Australian (ASX) Stock Market Forum

Growing pains

Since you don't trade exclusively off charts [I don't either], maybe forming a view on volatility might be of use to you.
 
Thanks everyone for your kind words of encouragement. It can be pretty lonely and scary trading on your own in front of a computer. Your words have helped me so much thus far, and have meant a lot me to. Options trading is definitely not a dinner topic conversation I can have with friends and relatives. I always get dismissed as a gambler or a fool, or (my favourite) a lunatic. My views are always contrary to the majority vote when the topic comes up.

Your replies have spawned so many new points of discussion. I'll reply to each of them in a separate post. I'll start with Grinder.

I prefer to trade as a risk manager, managing my positions as threats to my personal risk comfort zone approaches & making adjustments using the greeks. I know that sounds broad & quite vague but it's pretty much it.
I'm heartened to hear that this approach is plausible. It sounds like an approach that I can attempt.

Like you, I have been terrible at picking direction. In fact, I'm 90% of the time wrong. So, if I acted the opposite to my instincts, I could be quite successful :).

I have followed your old and recent posts with interest and know that your recent upgrade to portfolio margin is a good indication that you are not talking BS. I would love to hear more about your approach in future forum discussions.
 
As a further comment, those bid/ask spreads can be a killer when trading like this. Markets with wider spreads should be avoided or traded in a different way.
Good advice. I'll add that to my personal training notes.

I like your approach of adjusting close to the money ICs and flies. I've followed your trail of clues and used in conjunction with the option bible, I think I'm starting to discover more efficient approaches to adjusting these creatures.

Your gems are camouflaged in such innocent one liners in your posts. All the more satisfying when I discover them upon revisiting your posts. Cheers.
 
I chose to do my options learning in the Oz market due to it being daylight hours, but certainly paid a high price in fees for the privilege.
Aside from the daylight hours, I see positives in trading Aussie options. It is a bit like learning how to drive. If you can pass and survive driving in crazy cities like Bangkok, then driving in sleepy ole Perth will be a breeze. Likewise, surviving Oz will make the US seem like a walk in the park, I would imagine/hope.

... IB only give the first 10 minutes to close the assigned position out before they start randomly closing out positions in the account which may or may not have a thing to do with the assigned position.
Yikes! :eek: Mental note to self: Close out positions before expiry.

Although I have had a bit of T/A training over the years, I find I am really not much better than 50/50 in picking direction. I have found FrankD's ebook quite helpful and using his larger timeframe levels for option adjustments has worked reasonably well especially when they cluster with other T/A such as fib levels, range repeats, volume analysis, etc. Frank goes into a lot more than just his basic levels in his ebook, but personally, I found the basic, longer term levels a useful tool for me. I also watch put/call ratios for extremes together with unusually large turnover. I also often incorporate some basic time analysis - I did start a thread on it some time ago, but haven't had the time to keep it up. Time analysis can be quite labour intensive and doesn't always work.
Thanks for the tips re. T/A. I'll keep the name FrankD in mind when I get a chance to investigate T/A further. Can I ask what 'D' in 'FrankD' is? Thanks.
 
Since you don't trade exclusively off charts [I don't either], maybe forming a view on volatility might be of use to you.
Volatility is a part of the options frontier that I will have to confront. I have been putting it off but at some point, I know that I will need to know more. Right now, I have not reached a point to ask decent questions yet.

With what little I know, I can't figure out how you guys know that a vol rush is upon us. I understand that vol is mean reverting. In other words, after a period of low vol, one can expect vol to increase towards the mean. What I don't understand yet, is, how to determine when that reversion to the mean will occur. How do you guys know that a vol tsunami will be hitting us soon? Couldn't vol pick up in say a year's time?
 
..Yikes! :eek: Mental note to self: Close out positions before expiry.

Not only at expiry - also ITM short calls day before ex-dividend (or other corporate events) can place the position at risk of assignment. And at any time a short position is ITM, there is risk of assignment especially if there is little extrinsic value left.


Thanks for the tips re. T/A. I'll keep the name FrankD in mind when I get a chance to investigate T/A further. Can I ask what 'D' in 'FrankD' is? Thanks.

Frank has a thread here at ASF - https://www.aussiestockforums.com/forums/showthread.php?t=6588. The link to Frank's blog is in his signature - that should answer your questions. That said, T/A is not for everyone and I think deciding where to adjust becomes an individual thing. As Mazza pointed out, not all option traders even look at a chart!
 
What I don't understand yet, is, how to determine when that reversion to the mean will occur.

No one knows for sure, but forecasts/predictions are made, and then the appropriate entry & exits are executed [just like for Px].

w.r.t methodology there are several approaches you would have come across some in McMillan [iv//hv crossover, vol cones]. Also Wayne's blog has good vola discussion, well worth checking it out.
 
Like its biological counterpart, my A.I.D.S appears incurable.

Bungle #5: Fooled By Volatility

I'm at a stage where the significance of volatility is starting to hit home. Honestly, I have been avoiding it like the plague ie. I have been assuming that

1. volatility does not change significantly during the period my position is open.
2. the effect of volatility is negligible on my position.

The truth was that I was not able to cope nor conquer volatility. Volatility had free reign over me and wreaked havoc to my fly. Why do I think that? Let me tell my story.

From my previous episode, I learnt a few lessons from my CBA fly. I came back for more this month, determined to improve upon the previous attempt. Changes I made were:

1. Increase my profit zone.
2. Minimize my number of adjustments.

Plan to increase my profit zone
How should I increase my profit zone a.k.a goal posts? How wide apart should the goal posts be? How about at least 1 standard deviation (sigma) for a weekly period? OK, I'll try that since I don't know any better. 68% probability is something I can stomach.

To determine sigma, I needed the a volatility figure. IV was trading at 27% and HV was shown to be about 24% in my IB TWS option trader platform. To be on the safe side, I'll assume the greater of the two figures and use 27% as my volatility figure. (I now know that IV and HV don't mean sh&t*. Only future volatility matters. But I need a crystal ball for that. Anyone able to sell me one?)

Weekly volatility is derived by
Code:
27%/sqrt(52) = 3.74%.
Convert that to price movement by
Code:
3.74% * $50 = $1.87.
I expect a 1/3 chance that CBA will fluctuate +/-$1.87 from the current price of $50 by the end of the first week.

To be safe, I planned to initiate my new CBA fly to have a profit zone of +/-$2.50.

Plan to minimize my number of adjustments
I will put in practice what I know about dissection and synthetic equivalence to minimize the number of legs to trade per adjustment. In fact I had great success on this front. During my second adjustment, I managed to reduce my transaction count from 4 to 2 to achieve the new position I wanted. (Woohoo! Well done Fox!!! Your're not just a pretty face :)).

Apart from more efficient adjustment methodologies, I believed that my carefully designed profit zone of +/-$2.50 will allow me to sit back and relax for the week. My approach was to:

a. target an average of 1 adjustment a week (and more if necessary) and
b. make adjustments if the price moved greater than 1 sigma for that week

Since I expect 1 sigma events to occur once every three weeks on average, I should be sitting quietly for two of those weeks. Sounds like a fantastic plan to me. I will avoid the mistake of over adjusting.

The body is willing but the mind is week
I know that price movements I calculated were based on expected movements at the end of each weekly period. This would mean that I needed to fill my order for a CBA fly and then go fishing for the rest of the week. Like the scab of a healing wound, I could not resist peeling of the protective covers of my fly. Instead of fishing, I watched nervously as CBA prices darted 1 to 1.5 sigmas during the week. Out of fear, I just could not resist adjusting my positions during those wild sigma swings, only to find that prices whipsaw back to the starting point of $50 after two weeks.

In the mean time, I have locked in a substantial loss of more than 1/2 of my anticipated profit. I feel like an innocent kid spilling milk on the floor and mummy will be bring out the whip really soon. "I really, really, really tried to hold the bottle with two hands, Ma. But the bottle just slipped out of my tiny hands". "That's no excuse Fox junior. You're a stupid, stupid boy! If you spill the rest of the milk, you will have your weet-bix dry tomorrow morning". Fox junior runs off to this room whimpering like a little girl, misunderstood and unloved by mummy.

zzzzzzzz .... Huh! I must have dozed off. Why are you all staring at me like that? Back to option reality again.

The devil is in the detail
How could this be? Should I have stuck to my plan and gone fishing? Would that have helped? In hindsight, yes of course. But were there other more compelling reasons?

Fortunately, I'm in the middle of reading Natenberg. I read a simple test he proposes to verify realised volatility. As described above, I should expect 1 sigma moves 1 week out of 3. I downloaded the weekly prices from yahoo and did a bit of detective work. The weekly price moves using figures from Yahoo are shown below.

-2.57%
4.40%
3.08%
4.12%
1.17%
3.61%
-7.67%
5.62%
3.78%
7.39%
0.66%
5.30%

The numbers in red show that prices moved more than my 1 sigma calculation of 3.74% for 7 out of the last 12 weeks :eek:! Not the 3 out of 12 that I had expected.

I started to have doubts about the volatility figures I adopted. I chose 27% by assuming that the market is intelligent and therefore by voting with their pockets, 27% was a good forecast. I also assumed that the HV from IB was as good as any. I then turned to Hoadley for HV and to my horror, his figure was close to 30%. When I reduced the "window size" to 5, the figures were fluctuating between 20 to 40% in the last two weeks.

So will the real volatility please stand up? I think we will never know. The elusive Miss Volatility has too many multi-personalities, alter egos and proxies that she seems to be everywhere and no where. Too many commentators pretend to tell you that volatility is tangible/concrete. I'm starting to believe that this is not the case. At best, volatility is something you learn to cope and understand through observation. At worst, it is the escape clause for authors when their strategies fail. Instead of saying that you, the mug punter, failed because you did not have faith, the escape clause is that you failed because you could not forecast volatility. The strategy is right but your forecast was just not good enough. He wins, you lose.

I was well and truly fooled by volatility. I'm only just a pretty face after all :(. Time to read and learn more. Back to the drawing board.

Lessons learnt:
1. Volatility forecasting is paramount. Unfortunately, I don't know how to do it (yet).
2. I need to believe that volatility is forecast to fall during the life of my fly. Otherwise, there is no point putting on a fly. The odds are against me, and you can't beat the odds.
3. If volatility forecasting is only a bet, make sure that I can afford the loss when I'm wrong.
 
w.r.t to vol forecasting [I'm assuming with the engineering bg, you've had exposure to statistics and regression] you may find literature on GARCH, ARMA, EMWA models for forecasting vol interesting.

Most of the practical application is not published in academia for competitive edge reasons ;)
 
Thanks Mazza, for throwing me a large juicy bone re. vol forecasting. Things are starting to look brighter already :).

Not the 3 out of 12 that I had expected.
Correction to my previous post ... I meant "4 out of 12".
 
lol, sorry about the open ended reference:eek:
Just wanted to add that its great to see you articulate your endeavours honestly and objectively with great detail, and learning from your experiences.

You've pretty much covered most of the 'gotchas' IMO, so not much extra input here.

Keep up the good work :xyxthumbs
 
Bungle #5: Fooled By Volatility

Fox, being a non-mathematician I've wrestled with this problem for quite some time. i.e. extrapolating sigmas into a probability of directional movement in a set time frame, as you do with flies and condors.

But there is a problem at the heart of the way volatility is calculated which makes it unsuitable for this purpose, viz, the annualized standard deviation of logarithmic daily change in price.

This is what I wrote in a blog post a while ago -

Option volatility takes no account of the trendiness of the underlying instrument however.

Let say instrument (a) moves 1% up and 1% down alternately for 20 days in a row, but instrument (b) moves up 0.5% every day for twenty days.

Instrument (a) finishes very close to where it started, hardly any nett movement at all, yet (b) is up over 10% higher after the 20 days.

According to the option volatility formula, the volatility measured over this time frame is much higher for (a) than for (b).

But depending on your precise option positions (b) could probably *feel* a lot more volatile than (a).

Although the standard practice is for volatility to be annualized (and sometimes extrapolated to "monthized" for we d-neutral traders), in my opinion this is a complete furphy, and a highly dangerous assumption.

This is the reason I abandoned so-called "high probability" condors in favour of flies and low probability condors and adjust/morph as necessary.

I used to take some losses on my high prob condors, but so far have not had a loss on the low prob variety (it will come one day I'm sure). Total risk is also much lower in the low-prob variety in the event of a black swan.

I now only use volatility for pricing and largely ignore it for determining probability of price swings over time periods. It is using a ring spanner to hammer in a nail IMO.

Witness low volatility (as measured) trends that can move one helluva a long way in any number of instruments; the grinding upward move that can go 20% or so as you watch statistical volatility sliding the slippery slope to historic lows.

FWIW
 
I used to take some losses on my high prob condors, but so far have not had a loss on the low prob variety (it will come one day I'm sure). Total risk is also much lower in the low-prob variety in the event of a black swan.
Wayne, I really appreciate your insights into low prob flies/ICs. The "hands on" experience that you are sharing will never make it to option text books and someone like me will have a slim chance of knowing them. Your positive experiences with these strategies gives me confidence that they are viable.

I'm currently half way through Natenberg. His view is biased towards the theoretical ie. unless you identify market discrepancies in option pricing, you will not have a mathematical "edge". That idea does appeal to me, as I like to take comfort that there is a mathematical basis to being profitable in this game, much like card counting is to blackjack.

On the other hand, I also like to believe that option theory has its loopholes. Your take on volatilities is one such example. Comments from Grinder and Sails also indicate this to be the case. They talk about managing the greeks as being their "edge". I would like to prove to myself that this management "edge" is real. In my case, only time and practice will tell.

Whether the "edge" is mathematical, management (human) or due to deficiencies in option theory (or perhaps all 3), I'm just glad that there are real players out there, who are able to profit from them.

Witness low volatility (as measured) trends that can move one helluva a long way in any number of instruments; the grinding upward move that can go 20% or so as you watch statistical volatility sliding the slippery slope to historic lows.
Being a non-mathematician myself as well, my humble opinion is that the problem stems from the fact that sigma is derived from magnitude of price movements. Direction of movements are not factored in. Therefore up,up,up,up vs. up,down,up,down movements are (incorrectly) given equal weight when determining sigma. Much like EWMA is better modelled with appropriate weights, as compared with standard HV calculations. (Thanks Mazza, for pointing that out)
 
But there is a problem at the heart of the way volatility is calculated which makes it unsuitable for this purpose, viz, the annualized standard deviation of logarithmic daily change in price.

As above, daily frequency of close-to-close hv, causes sampling errors [small sample] or irrelevant info incl. [large sample].
Improvements via high frequency data [expensive for retail] or estimates that incl. combos of OHLC e.g. Parkinson, Garman Klass

Being a non-mathematician myself as well, my humble opinion is that the problem stems from the fact that sigma is derived from magnitude of price movements. Direction of movements are not factored in. Therefore up,up,up,up vs. up,down,up,down movements are (incorrectly) given equal weight when determining sigma

Try profiling price 'behaviour' by quoting price changes in sigma units. E.g. XYZ & ABC stat vol =50%.
Most price changes [up/down] of XYZ [=/<]2 sigmas. ABC price changes are primarily [=/<] 1 sigma, but monthly has moves > 4 sigma.

Range plays are > suitable for XYZ than ABC [appears better to be long gamma/vega] despite annualized stat vol of both being 50%. [Independent of other analysis]
 
I'm currently half way through Natenberg. His view is biased towards the theoretical ie. unless you identify market discrepancies in option pricing, you will not have a mathematical "edge". That idea does appeal to me, as I like to take comfort that there is a mathematical basis to being profitable in this game, much like card counting is to blackjack.

On the other hand, I also like to believe that option theory has its loopholes. Your take on volatilities is one such example. Comments from Grinder and Sails also indicate this to be the case. They talk about managing the greeks as being their "edge". I would like to prove to myself that this management "edge" is real. In my case, only time and practice will tell.

Whether the "edge" is mathematical, management (human) or due to deficiencies in option theory (or perhaps all 3), I'm just glad that there are real players out there, who are able to profit from them.

I figured awhile back (after taking some hefty losses) that you need to play to your strengths not your weaknesses if you want to survive, and that the math has to make enough sense to justify your postioning for say positive expectancy or understanding how the greeks represent in your portfolio but should'nt (at least for me) be the basis of all the decision making. I'm sure theres quants (mazza) that can get it done on the math alone, that aint me. So I do what works for me & thats strictly managing my low prob ICs & letting theta grind out the profits.
 
Try profiling price 'behaviour' by quoting price changes in sigma units.
Mazza, Your idea of profiling (see above) lead me to think that comparing

a. HV calculated using daily gains VS
b. HV calculated using weekly gains

may lead to similar conclusions. I experimented with real data from CBA and BHP which are now both showing HV of 23%, but are different in nature. The results are are below and they are startling. That's why you're the quant, and I'm not.:bowdown: Brilliant!:thankyou:
Code:
[FONT="System"]
Underlying          BHP      CBA
Daily(30)            23.0     23.4
Weekly(10)         23.3     30.7
[/FONT]

... unless I'm jumping to conclusions too early :eek:, in which case, shoot me (compassionately).
 

Attachments

  • hv.jpg
    hv.jpg
    55.1 KB · Views: 68
As above, daily frequency of close-to-close hv, causes sampling errors [small sample] or irrelevant info incl. [large sample].
Improvements via high frequency data [expensive for retail] or estimates that incl. combos of OHLC e.g. Parkinson, Garman Klass

Try profiling price 'behaviour' by quoting price changes in sigma units. E.g. XYZ & ABC stat vol =50%.
Most price changes [up/down] of XYZ [=/<]2 sigmas. ABC price changes are primarily [=/<] 1 sigma, but monthly has moves > 4 sigma.

Range plays are > suitable for XYZ than ABC [appears better to be long gamma/vega] despite annualized stat vol of both being 50%. [Independent of other analysis]

I had to read that three times mazza.

Once converted into wayneL language, yes - makes perfect sense and something I've actually been trying to quantify in "retail trader style" for some time.

It's given me a couple of ideas for a retail muppet's indicator, thanks.
 
Yeh, the math is only to quantify ideas for analysis. Models are like any other tool, you still need market experience & mgmt. That's why I like learning and listening to more experienced traders; Cottle, Wayne, sails, boss etc.

a. HV calculated using daily gains VS
b. HV calculated using weekly gains

Although not exact, this is conceptually a variance ratio test of random walk [trend/mean reversion indicator].
IMHO it doesn't adequately describe the 'behaviour' of the underlying. You'd like to know the range and frequency of its stdev moves.

I'd add using different time composites to account for trends.
E.g. XYZ daily moves [=/<] 2 sigmas - sounds great for range play. But, also assume XYZ often moves [=/<] 2 sigmas for 3 consecutive days, this is a 6 sigma move - killing any IC, lol.
IOW, profile x-day price changes as well.

edit: sorry guys about the writing. I tend to be too verbose or too concise :eek:
 
Postscript to Bungle #5 Fooled by Volatility

After giving much thought to all the comments from Mazza and Wayne, I think I have traced the source of my mistake. The mistake I made was to equate an annualised volatility with the annual volatility itself. The former is actually the daily volatility measure expressed in an annualised form. The latter is the volatility measure of annual price movements. They are NOT the same thing and there is grave danger in interchanging both terms. Well, if you have an option position, let me assure you that you will feel your error :).

As pointed out in Mazza's last post, the numbers are used to quantify ideas. They do not reveal the complete and true nature of those ideas.

Let's turn to nature to illustrate my bungle. Join me on a leisurely walk down to Cottlesloe beach and consider a ripple, a wave and the tide. The ripple rides on the wave. The wave rides on the tide. The tide goes up and down each day like clockwork.

I'll pretend to be an oceanographer and measure the volatility of the ripple by recording the height of a ripple. I would say, it is about a centimeter. Next I measure the volatility of the wave and it is about 1 meter. I'm too lazy to use the word "centimeter" because it has 4 syllables, I decide to use the term "meter" instead. So I say that the ripple's volatility is one hundredth of a meter. But being the lazy bugger, "one hundredth of a" is a mouthful. I start calling it a wave equivalent or "wavethised" 1 meter. "1 meter" - that's nice and easy to describe volatility of a ripple.

I get all excited and start reporting my findings to my fellow geeks at the local amateur oceanographer club. Soon, I convert everyone to be lazy buggers and we all agree to refer to ripple volatility measurements in "wavethised" units of X meters. We all had a jolly good time measuring ripples, waves and tides in "wavethised" units, proudly displaying and interpreting our charts.

Notice that all I measured was the height of the ripple. Being an amateur, I did not know how to quantify the ripple's beauty, its colour, its foaminess, it's roundedness etc. So the ripple has now been diminished to be one dimensional in nature ie. short or tall. Sorry ripple, I have wronged you.

A few years passed and a new member (Baby Fox) joined our amateur oceanographer club. Baby Fox listened with wide eyed wonder about our ripple and wave measurements. A few scientists among our members suggested that we can estimate wave volatility from ripple volatility. Baby Fox was fascinated by this. Besides everyone in the club referred to ripple volatility in meters. Poor Baby Fox got confused but was too embarrassed too ask. So, Baby Fox just assumed that ripple volatility and wave volatility were one and the same thing.

The official designated club measurer measured ripple volatility to be half a centimeter one day and recorded it as half meter in "wavethised" units. Baby Fox read it to mean a calm day because he interpreted the figures to mean that the wave was only half a meter high. A good day to paddle out to sea in his plastic dinghy, he thought. A ten meter wave came along and the sharks had a tasty feast that day.
 
Top