SPY: Overnight Black Swan Waiting Time?

tom

Well-known member
Administrator
SPY_overnight gaps.png

A graphic that displays historical opening gap-downs in the SPY. Data from YAHOO! Finance.

Posted by Ice101781
 
Two things to consider: 1) what does the graph really tell you, and 2) how to deal with large opening gap-downs.

The graph above shows three major gap-downs and the time spans between those. The human mind is very good at extrapolating such data and draw the conclusion that a next episode must be near. But is that really the case? First, the interim periods show no consistency. Second, there is no data shown prior to 1987. Third, and most importantly, what would be the rationale for having a major gap-down at certain periodicity? In other words, I think this graph is suggesting something that is not there. We can discuss the specifics for all three events shown and agree that those conditions are not present currently.

Having said that, there is always the possibility for a major event resulting in a gap-down opening. Thus, it is prudent to be aware of this and have a plan for dealing with it. This could involve always having black swan protection in place (proactive) or an action plan during such an event (reactive). That would be a useful discussion to have.
 
Is it possible to gap down over 5% now with the market circuit breakers?
 
Is it possible to gap down over 5% now with the market circuit breakers?

I think it must be possible since there was a < -5% opening gap that occurred in August, 2015. Looks like the 5% rule only applies to after-hours trading.

From Wikipedia:

"On the New York Stock Exchange (NYSE), one type of trading curb is referred to as a "circuit breaker". These limits were put in place after Black Monday in 1987 in order to reduce market volatility and massive panic sell-offs, giving traders time to reconsider their transactions. The most recently updated amendment of rule 80B went into effect on April 8, 2013, and has three tiers of thresholds that have different protocols for halting trading and closing the markets."

"At the start of each day, the NYSE sets three circuit breaker levels at levels of 7% (Level 1), 13% (Level 2) and 20% (Level 3). These thresholds are the percentage drops in value that the S&P 500 Index would have to suffer in order for a trading halt to occur. Base price levels for which these thresholds will be applied are calculated daily based on the preceding trading day’s closing value of the S&P 500. Depending on the point drop that happens and the time of day when it happens, different actions occur automatically: Level 1 and Level 2 declines result in a 15-minute trading halt unless they occur after 3:25pm, when no trading halts apply. A Level 3 decline results in trading being suspended for the remainder of the day."

"Circuit breakers are also in effect on the Chicago Mercantile Exchange (CME) and all subsidiary exchanges where the same thresholds that the NYSE has are applied to equity index futures trading. However, there is a CME specific price limit that prevents 5% increases and decreases in price during after hours trading."

Posted by Ice101781
 
First, the interim periods show no consistency.

I wouldn't expect them to, otherwise option prices would reflect it or there would be an arbitrage opportunity.

Second, there is no data shown prior to 1987.

SPY only began trading in 1993. I tried to analyze YAHOO! data for the SPX, but the data has different properties. It didn't reflect the ~9% opening gap-down moves that traders experienced in 1987, 2001 and 2008. Does anyone know why that would be the case?

Third, and most importantly, what would be the rationale for having a major gap-down at certain periodicity? In other words, I think this graph is suggesting something that is not there.

How so? No predictions are made. It only reflects what actually happened. Color-coding was used to highlight the disparity in the frequency of events between -2% and -3% and those < -3%. Days where the markets gaps-down < -3% are far more rare. The effect is highly non-linear. In fact, it only happened 11 times since 1993. The vertical bands do suggest a clustering effect for market shocks, but that's well-known. I added them because they hint that opening gaps < -2% may sometimes be 'warning shots', which is a very useful concept, IMHO.

Posted by Ice101781
 
There is a good case for employing Occam's Razor (law of parsimony) here, I think. The graph (and history) SIMPLY shows that bad things happen in the market occasionally, but without any discernible periodicity. In other words, we over-complicate the markets by reading too much into historical data. Can the "end of days" come? Of course, but remember that in the event of a nuclear holocaust money will have no value. Yet, we don't routinely concern ourselves with that possibility. So, personally speaking, I am vigilant but do not have an apocalyptic mindset. I periodically see pamphlets in the Starbucks rest room dealing with the Beast, 666, and quotes from Revelations. But, we are still here. My mindset has always been (check with my former students at SOM): "trade the trade and be ready for big down days".
 
Last edited by a moderator:
How so? No predictions are made.

I'm sorry if I misinterpreted your original posting. Given the title of the thread, I assumed you were implying a major down occurrence is pending. I just tried to indicate that such a conclusion can not be drawn from the graph. I fully agree with Dan's view above.
 
Perhaps we are headed to one. Seems that human emotions might be controlled by something outside of what we think is all that logical. Apparently there has been research on major battles in the world. Every few years the world goes nuts and people go to war. Some call point to the Shmita year and other point to the 8 year cycle. Then there is the Jubilee year.

I used to think we are logical but unfortunately we are can be stimulated and affected just the same. This is not a stock market graph but I am leaving this here for folks to ponder upon. I have always wondered why governments spend so much money on this.

sun.GIF
 
There is not a large enough data set to make any sort of statistical conclusion except what we already know which is that financial markets have fat tails.
 
There is not a large enough data set to make any sort of statistical conclusion except what we already know which is that financial markets have fat tails.

Agreed. We can't make a reasonable prediction of the waiting time until the next ~9%+ black swan due to the limited sample size, but I don't think that means we shouldn't be discussing black swan risk management methodologies. At this moment, based on the three data points mentioned, I think it's prudent to assume these events have at most 14 years between them, even if that's not true.

Many traders on CD seem to have at least 25% of their accounts dedicated to strategies that can withstand a 5% downside move, possibly with some sort of manageable loss. What does that loss look like at ~9%+ though? How much of your account size would you lose if, sometime in the next 5.5 years, you woke up to a market that's gapped-down 10% on some completely unforeseen geopolitical event?

What are people in this community doing today, if anything, to ready themselves for something like that? The purpose of this thread is to motivate that discussion.

Posted by Ice101781
 
So-called "teenies" (i.e. cheap low delta puts) are actually a reasonable hedge at least for BWBs like the RTT. Even if it's not a complete hedge it would still have the effect of significantly cutting your max loss in a crash. Conditional orders to buy a put can do the same thing but what if the market gaps past the trigger price?
 
Conditional orders to buy a put can do the same thing but what if the market gaps past the trigger price?

This. I feel you already need to be net-long puts when the proverbial fan gets hit - it's too late when the market is free-falling. Puts will be very expensive then. One of the things Tom Sosnoff and I agree on.

Posted by Ice101781
 
I don't know about backratios but I seem to recall discussion about a "sea of death" when using put backspreads.
I'd stick with teenies as crash insurance.
 
I know Ron Bertino is a proponent, but in my simulations, I have yet to construct a backratio that offers any meaningful additional benefits over corresponding 'teenies' other than a slightly-higher downside B/E at expiry and a slightly-lower initial debit. Has anyone been able to take advantage of skew when constructing backratios? I welcome any comments on things I might be missing.

Posted by Ice101781
 
Has anyone been able to take advantage of skew when constructing backratios?

There's no getting away from the fact that back ratios mean buying (very expensive) skew ... the only slight skew benefit is that being long the two means you can sell one teeny against it, if the back ratio is close in enough (that is, net you are selling a BWB with a very wide lower leg) ... doing this, you cap the left side of the t+0 but you can cap it at a positive terminal value.

IMO back ratios are best constructed near strikes where your fly lower leg is, roughly and plus or minus ... in this area they can be significantly less expensive than equivalent long option protection, but they need to be managed (rolled, contracted) to avoid issues of sea of death and getting-too-expensive-to-hold. Most traders don't like this idea because effective hedging is very expensive ... you can't just blow a few hundred on unit puts and/or VIX calls and expect to have a good hedge for a 10% gap (30% yes, 10% no) ... rather, you have to get serious and say, for example, "I'm going to blow a full 33% of my expected trade profit on a hedge that will reduce my max loss by 66%, raise my far left p&l to about $0, and make my near t+0 line quite a bit better, but still very painful on a big gap down."

If you prefer to live the lifestyle of always being hedged for the big one, check out Bertino's group ... the ideas are (1) have your income structures far down enough so the hedges necessary to protect them are effective in spite of being very far OTM, and (2) as a way of life, always be blowing a chunk of your potential profits on building up your hedge portfolio ("blowing a chunk of your potential profits" is what Ron calls "getting free hedges"). It's quite impressive to see the hedge portfolios these guys are building up without spending much upfront.
 
Snap1.jpg

There's no getting away from the fact that back ratios mean buying (very expensive) skew ... the only slight skew benefit is that being long the two means you can sell one teeny against it, if the back ratio is close in enough (that is, net you are selling a BWB with a very wide lower leg) ... doing this, you cap the left side of the t+0 but you can cap it at a positive terminal value.

IMO back ratios are best constructed near strikes where your fly lower leg is, roughly and plus or minus ... in this area they can be significantly less expensive than equivalent long option protection, but they need to be managed (rolled, contracted) to avoid issues of sea of death and getting-too-expensive-to-hold. Most traders don't like this idea because effective hedging is very expensive ... you can't just blow a few hundred on unit puts and/or VIX calls and expect to have a good hedge for a 10% gap (30% yes, 10% no) ... rather, you have to get serious and say, for example, "I'm going to blow a full 33% of my expected trade profit on a hedge that will reduce my max loss by 66%, raise my far left p&l to about $0, and make my near t+0 line quite a bit better, but still very painful on a big gap down."

If you prefer to live the lifestyle of always being hedged for the big one, check out Bertino's group ... the ideas are (1) have your income structures far down enough so the hedges necessary to protect them are effective in spite of being very far OTM, and (2) as a way of life, always be blowing a chunk of your potential profits on building up your hedge portfolio ("blowing a chunk of your potential profits" is what Ron calls "getting free hedges"). It's quite impressive to see the hedge portfolios these guys are building up without spending much upfront.

Not necessarily true. Following is chart (from the book "Second leg down") selling 1x2 put ratio on delta neutral basis weekly gross of costs. Average premium cost is 6 bp/week or about 3.1% per year.

Edit: just to clarify nomenclature of the structure for any new visitors of the site. Author says short put ratio, which is same as buying backspread as you mentioned. Short near the money, buy more wings OTM
 
Not necessarily true. Following is chart (from the book "Second leg down") selling 1x2 put ratio on delta neutral basis weekly gross of costs. Average premium cost is 6 bp/week or about 3.1% per year.

That is a suspicious looking chart for a long-vol profile trade. Where is the short strike placed? How many DTE? Does "delta neutral basis weekly gross of costs" mean that you put in on delta neutral, then close out/reposition one week later?
 
Top
Contact Us