VIX butterfly trade - post-mortem

[Kim]

After 500+ trades we recorded today our first 100% loss.

 

SteadyOptions members know that I place full transparency as one of the top priorities of the service. Unlike many other services that publish only the track record at the best, we list ALL our trades on the performance page, good and bad. So it is only natural that our first 100% loss would require a special topic to analyze what went wrong. 

 

The Trade

 

Members can see the history of this trade here and all the discussions (over 230 posts!) here. Here is the history of the trade:

 

On July 1st, VIX was trading at 16.20 and we purchased the VIX July butterfly at 1.46 debit:

 

Buy to open 1 VIX July 16 15 call
Sell to open 2 VIX July 16 20 call

Buy to open 1 VIX July 16 25 call

 

On July 16 we rolled the butterfly to August for 0.80 debit, and on August 19 we rolled to September 16/20/24 fly for 0.40 debit (less 0.30 credit). It expired worthless today.

 

The thesis

 

The idea behind the trade was to take advantage of any IV spike. Take a look at the VIX chart:

 

 

The entry point is marked in blue. As you can see, VIX went down from 20+ to ~16 and I considered this area low enough. In addition, we also had some VXX puts at that time, and the VIX trade was supposed to partially hedge those puts. Any spike back to 19-20 area would cause the fly to widen to 2.00-2.50 and provide us with very nice gains.

 

What went wrong

 

Well, as we can see from the chart, VIX just kept falling like a rock. It did briefly spike back to the 17 area, but it was not enough. I had a mental stop loss around 30%, but VIX just continues to collapse very quickly and 30% turned to 50-60% loss very quickly and never recovered. The prices moved so quickly that I'm not sure we could get out at 30%. When the loss became 50%+, I didn't see a point to exit.

 

To demonstrate the magnitude of the move, take a look how this trade developed in just one day:

 

Morning July 5 (Friday):

 

 

End of day July 5:

 

 

Morning July 8 (Monday):

 

 

Basically we didn't have a chance to get out at 30% stop loss.

 

So, the first mistake was entering when VIX was around 16. This is just not low enough to give us any margin of safety.

Second mistake was obviously not honoring the stop loss. When VIX went down like a rock, the right thing to do was just exiting for whatever we could get for the trade, but the window of opportunity to exit at 30% loss was very short (about last 15 minutes before the close on Friday, so not sure that all members would even get the alert on time). The reasoning for not exiting was experience of some of our previous VIX trades that were down and recovered nicely. However, the difference was that the VIX was much lower in the previous trades. In hindsight, exiting at 55-60% was the only right decision.

Rolling was obviously a bad decision as well - but again, we did it few times in the previous trades and they recovered nicely.

 

Lessons learned

 

Obviously the main lesson is not to go long when VIX is above 15. In fact, to get a good margin of safety, I would even lower the minimum value to 13-14 area. Following this rule provided us some nice gains this year. Sticking to what was working in the past is usually the smart thing to do. Adding significant capital to a losing trade is also a bad call most of the time. Another lesson is that opening those trades only 2 weeks before expiration is not a good idea. You need to give the trade some time to work out.

 

My advise to members: if you don't understand the trade or think it is too risky, don't take it. To reduce the risk, you can reduce the size or not add to a losing trade if you feel that you are throwing good money after bad.

 

There is not doubt that this was a brutal trade. I feel your pain as I'm in the same shoes. Unlike other services, I trade EVERY single recommendation I send - in fact, all trades come with screenshots of fills from my broker. I also open a trade discussion before the trade is made and explain the thesis and the rationale behind the trade. Again, if you don't understand the rationale or don't agree with it, don't take the trade.

 

To put things in perspective, we did 8 VIX trades before this trade and booked 8 winners. Average winner was around 30% - that's 240% cumulative gain. Another proof that you should stick with what is working.

[thanitp]

I don't know if you agree with this, but I think another key point is you were trading in a directional way (even though you usually don't) for this trade.

 

VIX, more or less, inversely correlates with broad indexes. Therefore, trading butterfly on this is like you indirectly bet the direction of indexes because butterfly trades bet a stock (or whatever) to stay in a given range (if I don't get it wrong, we normally don't consider vega theta etc in VIX because it's already a secondary security). What went wrong was that, after you're on the trade, it's even more important that you have to analyze the directionality of the indexes. I believe this was a reason you kept rolling because you chose not to care about the direction of the broad indexes and that's why you just stayed, hoping to recover.

 

Nowadays, VIX can stay well below 13-14 for quite a long time. Earlier this year we got 12-ish for several days (for VIX spot). Thus, even though you stick with VIX around 13-14, you may still have to roll with some extra capital and gain nothing yet in that month. Do you think this is a good idea?

 

My point is, should we limit ourselves to similar trades in the past like spreads (where you traded the spread or "gap" between two futures contract)? Containing ourselves to what we (well - mainly YOU) do the best?

 

Thank you.

[Kim]

This is a good point, and it is mostly relevant to this trade due to its structure and where VIX was when the trade was initiated. It is less true for our previous trades - even when indexes kept going up, the VIX calendars were losing very little, if at all. 

 

If we initiated the trade around mid July when VIX was around 13, the spike to 16-17 area a month later would be more than enough to provide some nice gains, even with this structure. What would work even better is a structure like calendar or risk reversal that would not lose (or lose very little) even if VIX stays low for extended period of time.

[bam1960]

Kim

Thanks for the post mortem. After the fed anouncement and the vix collapse I would entertain the idea of another fly once market shakes out a little bit. We will get it back.

thanks

Bret

[Kim]

I will evaluate it in the next 1-2 days to see when to enter and which structure.

[Hannes Kury]

Kim,

 

you recorded a 100% loss? For me this trade, which on 7/1 and ended on 9/17, produced almost a 200% loss (comm. included). 

[Kim]

Dear Hannes,

 

I assume you are referring to a loss on the original investment?

 

The P/L is always recorded on the maximum investment, excluding commissions. We increased the investment from 1.46 to 2.50, or ~70%, so in any case, we are nowhere near 200%.

 

This works both directions. In July put calendar http://steadyoptions.com/forum/topic/1207-trades-vix-july-put-calendar-2/?hl=vix we started with $102 margin, increased it to $172 and made a $28 gain. We recorded it as 16.3% gain (28/172) since $172 was our maximum investment. The gain on the original investment would be almost double. Same holds for many other trades - for example, the July RUT IC http://steadyoptions.com/forum/topic/1272-trades-rut-july-2013-iron-condor/?hl=rut.

 

Obviously when the recorded gain is smaller than the gain on the original investment, nobody complains.

[Hannes Kury]

Dear Hannes,

 

I assume you are referring to a loss on the original investment?

 

The P/L is always recorded on the maximum investment, excluding commissions. We increased the investment from 1.46 to 2.50, or ~70%, so in any case, we are nowhere near 200%.

 

This works both directions. In July put calendar http://steadyoptions.com/forum/topic/1207-trades-vix-july-put-calendar-2/?hl=vix we started with $102 margin, increased it to $172 and made a $28 gain. We recorded it as 16.3% gain (28/172) since $172 was our maximum investment. The gain on the original investment would be almost double. Same holds for many other trades - for example, the July RUT IC http://steadyoptions.com/forum/topic/1272-trades-rut-july-2013-iron-condor/?hl=rut.

 

Obviously when the recorded gain is smaller than the gain on the original investment, nobody complains.

Dear Kim,

 

I know you as a smart, honest, straightforward guy. I am not trying to start an argument with a long winding discussion to follow. However, I want to share some of my thoughts about this really bad trade.

 

profit and loss

Fact is that this trade, following the official adjustments, produced a loss of 181% (incl. commission) on my original investment. I understand you are recording P/L on max. investment and without commission. You say it works both ways. I agree with that. However, I feel the math is flawed because it hides real losses/gains. I think the investment in a trade is increased more often when a trade gets in trouble - less often when things go our way. Because of that I tend to believe that we hide more losses than gains. Also, when we roll, like we did in this trade, we basically close the trade and open a new trade. The math based on max. investment does not fully account for the real losses. If one take the individual SO trade performance records and applies them to a certain account balance over a certain period the resulting balance would not match the real numbers of that period. One could actually blow out an account and then  show one 100% loss on a 10% allocation. That's what I think but I am not sure if I am missing something.

the trade

I acknowledge that the post-mortem does point out a number of mistakes. However, it also has a defensive note which lessens the substance of the analysis. I reread some of the trade discussion comments: many members, especially Ice, brought up issues and concerns. Kim, let me be clear, you are a very knowledgeable and intelligent trader with some superb skills. Your productivity absolutely amazes me. That said, the recent AAPL and VIX trades maybe also point out some weakness. So, take this from a friendly admirer: these trades are not just bad they are totally unworthy of you. In both trades you are convinced that it cannot get much worse and we are increasing the amount of capital at risk - that is not a mental stop loss, that is a double down.

You write: """My advise to members: if you don't understand the trade or think it is too risky, don't take it. To reduce the risk, you can reduce the size or not add to a losing trade if you feel that you are throwing good money after bad."""   Well, I did not appreciate that comment in the post-portem of this trade. When asked by members at the beginning of this trade in late June your recommended 10% allocation and you saw the potential loss at 20-25% and the potential gain at 30-50%. That was the plan we did not trade. Again Kim, I think this trade is indefensible. I assume that most members, me included, do not fully understand all the trades, and maybe more important, a lot of us are not good traders (yet). It was not the complexity of the VIX or the trade which lead to the big loss, it was poor trading. In hindsight it is easy to think the entry point was the problem but I disagree: we assumed that the VIX would go up from where we entered - our trade was directional and we were wrong. When we enter a directional trade we don't know where the underlying is going to go - we will always have good/bad entries, our focus needs to be on good trading.

 

I trust that you will probably learn more from this devastating trade that most of us. I hope that you do not misunderstand my comments. I don' want to offend you and I do not expect an extensive response. Just take it as a friendly comment.

 

Thanks

Hannes 

[Kim]

Hannes, your points are well taken and I appreciate your feedback. Lets agree once again that it was a terrible trade, and the fact that it is only one trade out of 500 trades is a poor excuse. Yes, this trade is indefensible - I did not try to defense it, I tried to provide an objective analysis of what happened and my logic behind the decision making process (which was obviously extremely flawed and poor). I did not try to blame the markets, Bernanke, etc. for my bad decisions. The decisions were mine and I take full responsibility.    

 

I don't want to get into that discussion about performance reporting again. I'm not sure I agree with your argument that "the investment in a trade is increased more often when a trade gets in trouble - less often when things go our way". This happened with virtually all condor trades, and we had 5 condor winners out of 7 trades, so in 5 cases out of 7 the return was understated. Reporting return on the original investment would cause much higher returns for most condor trades. The July put calendar is another example. Except for those (and the AAPL trade), I don't recall any other trades where we increased our investment.

 

The only alternative is to report based on the exact number of contracts - since I don't include the number of contracts in the trade alerts, this would be difficult to do. If we did this trade as 3 separate trades, the allocation would actually increase when we opened each new trade since the prices of the rolls were much lower than the original trade.

 

Since the model portfolio is compounded, this trade will actually have much higher impact on the balance that 10% ($1,969 to be precise which is actually even higher than 181%). This is due to the compounding effect that works both ways, and due to the fact that this loss happen after 9 months of accumulated gains. 

[Alturys]

I would follow Kim's logic when it comes to calculating the % return on the investment but I would agree with Hannes that maybe the SO performance doesn't reflect the real impact on his portfolio if you apply that %'age on a 10% allocation if you significantly increased the investment when rolling/adjusting the trade. You should either adjust the size of the trade to bring it back to the original 10% allocation (which might be impossible for the 10k model account sometimes if the original trade was 1 or two lots and reducing allocation would mean to sell half a lot) But maybe the other way to reflect that would take the profit or loss on the increased allocation.

So if the trade grew from 10% allocation to 17% of your portfolio then apply the %'age P/L to 17% rather than to 10%

I still like the idea that I floated before (even though I understand why it won't happen) to use a real IB (or other broker) portfolio and trade all SO trades with a real 10k$ portfolio including all the challenges that a real subscriber with 10k faces. Fees, allocation above/below 10% (as he chooses to trade one lot and 7% allocation or 2 lots and 14%) or an increased allocation as the trade is adjusted.

It would also address the imprecise calculation due to fact that performance is added to or subtracted from the SO portfolio when a trade is closed so if you close a 2 month trade that started when the SO portfolio was at 14,000 it will have a different impact if you close it after a series of winners (portfolio is at 16,000) or losers (portfolio at 12,000) surely works both ways so not sure whether this ends up over or understating SO performance or on balance doesn't make a big difference. Thing is we'll never know.

So a real portfolio would be über transparent address all these issues and show 'real' performance and would stop a number of discussions that come back again and again.....

 

Good points here. But a 10k account with allocation issues (7% for 1 contract/14% for two...) will not reflect the performance of a 50k or 100k account. With a larger size account you have no problem of allocation above/bellow 10%...

 

I think the actual formula is fair as you can extrapolate for bigger account.

 

For the VIX trade, Kim have made a good post mosterm analysis. This trade is an horrible one. The mistakes are obvious, especially the stop one. Even with a 50% loss you must exit if your stop loss was at 30%. The 20% extra loss are a sort of "slippage"... A stop is in place to protect capital...

 

The only suggestion i could make to Kim: In your trade discussion or in the alert, maybe you could indicate a hard stop level. Above this loss, any member should try to salvage the trade. I know this can be brutal and cause some additional loss because some trades could recover. But i don't see other solution to protect capital from disaster. The stop should be large enough to be triggered only a few time per year but could avoid massive loss...

 

By this way you can be sure that members will not miss the information, as the "catastrophic" stop level is known before entering the trade.

 

And if a trade is stopped, we can make the roll after market condition are going better. Without any pressure because expiration or unrealised loss to recover... SO members should never forget that a roll is nothing more than a new trade and should be treated like an independant trade.

 

I hate the "rolling sensation" as you invest more money to a looser who should have been cut earlier. It can give the sensation that you can alway recover a loss wich is not true. And because the first trade make your judgement biaised you will probably miss the proper exit on the roll as you have a "loss to recover".

 

For the VIX Trade, let's say that prefined stop loss is 30%. The market dropped super fast and members exited with a 50% loss. Let's say 500$ per 10k account for this trade. As we are out of market, we could have done the "roll" with vix at 12/13. Probably something like a 13/18/23 butterfly. The new butterfly would have give us some nice gain with the VIX spike to 17. As we ARE NOT in the same trade, we have nothing to recover, no pressure. And we could probably have recovered our initial loss. Maybe with a final profit, maybe not. But we are not blocked in a bad trade. In this scenario we had two trade. A stopped one (-50%) and a good gain (+30 % ?).

 

My two cents...

 

 

 

.

[Kim]

Alturys, those are excellent points and excellent analysis, I agree with ALL of them. My only comment is about hard loss - with options, especially multi-leg positions, I would not recommend hard stops due to slippage. Mental stops should be sufficient. In fact, this is our fist trade after 500+ trades which moved so quickly that we would have hard time to honor the 30% mental stop, but we still had plenty of opportunities to exit at 50%.

 

As for allocation and performance reporting - we are talking about maybe 2% of the trades where we increased our allocation. Usually when we do that, it might be by 20-25%, VIX trade was really unique and not typical. I agree with Marco that 10k account would have some challenges in terms of allocation, but we would not even have this discussion if the model portfolio was 40-50k. 10k was selected randomly, just to demonstrate the account growth, obviously many trades could not have the exact 10% allocation in 10k account since we cannot buy partial contracts. But for tracking purposes, we still need to have some rule and stick to it.  

[Alturys]

Alturys, those are excellent points and excellent analysis, I agree with ALL of them. My only comment is about hard loss - with options, especially multi-leg positions, I would not recommend hard stops due to slippage. Mental stops should be sufficient. In fact, this is our fist trade after 500+ trades which moved so quickly that we would have hard time to honor the 30% mental stop, but we still had plenty of opportunities to exit at 50%.

 

As for allocation and performance reporting - we are talking about maybe 2% of the trades where we increased our allocation. Usually when we do that, it might be by 20-25%, VIX trade was really unique and not typical. I agree with Marco that 10k account would have some challenges in terms of allocation, but we would not even have this discussion if the model portfolio was 40-50k. 10k was selected randomly, just to demonstrate the account growth, obviously many trades could not have the exact 10% allocation in 10k account since we cannot buy partial contracts. But for tracking purposes, we still need to have some rule and stick to it.  

Thanks for your answer Well, when i say "hard stop" i mean:

 

- A rough "catastrophic" stop level known by members before the trade. Maybe you could indicate a target max loss in the trade alert, especially for highly speculative trade... I think we could also measure the maximum adverse mouvement in backtest/archived trade to have good stop who protect us and don't hurt too much our profitability.

- A stop respected... even with a big slippage in case of exeptional move. I don't think i can find a justification to a 50% loss when i would gladly exit with 30%. It is "hope mode", a very bad way to trade...

- I encourage members to take their loss even if SO don't stick to the plan. We are ALL in FULL responsibility for our account as we are the final winner/looser. SO is only a way of learning faster, to have some trade "blueprint", an acces to the forum and to have a lot backtest and research done for us at a very fair price.

 

I known that with complex spread we cannot execute market orders/stop orders at a fair price. But this fact should not be a reason to let a trade going too much against us. At some point, defined before entering into position, we, members, should take our responsibility and salvage the trade without waiting for a signal from SO.

 

I really think this invalidation point should be given before the trade, in the alert, so:

 

- Members have a clear plan with a profit target and a stop. Of course these stops should be large enough, to let market fluctuate.

- You, Kim, will have to respect these stop because they will be "official limits" and we'll be all watching over your shoulder I think it could be a good way to stick with discipline.

- Even if Kim doesn't want to indicate a stop, each member should define a "stop level" for himslef. Idem for the adjustement point, target profit...

[fradav]

Good analysis and discussion.

 

Kim, that you discuss this trade on a forum open to general public, and thereby draw attention to it, and also are capable of admitting that mistake were made, sets your service apart from at least 95% of other services (and traders) out there. Almost nobody talks about their losing trades. 

[bam1960]

That is a fact. Good luck getting this kind of transparency

somewhere else. We can all learn something from this.

Bret

[Kim]

Thank you for your kind words guys.

 

I think the fact that we have an active forum where members can discuss all trades and strategies already sets up apart from most services. We are really trying to be different, and transparency is one of our top priorities. 

[samerh]

Just want to add my 2 cents for what they are worth and also thank Kim for the transparency and everyone else for their comments which read together provide valuable learning.