VIX systems got crushed

Trading and Markets
21

Most of the drop in XIV was intraday. So any volatility system with a good stop loss would have avoided most of the drop even if it was in XIV.

It’s crazy not to have a stop loss combined with a system that only issues end of day (or beginning of day) signals, for an instrument that could fall 30% or more in one day. Some developers say their backtests work better without stop loss on XIV. However the future is not the same as the past and if your one signal per day system is of by just 1 day the results could be catastrophic without a good stop loss. To assume your system will always avoid the massive down days in the future means you have a overfitted system that works too well only in the past.

A XIV / VXX system should either have a good stop loss strategy or be able to issue signals any time of the day (adapt quickly to the daily situation) to be really robust.

22

We think so too.

[https://collective2.com/details/110186894]

23

[quote=“MaxTor, post:21, topic:10847”]
Some developers say their backtests work better without stop loss on XIV.[/quote]

Indeed, that’s a code phrase for “was too lazy to develop intelligent stoploss logic and don’t care my clients will lose money when the inevitable intraday rout happens”.

Overnight drops in XIV are much harder to avoid and thus are acceptable in a sense.

2 Likes
24

Here is why stop losses generally don’t work too well in long XIV volatility strategies.

Last month (late July 2017), I ran a year’s worth of data on intraday XIV DDs from recent highs.

On 46% of days in the last year, XIV was down 5% from a recent peak. In other words, 5% DDs were so common in XIV that they occurred on almost half of trading days! And the last year was a good year for XIV.

Days including 9% DDs from a recent peak in XIV occurred on 21% of trading days, averaging about once a week. Days containing 14% DDs in XIV occurred about once every 2 weeks. I understand that these DD days tend to be bunched together, but still they are exceedingly frequent.

When I modeled many traditional stop losses (e.g., 5%, 10%, 15%) on XIV last fall, they not only reduced returns substantially, but they INCREASED DDs and INCREASED maximum DDs. They locked in frequent losses, and avoided too many frequent bounce-backs.

I think the better approach is to buy VIX, VXX, or UVXY calls as hedges (my favorites are long-term out-of-the-money VIX calls or long-term out-of-the-money UVXY call spreads . With VIX calls, DDs are still extremely large, but catastrophes, such as the VIX jumping quickly to 40, or 50, or 80, can be guarded against quite effectively. On a jump in the VIX from 10 to 16, however, they don’t provide much protection, and the losses from being long XIV until the indicators change are still very large.

As for holding overnight and over the weekend, that’s when a disproportionate amount of the gains are earned.

3 Likes
25

As bad as the recent downturn has been for most volatility systems, much bigger challenges will come sometime in the next few years.

Back in March, I posted this on the Forum referring to what would happen to vol strategies in a true bear market:

When the hard times come, and they will, there will be a big shakeup here, but IMO the strong returns to most volatility systems will return after some huge drawdowns. My guess is that, except for the few funds that manage the downdrafts well, most volatility funds will lose most of their subscribers when the DDs get large [i.e., -50% or worse]. But the vol. returns earned after the bear market ends will again be huge, and draw lots of new (and returning) subscribers and developers.

So–off and on–I expect vol. systems to be among the high-flyers on C2 for a long time to come–unless the stock market creates more neutral ETfs for volatility than XIV, VXX, and the like.

26

Yeah, a stoploss rule will typically reduce raw CAGR but can massively improve CAGR/MaxDD when re-entry is handled properly. That’s why I said intelligent stoploss logic. My own pure price XIV system (outside C2) is set to 10% trailing stoploss, plus initial position stoploss. Maximum draw down is about 3 times that in bear market before another “system is losing money rule” comes into play shutting it off, truly cataclysmic events (e.g. XIV termination) excepted. Yeah, pretty hefty, but it’s what I’m willing to risk for 100% CAGR. It’s bad compared to best C2 systems but I decided that XIV is serious business and enough to develop my own backtested strategy so I know what it does. I’d rather not get surprised. Currently my system hasn’t gone out of stoploss since last weeks’ drops and is hiding in GLD. Yeah, I’m missing some rebounds.

I have yet to study options, hence my choice for unhedged strategy. Will be looking into that the coming year. Anyway, the reason for my post is to explain my position, but I guess that with options hedging you maybe it’s a different story.

1 Like
27

Long term Vix calls do not provide good short term Protection. This is a very common misunderstanding. These options move like the vix Future prices e.g. the further away your expiration the less Price movement due to recent events.

Regarding stop loss I come to a different result. My underlying stop logic actually increases Profit and decreases drawdown compared to XIV buy&hold. However, I don’t use trailing stop. That’s one of the worst stop logics anyways.

28

I don’t think it’s accurate to call trailing SL “one of the worst logics”, as I’ve seen plenty of claims otherwise. It entirely depends on the rest of your strategy, for some strategies I’ve tried it works great and in others it doesn’t work at all.

1 Like
29

Here’s what happen when you don’t use stops https://collective2.com/details/101429306

4 Likes
30

That´s right, generalization is never accurate. Though I never found classical trailing stops useful in developing strategies. Just personal experience.

31

I think that in the volatility world, strategy without a stop loss it’s like jumping from airplane without a parachute… It may takes some time to crash but it will crush for sure…
I think that stop loss should be a major part in any risk management and lack of stop loss means that there is no risk management…

2 Likes
32

It’s hugely important to have an effective risk management plan when shorting volatility. But stop loss is definitely not the only way to do it. I would argue that options protection is a much more “complete” form of protection. If you’re short volatility overnight and the VIX triples before the market opens, your stop-loss isn’t going to do much good…but with effective options protection, once the options strike price is reached, every dollar lost in the underlying is gained in the option - no matter what time of day the VIX spike happens. Sure, there is a cost for the protection - just like the premium for home owner’s insurance. I wouldn’t own a home without insurance…and I wouldn’t want insurance that only covers my home during market hours. :slight_smile:

At least that’s the way I see it,
David

8 Likes
33

You’ve got a good point there, and it’s important to be prepared for extreme events. Going through historical values though, I didn’t find a case where VIX tripled before market hours. Even in 2007-2008, it is up to 30%-40% increase between close of previous day to open, so a stop loss at that area is probably a reasonable measure too.

1 Like
34

Having 100k capital and possibility of overnight 30k-40k loss with out having chance to do anything, for me is unacceptable. There are more control ways to make money. In case of volatility instruments, hedging with options is the right way in my opinion. Trading volatility people do not use insurance protection because they think they can afford it, but because they think the event not happen for them. And it will happen. Do you know insurance business that does not use re-insurance of potential liabilities?

35

At which strikes do you buy your insurance, marekj? If it’s very near options. there will be no point to trade the system at all due to the high cost.
If it’s far options, then they will only protect against a very large loss - in those scenarios the 30-40 percent you will lose anyway.

2 Likes
36

I guess it’s not a coincidence tripling of vix was mentioned. When you buy
insurance for such extremes cases, it will be cheap. You will still get the
30-40 percent damage to your account or more.

2 Likes
37

Question for DavidJuday, he is the expert in this subject and so far he is doing OK.

On second thougth, the same dilemma has an insurance company buying re-insurance. Possible answer is targeted profitability margin. In case of volatility traders, targeted profit margin is way to high and this eliminates possibility of re-insurance as a high cost burden. Professionals trading volatility do not target 3 digit returns. And for developers on C2, 20-30% target will not provide single subscriber in first 2-3 years of running strategy.

38

I just want to put another important points in here, such as svxy in the past 2 weeks, drop significantly from 92.5 to 75 and from 80 to 67 in one day. This trade is not only deal with loss of significant money but also margin requirements jump significantly 200%-300% n will cut positions automatically without notice. Volatility Returns is the best example from volatility model without protection. David Juday is the best model with protection. Drop 16% is still acceptable in volatility model but drop 50%, that’s not good.

That’s the real situation people have to deal. Since I am a subscriber, I’ll just let a professional trader to figure out how to apply some protection before this event happen by using Options. Another way, you can also hedge with gold or Tbonds but that’s more complex. Most of us as subscribers, we may know in theory what instrument we should use for hedge or protection but thats make our life become difficult. Why don’t choose a model that’s has already proven, just make our life easier when crazy event come without notice.

39

Point of view. As for investor of volatility strategy overnight drop 16% in low correlated strategies portfolio is acceptable, overnight drop of 16% for hedge fund trading volatility is not acceptable. Everything comes to targeted profit and risk model.

1 Like
40

Yes, some numbers here - an example of strikes, premiums - will be useful
for a deeper discussion.
I’m not saying don’t use options, but an understanding of what they
actually protect is important. Apart from that, it is possible to use stop
(at a level it normally doesn’t get) in any case - it is free :slight_smile: