How to test my new stock trading idea

I am a hobby astronomer and also a beginner hobby stock price observer. I suspect there is pattern in stock price during some planet and sun cycles. I cannot tell you the pattern. It is secret at the moment. I dont need you to mock me for doing "pseudoscience astrology". All I need is methods to prove or disprove my theory patterns. Using scientific methods, numbers will speak for themselves. What are common methods for proving or disproving stock price pattern hypothesis? I know how to do a bit of computer programmation, so I can do automation if necessary.

• You will want to do backtesting: simulate a portfolio using historical data. Doing that multiple times over randomly selected time periods could give you some evidence for whether your method could work. You should compare your returns to another strategy, e.g. by comparing to the S&P 500 index. Most likely, your method will have lower returns and/or higher volatility.
– amon
Commented Nov 12, 2020 at 8:39
• Cut the guy a break with the downvotes. You may think that his premise is nonsensical but it's a sincere question seeking a financial/mathematical solution regarding "stock price pattern hypothesis." Commented Nov 12, 2020 at 13:11
• Sounds like astrology, not astronomy to me. Is that what you meant?
– JohnFx
Commented Nov 14, 2020 at 16:08
• I used to work in solar physics (doing IT support, not as a solar physicist). What you suggest is actually a known correlation, and was used as an example of "correlation is not causation" by my boss, as he would then show how you could take almost any cyclic pattern with a period in a given range and get it to align to the SIDC Sunspot Numbers.
– Joe
Commented Dec 8, 2020 at 15:14
• I agree with @BobBaerker here. The merit of the question should not be based on whether the strategy is reasonable or not. Suppose the question was reworded to say, "My friend thinks there is pattern in stock price during some planet and sun cycles. How would I go about disproving it?" The answers would be the same (except for the one that says to use real money.)
– TTT
Commented Dec 8, 2020 at 15:23

You can test your theory by simulating it.

1. Give yourself a million imaginary dollar (or other preferred currency).
2. See which stocks you should buy right now with that imaginary million according to your system.
3. Create a spreadsheet for the imaginary stock portfolio you would build with your imaginary million (which stocks you would own and how much)
4. Wait and see what happens. When your system tells you to make trades at specific times, update your spreadsheet and your imaginary account balance according to the current stock prices.

After a couple years you should be able to tell if your imaginary portfolio beat the market or not.

(Possible weak points: This simulation assumes that you will be able to trade any stock at any time for its current market price. With popular stocks that's usually the case, but with more obscure stocks it might actually not be possible to buy/sell them whenever you want because nobody wants to sell/buy them. Also, your actions do not influence the stock market. But again, as someone with "just" a million to invest, your influence on most mainstream stocks would be negligible).

When you don't want to wait, you can also do that retroactively by using historic stock prices. Apply your system to the stock prices from 5 years ago, and see how much money you would have if you followed your system until today. But when you do that, then it is important to only use stocks which you did not consider when you formulated your hypothesis. Otherwise you are picking stocks you already know will conform to it, which is a pretty useless method to prove a hypothesis (known as HARKing among academics).

And you should also be aware that backtesting is a much weaker test than prognosticating. Anyone can predict the past. A theory is only useful when it can predict the future.

• If one has access to the historical trigger points then one can backtest large sets of data (across different market cycles) on a variety of stocks and indexes to see if the trading system beat the market or not. If it does beat then one can then work the system in real time with paper trading to hone some trading rules and see if disciplined risk management can improve the results. You can buy/sell more obscure stocks at any time. The problem with them is the size of the bid/ask spread and getting size Commented Nov 12, 2020 at 15:57
• @BobBaerker What are "historical trigger points"? Are they stock market crashes?
– Flux
Commented Dec 8, 2020 at 8:16
• A better description would have been historical buy/sell signals for the system. Commented Dec 8, 2020 at 11:32

Price related patterns can be tested in Excel as well as commercial software. The latter is far more adept at testing a strategy on multiple datasets and much of the programming is already built in.

Either way, if you can write a mathematical formula, both will facilitate strategy performance.

Testing visual patterns generally requires grunt work, looking at one chart at a time.

Welcome new user,

1. Your concept for stock picking ("astrology related") is no more or less whacky than almost all other "schemes" for stock picking.

2. What you're looking for is called "backtesting". It is a completely established field and you can instantly click on a backtesting suite (online, desktop, phone, whatever) and instantly "backtest" your concept. Again, you can instantly buy/find "backtesting" software and online systems, it's completely established. No programming is needed, no more than you would start programming a "text editor". Just click around online and start backtesting. Enjoy!

3. Note however that "backtesting" is psycho voodoo nonsense. It doesn't work, it's utterly silly on at least three different levels, and the fact that people still do it in the face of decades of overwhelming failure says it all about trading "technology".

Also...

"All I need is methods to prove or disprove my theory patterns...."

There is one and only one way to do this.

Take some money (what about \$5,000) and start trading your system. Do that for some years. You'll now have (some) evidence of whether it works.

I caution you that planets are an expression of the behavior of gravity, but markets are an expression of the behavior of consciousness.

. Do note that opinions differ wildly on trading ideas. And only extreme opinions exist.

• You don't need to take real money to try out a system. Just pretending to do so and trading stocks on paper is enough. Commented Nov 12, 2020 at 14:47
• "[backtesting is] utterly silly on at least three different levels..." Could you elaborate? What are these three "levels"?
– Flux
Commented Nov 12, 2020 at 15:29
• @Fattie - "backtesting is psycho voodoo nonsense. I'm going to respectfully disagree with you on that one. The purpose of backtesting is to determine how robust a strategy is across different time periods and securities. It weeds out losing strategies quickly. If one finds one with an edge then the next step is real time, working on evolving trading rules and utilizing disciplined risk management. You don't have to believe me but I've achieved this once and it paid off handsomely in 2008 and 2009. The number crunching in backtesting led to the AHA moment. Commented Nov 12, 2020 at 15:31
• "going to respectfully disagree" - for sure, I'd expect no less! (In terms of my presentation to the OP, the amount of money that has been lost due to backtesting would be on the level of the amount of money lost due to, say, the Black-Scholes fisaco!) Do please note my Final Footnote to our friend the OP ! Commented Nov 12, 2020 at 15:45
• Paper trading offers several benefits if done on a legitimate platform that replicates actual market conditions (buy at ask, sell at bid, with trade execution only if actual trades occur at those prices. It familiarizes you with your trading platform, to learn how to avoid basic mistakes, to see how real time performance is different than looking at close to close results, and to develop and implement a set of trading rules. However, it typically fails because the money isn't real - it lacks emotion. It enables you to stick with losing positions that you would never ride that far down IRL. Commented Nov 12, 2020 at 15:48