentum trading on this asset” etc etc

So, how to come up with hypotheses? For me, there are roughly two parts:

1. Consistently consume diverse, medium/high quality content on the subject

2. Data exploration primarily through data visualization

Part 1:

I would highly recommend consistently consuming some sort of content about trading (not huge amounts but like little intellectual appetizers). Whether it’s blogs (Medium), forums (Reddit), podcasts (Chat with Traders and Better System Trader on Youtube), lectures (Hudson & Thames on Youtube, Ernie Chan lectures) or books (Marcos Lopez de Prado). The diversity here is equally, if not more important, than the quality. In my opinion, Marcos Lopez de Prado’s books are very high quality but those alone won’t just hand you a million dollar trading strat. Consume a wide variety of content to get a variety of perspectives, jot down interesting/fun/appealing ideas, explore and validate them. I say “consistently” because this is an area where the problems we’re solving are very difficult — so it’s likely you’ll need to spend a lot of time thinking about them. If you consistently consume material on this subject, it’ll keep your brain whirring in creative ways so that indeed your shower thoughts are you trying to solve this, even on a subconscious mental level.

This works (for me) because:

1. It keeps me motivated. If I’m excited, I’m going to have better ideas, be more creative and spend more mental time on this without even trying.

2. It provides legitimate mental models/approaches for you to adopt, sometimes

3. You will start synthesizing new and interesting ways of looking at your data when you can draw upon the experience of others. Cool idea here, interesting approach there, didn’t even know that data existed, never thought you could do that etc.

The point here is NOT to try to find a strategy someone else made so you can copy for free. This is a road to nowhere. The point is basically to have context on what people are doing and trying, what range of possibilities exists etc. It’s like… if you’re trying to cook a cool new recipe, reading a bunch of recipes online might be a good starting point to get some ideas/inspiration (note: I am not a professional chef lol). I imagine it would be hard to come up with a great novel recipe if you’ve never read a cookbook, never read anyone else’s recipe, and you just had to come up with something from scratch in a bubble.

Part 2:

For me, the by FAR most effective thing to do this is to combine Method 1 with good data visualization. Your brain is a complex pattern-recognizing machine and if you have SOME kind of vague idea/hypothesis of what to look at (bid-ask spread vs. volatility, moving averages vs. trends, volume-weighted returns vs. length of trend whatever whatever) you should absolutely try to visualize it. Look at charts and plots. Whether it’s price charts with indicators on it, or correlation plots between variables of interest, or anything else, try to find easy/quick ways to visualize the thing you’re interested in and really sit down and just study those charts. Let your brain soak in them for a while. Don’t immediately try to implement a trading strategy, just try to UNDERSTAND the data you’re look at. “Huh, why does volatility go up a lot faster than it comes down?” “Huh, it’s interesting that price responds in ____ way following a large order”. Try to really explore and dig into your data. I believe visually is the best way to do it because any kind of quantification at this stage will leave out too much information (correlation coefficients and other singular values will ALWAYS be less informative at the exploration stage than if you take the time to look at the chart and really absorb the information there).

Side note:

I believe that this data exploration stage is absolutely crucial in quantitative trading and in order to really do this effectively, you have to find a way to make it easy for yourself. It shouldn’t be a 3 day painful process to be able to generate a chart of your variables of interest. Sort out ways to 1) get the data you need and 2) have ways to easily process it so that you can rapidly, dynamically, interactively play with it in different ways to quickly iterate through your hypotheses, see new perspectives and get new ideas.

Once you think you’re onto something, then perhaps it’s time to do some backtesting/tuning/training etc

It’s not a linear process, you’ll be bouncing around a lot and that’s totally fine. But having some ways to draw inspiration, spending time on your own contemplating, spending time studying (visually) charts to understand what does the market feel/look like from a hundred perspectives, that will help you gain a deeper understanding of the possibilities as you start coming up with your own “what if I try…”.

If you’re rule-based oriented, these hypotheses will likely be ideas for trading signals or new ‘rules’. If you’re ML oriented, these hypotheses will likely end up being features to feed your models.

I hope this is helpful, would be curious what reactions and thoughts are, what other people’s approaches are.

You can save up to 100% on a Tradingview subscription with my refer-a-friend link. When you get there, click on the Tradingview icon on the top-left of the page to get to the free plan if that’s what you want.