Emergency Exit for Weekly System

[MovingAverage]

G'day all,

Hope the markets are treating you well.

Being a person to try and take something positive from a bad experience I've been reviewing how my systems performed during the COVID March sell off to see if there is anything I could do to improve my system performance for similar future events. I live trade several different system but the one that troubled me the most during the March sell off was my weekly system. Because of the rapid downturn in March my weekly system experience some very significant drawdown that didn't sit well with my risk profile. The thing about the March downturn was that it was extremely rapid and deep and XAO has not seen anything like that before (at least over the 25 years of historic data that I use)--even the 2008 sell off was pretty slow compared to the March COVID sell off. Of all of the back testing I've done on my weekly system it had never been subjected to the conditions we experienced back in March of this year because XAO hadn't behaved like this before. So, with this in mind I set about augmenting my weekly system with an "emergency exit" that would dump all open positions if there is a significant and rapid downturn in XAO. To do this may not suit everybody but it suits my personal trading style.

A little about my weekly system: first it is not the Radge WTT but it is vaguely related to that. What is relevant for this post is that my weekly system uses a similar index filter and a similar stop loss strategy. For those not familiar with the WTT the index filter is only used to prevent new positions from being entered if XAO is in a down trend and the stop losses are used to exit open positions when the close drops below the stop loss. I wanted an additional exit of all positions if the market was experiencing a rapid downturn.

The below images represent the current simulation behavior of my system that I have been using since 2015 and live traded through March--AB simulation results.

You can see in the below equity chart that there was a nasty drawdown in early 2020. Mentally this was very tough for me to live trade through.



The drawdown is shown in the below drawdown chart in which the portfolio experience around a 20% drawdown.



And the below chart chart shows the simulation performance of my system since 2015



And finally the below chart shows the Monte Carlo simulation results for my system



So as I mentioned earlier, having live traded my weekly system through March the one aspect of my weekly system I wanted to improve was for it to completely exit all positions in extreme downturns (an emergency exit) like we lived through in March. While I've been experimenting with a number of different emergency exits I binned a lot of the techniques I was working on because they became too complicated and convoluted. I'm a firm believer in KISS when it comes to systems so through some trial and error I landed on an emergency exit that simply sold all open positions if for any given week the close of the XAO index was 4.5% down on the open for the week. After running some simulation...wow, what a difference. The system would now deal with a March downturn in a way that is much better suited to my trading style.

The below images are the performance of my system with the new emergency exit. I'm pretty happy with the improved performance and will include in my live system after I do a few more simulations.

The below equity chart clearly shows that the significant drawdown in early 2020 has been greatly reduced.



The reduced drawdown is shown in the below drawdown chart and the COVID drawdown is only around 8% compared to the above standard system that experienced around a 20% drawdown.



The overall system performance since 2015 is shown below



And finally the overall Monte Carlo simulations are below

 

[over9k]

It's a classic case of taking the stairs up & the lift down.

I mentioned trading daily vs weekly in skate's thread, and he pointed out that post-march, weekly vs daily was very (very) similar.

You were obviously stung in march, but you need to ask yourself whether you expect such an event to be repeated. I'm doubtful.

With that being said, it'd pay to watch the vix like a hawk (i.e daily) even if you don't trade daily.

 

[MovingAverage]

It's a classic case of taking the stairs up & the lift down.

I mentioned trading daily vs weekly in skate's thread, and he pointed out that post-march, weekly vs daily was very (very) similar.

You were obviously stung in march, but you need to ask yourself whether you expect such an event to be repeated. I'm doubtful.

With that being said, it'd pay to watch the vix like a hawk (i.e daily) even if you don't trade daily.

Yes, got hit in March but I've since recovered it all and some so all good. Would I ever expect such an event to be repeated--well who knows because I don't have a crystal ball, but a very wise person once told me "plan for the worst and hope for the best" and that's exactly why I've done what I have. I'm not sure why you're doubtful it will happen again, but adjusting my system on the assumption it will doesn't come at a cost.

 

[CNHTractor]

The reduced drawdown is shown in the below drawdown chart and the COVID drawdown is only around 8% compared to the above standard system that experienced around a 20% drawdown.

Hi, if you can get your weekly system limiting to a 8% drawdown, you are laughing.

my weekly system experience some very significant drawdown that didn't sit well with my risk profile.

The "risk profile" is the key to any trader to understand

From my experience a 20 to 25% drawdown would not be deemed unacceptable [although painful]. I see from your original posting that you had been able to recover from the March Covid drawdown.

I'd be interested in seeing what parameters you have brought into play to assist your system drawdown.

I will watch with interest how you go with further simulations. You may be in a position to test in a live market now as volatility ramps up again.

Good luck

 

[MovingAverage]

Hi, if you can get your weekly system limiting to a 8% drawdown, you are laughing.


I'd be interested in seeing what parameters you have brought into play to assist your system drawdown.


Good luck

Thanks.

In relation to the system drawdown, I haven't adjusted any system parameters. The improved drawdown is only the result of including my emergency exit, which is sell all open positions if a single bar in XAO (keep in mind it is a weekly system) sees a close that is 4.5% down from open. I'll post up my further sim results soon.

 

[MovingAverage]

Hi, if you can get your weekly system limiting to a 8% drawdown, you are laughing.

My emergency exit is far from the holy grail of exits. It is only designed to exit in rapid declines in a short time but will not trigger for slower declines. See below drawdown chart which starts 1/1/2000. You can see it doesn't do much around the slower market declines of 08/09.

 

[CNHTractor]

So, with this in mind I set about augmenting my weekly system with an "emergency exit" that would dump all open positions if there is a significant and rapid downturn in XAO. To do this may not suit everybody but it suits my personal trading style.

This concept brings to mind an article by Marcus Padley "The Collins Class Rule". This article refers to a daily movement in the US market, but is in line with your thinking " corrections start fast and then trend and you had to have an unemotional mechanical strategy that pulled you out early on big market falls "

You are correct that the number of occasions where such an emergency exit would be enacted is quite rare. You have twigged my interest!

 

[MovingAverage]

This concept brings to mind an article by Marcus Padley "The Collins Class Rule". This article refers to a daily movement in the US market, but is in line with your thinking " corrections start fast and then trend and you had to have an unemotional mechanical strategy that pulled you out early on big market falls "

You are correct that the number of occasions where such an emergency exit would be enacted is quite rare. You have twigged my interest!

I'm going to experiment with a few changes to my emergency exit to see if it can mitigate a 08/09 decline so I'll see how I go with that, but my priority is to handle a March style crash. I'll let you know how it goes. Thanks for the link to the Padley article--I'll read it.

 

[over9k]

The other problem you have is that most financial maths really only properly works under normal market conditions, and times certainly aren't normal. Black-scholes-merton and long term capital management getting wiped out in the asian crisis comes to mind.

 

[MovingAverage]

This concept brings to mind an article by Marcus Padley "The Collins Class Rule". This article refers to a daily movement in the US market, but is in line with your thinking " corrections start fast and then trend and you had to have an unemotional mechanical strategy that pulled you out early on big market falls "

Very interesting article, this was most definitely my thinking although I'm doing a complete sell down and not partial.

***

 

[MovingAverage]

The other problem you have is that most financial maths really only properly works under normal market conditions, and times certainly aren't normal. Black-scholes-merton and long term capital management getting wiped out in the asian crisis comes to mind.

I'm not so sure it's a problem with financial maths--not that my system uses any "financial maths". My system is nothing more than a bunch of buy rules and sell rules--no maths. Fundamentally the problem is all about the historical data being used and the fact that my buy and sell rules are based on the behavior of XAO over the past 20 odd years. When XAO behaves in a way it hasn't done in the past 20 odd years (like we saw in March) I have no buy/rules to account for that behavior. To me this is one of the major flaws with back testing systems on historical data. Even walk forward testing on historical data doesn't address this issue. If I could have one wish for system testing it would be that someone like Norgate could provide an artificially created XAO database that models a lot of different and extreme market conditions. Then I could develop a more comprehensive set of buy/sell rules.

 

[frugal.rock]

Probably a good time to be considering the subject matter all in all IMO.
FWIW,
I am at unease with certain world conditions ATM.
VIX levels, gold falling, oil stalling, materials metals mining behaviour (boom), tech boom, banks at lows, world virus levels still peaking, 2nd waves striking, fiscal stimulus measures waning, world GDP well below stock price equivalents, US elections and vested (superpower) interests in detrimental outcome.
It just all adds up to the question, why are we experiencing a "euphoric" bull market?

Personally, I am pulling out of the market trade by trade, not in a great hurry, but a concerted effort to only be about 20% in, at any one time. Just keep the big toe in the waters, so to speak.

On the topic of the emergency exit or GTFO, which is Skates Get The Funds Out filter (clean version), I don't see why weekly traders can't have a more regular check on emergency exit conditions in periods of heightened volatility... such as now.
Cheers.

 

[CNHTractor]

You have twigged my interest!

I have run a test on one of my systems this morning, unfortunately it hasn't added value to my "GTFOH" filter. When I have time I will run it though some of my other systems,

 

[MovingAverage]

I have run a test on one of my systems this morning, unfortunately it hasn't added value to my "GTFOH" filter. When I have time I will run it though some of my other systems,

How does your GTFOH filter work? Just high level, not needing details

 

[quanttraderx]

With my weekly stocks portfolio, I also have a leverage treasury etf strategy running.

I mainly trade the US markets but the concept should work just as well in the Australian market. Unfortunately the ETFs on the ASX are very illiquid.

 

[CNHTractor]

How does your GTFOH filter work? Just high level, not needing details

The GTFOH filter I’ve used on this system operates at 2 levels.

A yellow flag is raised if the Index Gann Trend is negative.

The GTFOH filter ? is then triggered if there is sell signal on the symbol using the Jim Berg codes, whilst the Index yellow flag is waving. The sell trigger irrespective of normal buy/sell conditions.

When there is no index warning flag I do not use the Jim Berg buy/sell signals.

Using the index as a caution has allowed trades to be taken under normal buy/sell conditions if they have continued to trade with momentum.

 

[MovingAverage]

The GTFOH filter I’ve used on this system operates at 2 levels.

A yellow flag ⚠ is raised if the Index Gann Trend is negative.

The GTFOH filter ? is then triggered if there is sell signal on the symbol using the Jim Berg codes, whilst the Index yellow flag is waving. The sell trigger irrespective of normal buy/sell conditions.

When there is no index warning flag I do not use the Jim Berg buy/sell signals.

Using the index as a caution has allowed trades to be taken under normal buy/sell conditions if they have continued to trade with momentum.

Very Interesting

 

[frugal.rock]

Being a person to try and take something positive from a bad experience I've been reviewing how my systems performed during the COVID March sell off to see if there is anything I could do to improve my system performance for similar future events

G'day MA, was there anything in current events to push you to do this? or is it a case of just getting around to it?
I ask due to noting Skate and Cam both have index filters that are off.
That may change on this weekend scans?, but I think not yet.
Cheers

 

[MovingAverage]

G'day MA, was there anything in current events to push you to do this? or is it a case of just getting around to it?
I ask due to noting Skate and Cam both have index filters that are off.
That may change on this weekend scans?, but I think not yet.
Cheers

Howdy, I’ve been working on this for a few months but just a case of getting around to it. It all takes time to revisit my live trades during COVID, do research on the topic and trialing various alternatives in AB. But my weekly has been performing extremely well over past few months so make hey while the sun shines

 

[qldfrog]

Do you check that 4.5% fall daily or weekly.
If you do the check weekly, are you not at the mercy of timing?
And should you not do a daily check similar to:
Exit all if fall on the last n session is above xx %?