Should-I-stop-my-systematic-trading-strategy-when-the-world-goes-crazy

I stopped all strategies, except my VIX strategy with the RSI indicator, which works with Trump even better than the last 10 years.
I hate to say it, but I think you have to be a bit discretionary.

In my experience, discretion will help you avoid some bad trades but also make you miss some big winners. It's likely you won't be able to tell how effective your decisions were over time because no one is inclined to gather that data and do the analyses. I'm not.

I'll let you know when I decide to stop, because that will be the buy signal that everything is going to turn around :D

Obv. I can only speak for myself, but maybe our uncertainty in such times could be diminished with the following thesis.
I believe it is our lack of mastery in creating short trading strategies and finding a good (automated) timing when switching from long to short. If we knew how to handle this better, we could all profit from such "exceptional" and difficult times. We must spend more time developing short-trading-strategies, rather than squeezing out another half percent of an already well working long strategy and integrate them into our trading system. Maybe then, we wouldn't bother stopping our system manually in times like these?

Discretionary trading in such market phases does not work for me, as I'm waiting way to long to get in. once I grasp what really is going on, I'm worried I'm too late to enter and ie. start shorting at the bottom. ;))

So to round it up, once I have my sh*** together I will hopefully be in the team, "It is all backtested, don't interfere with ths system, let it run".

I agree. If there is something that happens technically that would cause you to stop trading a strategy, then surely that "something" can be quantified and detected and backtested.

QUOTE:
If there is something that happens technically that would cause you to stop trading a strategy, then surely that "something" can be quantified and detected and backtested.

I stumpled about two problems with this approach:

1.) Extreme crashes are rare. We need about 100 events to make statistical sense of any dependencies. With one crash every few years there are simply not enough "cases" to develp something meaningful.

2.) Every crash is different. It may be possible to find a mechanism that detects the financial crisis of 2008. But this mechanism surely fails in the Corona crash of 2020. Probably it is possible to find something that works in both scenarios, but I am pretty sure that this "thing" fails in the current tarrif chaos.

It all depends on your strategies. For long-only strategies based on technical analysis, it's really not the best time. But for scalping or mathematical ones like "one percent a week" - why not? This kind of volatility is actually a plus for them. Overall, in my opinion, the market feels like gambling right now. But as mentioned above, by limiting losses, you might also miss out on gains - you never know when that lucky day will come

I was involved in yesterday's discussion. It was very interesting. In my opinion, it is better to continue trading your systems. Ideally, the systems were developed in sample and out of sample over various market phases. This means that they have also ‘seen’ out-of-sample crashes in particular. You know how the systems have reacted to such events in the past. Although one crash is never like another, they are basically similar in some way. Vola rises and it goes down. The recoveries are then fast or slow and so on. In the end it doesn't matter, every exaggeration ends at some point. The problem with discretionary interventions (suspending the system) is that you leave systematic trading. If you decide to shut down systems, then you should also do so systematically. In other words, mechanisms should be developed that precisely define switching off and switching on. Personally, I have not yet found a good switch here. Most things cost a lot of performance and don't reduce the DD very much either, as you generally switch off too late and switch on again too late. Today is a good example of this.

When responding to the question, "Should I stop my carefully designed and backtested systematic trading strategy when the world goes crazy?" I believe many developers create systems—or groups of systems—that are inherently too risky from the start. Every system has blind spots, and eventually, the market will exploit them. Identifying systems and asset classes that can help mitigate these vulnerabilities is crucial for achieving higher risk-adjusted returns. Adopting a multi-timeframe and multi-asset-class approach can further enhance resilience. Many developers harbor unrealistic expectations about returns and the volatility associated with them. If a strategy seems too good to be true, it likely is. Developing a critical perspective—something I lacked as an early developer—is a skill honed over time. Factoring in elements like execution quality, data integrity, and the probability of black swan events can lead to a truly robust set of systems. If you find yourself needing to turn off your system, it may indicate that you haven't built a sufficiently robust set of systems in the first place.