Blog Post: Success Story 1: Stock Momentum Trading System

Hello all,

A while ago, I promised to share my little success story. I did in fact develop a stock trading system that strictly trades stocks on the NYSE and NASDAQ. (Also, there's no ETF's.)

I finally got around to writing about it. Happy reading!
https://www.vitaltrades.com/2020/07/07/success-story-1-stock-momentum-trading-system/

I'm happy to answer any questions you may have.

Andrew

Thanks & congratulations, AndrewAMD!

I appreciate the focus on simplicity. Hopefully I can follow in your (& jcl's very large) footsteps someday...

One question: Since the system is long-only - are you concerned about a severe, very deep, no-recovery crash, as many are currently predicting?

Thanks.

You're too kind!

Good question. Attached are the 2008 crash (in sample as of 2009) and the coronavirus crash.

I have put in place lots of risk management. Once a certain amount of equity of the overall system is drawn down too much since the last rotation, all of the losing positions are cut. If we are approaching a margin call, the losing positions are cut. The risk management checks for problems on every tick. So I feel pretty good about it.

And how many times were such "end-of-capitalism" predictions correct?

Thank you for the article, Andrew!

10 parameters with Brute force sounds too overfitted Do you plan to re-train it every couple of months?

Thx

Once a year at most. Just so we’re clear, it’s ten total, not per asset.

There are 200+ assets with at least eight years of data being trained. Assuming an average stock lifespan of 7 years out of 8, that at least 1400 years of total daily stock data being analyzed.

"no-recovery" was a bit of "poetic license". Should have been "very loooooooooooooong recovery"...

If by "capitalism" you mean the (free) market, then it's never ended, and can't. That doesn't, however, mean the (real) market doesn't ultimately reassert itself over whatever system has been running, causing tremendous loss, pain, & suffering in some mathematical relationship to how long and how deeply it's been distorted.

But I don't think this is the proper Forum for this kind of Educational discussion!

Thanks again!

Aren't you scared of Bruteforce optimising?
How often is the system retrained?

How many parameters are trained?
The trained parameters are within a small or a wide range?

How do you define "system receives a hit"?

I address these questions here and here.

Executive summary: I'm not concerned because of the tons of stock data being trained against and the type of parameters being trained (mostly broad risk management parameters). Also, I should refer you again to the N=1000 shuffle test, which was performed 100% on out-of-sample data.

About this... I just reviewed my code for the first time in months. Actually, it's only 4 parameters trained. This is why my blog said "10 or less"!

They are:
* Number of assets allowed to trade in the market at a time
* Nominal margin consumed (assuming Leverage = 2)
* Max portfolio allocated to any given asset
* Max portfolio excursion

Blog updated likewise.

Thank you for clarifying!