Most of you know me from the Sleep Well Portfolio and I am excited that so many of you have found a use for it. My other specialty is Options on Broad based indices and after talking to Tim Pierson we decided skew shape would be a fun topic. The recent highs in the SPX have created a huge skew shape disadvantage to options credit spreads over the short term. I am going to give a presentation on the subject on 12/3/20 with Tim on the Beginner's trading group. Before I do though I would love to hear if anyone reading this watches skew shape or uses a measurement of skew in there trading decision process? I look forward to talking more about this subject.
Skew Shape and How it Affects Options
- Thread starter Wayne
- Start date
Chuck Moore
Member
A lot of us trade some type of Butterfly; Balanced, BWB, ATM, OTM etc. Although they are a combination of credit spreads, a look at how Vol Skew affects the different Butterflies would be interesting. And what tools you may use to show the skew curve and compare it over different time frames.
I am not using the skew to make trading decisions since I don't see the correlation between the skew shape and individual option prices but I will watch the presentation and perhaps you can cover that topic
There are too many variables to make any decision on what option price to trade based on a certain skew shape which also varies in time so what may be a good entry at one point based on the skew shape it may not look so good when the market goes against the trade and perhaps the skew changes shape
I would consider the skew more like the VIX It's something to be aware off but not something that I would base a trade decision on
There are too many variables to make any decision on what option price to trade based on a certain skew shape which also varies in time so what may be a good entry at one point based on the skew shape it may not look so good when the market goes against the trade and perhaps the skew changes shape
I would consider the skew more like the VIX It's something to be aware off but not something that I would base a trade decision on
Hi Wayne, I have been working with a SPX skew. I'm very interested in your approach. I hope I will be able to participate live in Tim's group meeting.
I managed to put some tools around this topic. One of it I, partially, presented in one of TG1 meetings. There was no interest from traders but I still think it provides valuable way in understanding how skew impacts different trading setups.
If Tim and you agree (and I'm available) we can use this tool in your 12/3 presentation or one of next ones. I think topic is worth spending more time on it.
I managed to put some tools around this topic. One of it I, partially, presented in one of TG1 meetings. There was no interest from traders but I still think it provides valuable way in understanding how skew impacts different trading setups.
If Tim and you agree (and I'm available) we can use this tool in your 12/3 presentation or one of next ones. I think topic is worth spending more time on it.
We would love to see what you use and find a way to incorporate it into the subject. PM Tim and he can set it up.
Very good. I will try to find time to put new face to the program for easier presentation. I simply join the meeting unless you or Tim want to see it before - then we can talk in some quick Zoom, but again main idea was already presented on TG1.
(Again, I can't promise 100% that I will be able to participate live.)
(Again, I can't promise 100% that I will be able to participate live.)
wayne, i remember from a long time ago when i was writing my options analyzer, i read a paper that
said it's mathematically incorrect to add volatilities. for this reason i didn't implement it but i didn't
document the correct algorithm, nor the reference i read. i tried searching for it but i can't find it. do you
know for certain that it's correct to add them?
said it's mathematically incorrect to add volatilities. for this reason i didn't implement it but i didn't
document the correct algorithm, nor the reference i read. i tried searching for it but i can't find it. do you
know for certain that it's correct to add them?
All great points and all with merit. I am excited that the conversation has drummed up some research and conversation. I look forward to hearing what others use and if it correlates to trade results. I definitely am not saying that this is the best way. What I was pointing out is that by using a universal approach to measuring points on the skew vs ATM vol we can conclude that an environment is similar or different to other periods with the same ATM vol. This signifies a change in the options volatility structure and through my analysis has shown a correlation to my trade results when using a butterfly/condor structure. I would love to be corrected so that we could all get to the bottom of options mispricing for spreads in various environments.
Wayne, you showed examples of vol skew from 2015 and 2018. Did you do any analysis of vol skew behavoiur, or rather vol surface, for 2018?
I'm interested in methodology. Where (dte) and when sagging of vol skew was most dominant, how long it last, how 2018 compares to other periods etc.
I'm interested in methodology. Where (dte) and when sagging of vol skew was most dominant, how long it last, how 2018 compares to other periods etc.
I will revise my reference to word "improper" -- Observing the IV as shown is valid for more optimal strike choices for the referenced chain, but comparing to a different chain or time may be problematic.
I would like to clarify my thought process for my presentation on skew. As many mathematicians are likely cringing at what I said. I am always looking for relationships that apply to trade performance and am attempting to normalize variables as best as I can, without getting too detailed vs the net benefit. My hope is that everyone questions the methodology that I use, like we are doing here, so thank you. This can help us all get to truth in the common goal of adding value to everyone's trading and knowledge. Many people much smarter than I calculate options differently and I realize what I said was contradicting the very nature of the Black Sholes formula (something that won the Nobel prize). The simple math and concepts I presented have a relation to my own trading and were not meant to prove error in another form of calculating the "correct price of an option". I am viscerally aware that my observations do not conclude truth and am extremely interested in others approach to solving the same issue.
Marcas: My observations in 2018 and the vol surface behavior since have been that the "sag" and "doming" nature is much more variable and pronounced than it was in the past, AKA 2007-2015 when using the normalization approach I showed. I am still working on how or if it correlates to market direction and trade performance over the short run. There are two ways that I have been looking at this. First the Shape "Sag/Dome" we can also look at it as convex or concave in relation to the skewed side of ATM. The second is looking at a 1/2 ATM vol out into the skewed side and measuring the rise. In essence the slope of the skew at same DTE expirations. This is drastically steeper now "2018-2020" and more volatile than in the past "2007-2016". As for how long it lasts sometimes it is a few weeks and sometimes a few months. in 2020 the sag is showing up at new relative highs and and flattening out after the subsequent snap down only to come back as the market goes back up. This is not always there as in 2018 where it was rampant for many months until the bottom in DEC.
Here are 2 similar ATM vol environments and check out 1 "ATM vol down" and the overall shape. When I see this I think to myself how condors and butterflies gamma profiles look. The flatter the better has so far been beneficial. Have you seen anything similar?
Marcas: My observations in 2018 and the vol surface behavior since have been that the "sag" and "doming" nature is much more variable and pronounced than it was in the past, AKA 2007-2015 when using the normalization approach I showed. I am still working on how or if it correlates to market direction and trade performance over the short run. There are two ways that I have been looking at this. First the Shape "Sag/Dome" we can also look at it as convex or concave in relation to the skewed side of ATM. The second is looking at a 1/2 ATM vol out into the skewed side and measuring the rise. In essence the slope of the skew at same DTE expirations. This is drastically steeper now "2018-2020" and more volatile than in the past "2007-2016". As for how long it lasts sometimes it is a few weeks and sometimes a few months. in 2020 the sag is showing up at new relative highs and and flattening out after the subsequent snap down only to come back as the market goes back up. This is not always there as in 2018 where it was rampant for many months until the bottom in DEC.
Here are 2 similar ATM vol environments and check out 1 "ATM vol down" and the overall shape. When I see this I think to myself how condors and butterflies gamma profiles look. The flatter the better has so far been beneficial. Have you seen anything similar?
Last edited:
Wayne, I share your approach to trading research. I also try to go my way instead of following somebody else footsteps. Like you, I leave open all the time possibility to be wrong and would love somebody proved me wrong. Just I expect proof to be logical not sort of 'everybody knows'. I sense a lot of misunderstanding about IV graphs and it's meaning out in the public. I can't be sure though because discussion about this is practically non existent. Maybe I struggle with obvious things. So, do not take any critique personally. I hope Jim and Gary can explain more why they object to adding IV. I oppose it as well as a general indicator for structures evaluation, but for different than mathematical reasons. I didn't look at it in the past and I will now. There is a chance that it became quick and dirty, 'good enough' method in _some_ cases.
Re 2018. I'm not questioning your observations. I was simply wonder if you did any statistical analysis of this period. I did not, I do not see a value for my trading in doing that (I'm focused on shorter terms). This days I relay, like you, more on observation and understanding than on statistics. I don't use your '1/2 IV ATM' method. I generally do not like to operate in delta space (having delta or sigma as x axis) but what I be looking at is not only rise but strike distance and atm IV level and try to extract useful info from combination of all of that.
I'm happy that this conversation happens. I hope it will last for a while. I'm convinced that exchange of thoughts is an excellent way to push understanding forward. I want to point out to one issue we should take into consideration if we continue. You use ONE. Gary and I stopped using it while ago, I do not know about Jim. Issue is this:
This pic is taken from TOS' On Demand. Take a look at iv levels in comparison to ONE from the pic you posted above. I have about the same values as TOS from my database. I do not know what settings do you use in ONE but iv values are quite different. I do not care about perfect accuracy of iv and greeks (yet another river to jump into
) but what we see can not be ignored and should be addressed in the future if we to go beyond conceptual things.
Re 2018. I'm not questioning your observations. I was simply wonder if you did any statistical analysis of this period. I did not, I do not see a value for my trading in doing that (I'm focused on shorter terms). This days I relay, like you, more on observation and understanding than on statistics. I don't use your '1/2 IV ATM' method. I generally do not like to operate in delta space (having delta or sigma as x axis) but what I be looking at is not only rise but strike distance and atm IV level and try to extract useful info from combination of all of that.
I'm happy that this conversation happens. I hope it will last for a while. I'm convinced that exchange of thoughts is an excellent way to push understanding forward. I want to point out to one issue we should take into consideration if we continue. You use ONE. Gary and I stopped using it while ago, I do not know about Jim. Issue is this:
This pic is taken from TOS' On Demand. Take a look at iv levels in comparison to ONE from the pic you posted above. I have about the same values as TOS from my database. I do not know what settings do you use in ONE but iv values are quite different. I do not care about perfect accuracy of iv and greeks (yet another river to jump into
) but what we see can not be ignored and should be addressed in the future if we to go beyond conceptual things.
Last edited:
I've looked at net_iv. For regular butterflies this is a simple measurement of the vol skew convexity ("Sag/Dome" of the graph). For unbalanced structures it does the same just in not so clean manner.
Wane:
1. Do you use net_iv numbers in your trading?
I suspect that it is only a minor element you consider before hitting send button.
2. Did you came up with some normalization process of net_iv for various structures?
There must be some sort of normalization, otherwise one has to use tables for each structure separately. I can imagine using rBF (regular butterfly) placed over tested structure and taking net_iv number from this. But...
Also... net_iv seems to be quite sensitive to precise iv calculations...
Wane:
1. Do you use net_iv numbers in your trading?
I suspect that it is only a minor element you consider before hitting send button.
2. Did you came up with some normalization process of net_iv for various structures?
There must be some sort of normalization, otherwise one has to use tables for each structure separately. I can imagine using rBF (regular butterfly) placed over tested structure and taking net_iv number from this. But...
Also... net_iv seems to be quite sensitive to precise iv calculations...
Oh for sure net iv for a structure is all over the place. Maybe something to take note, like... For this structure with 0 delta I'm at whatever net iv. See if it correlates to expectancy. I have not used iv that way though. Strange about the ONE issue. I have cev and surface turned off. I will take some time this week to cross reference to tos.
I do normalize skew shift and atm vol to calculate the environment in respect to skew before sending a trade. Pending results I will shift from higher risk all the way to a net buyer of vega, (not net iv). I don't measure iv for a structure. Interesting though it is that in 2015 to get flat delta we had significantly higher net iv sold vs bought for a structure of spreads only. Compared to 2018.
I believe the more we pay attention to the relationship of iv of our options we will find that it does correlate to total edge/expectancy.
Tom is working on getting the video up soon.
I do normalize skew shift and atm vol to calculate the environment in respect to skew before sending a trade. Pending results I will shift from higher risk all the way to a net buyer of vega, (not net iv). I don't measure iv for a structure. Interesting though it is that in 2015 to get flat delta we had significantly higher net iv sold vs bought for a structure of spreads only. Compared to 2018.
I believe the more we pay attention to the relationship of iv of our options we will find that it does correlate to total edge/expectancy.
Tom is working on getting the video up soon.
OK, I thought too much of it. My primary focus is on structures, that's why I applied net_iv here. To see what is going on with the whole graph I might put a bunch of, let's say 50pw rBFs from ATM to far OTM and plot net_ivs to see where and how much iv graph convex (aka where to place a trade). Of course width and sample density can be adjusted to match trader's needs. I'm talking about net_iv not a steepness. Agree that applying net_iv to structures is pitb unless somebody traders only 1 or 2 structures constantly.
You mentioned expectancy. As a reservation: expectancy is something I have a beef with, or rather with how this term is often used/missussed/missunderstood - but let's not start it now, even if it is pretty darn important.
You mentioned expectancy. As a reservation: expectancy is something I have a beef with, or rather with how this term is often used/missussed/missunderstood - but let's not start it now, even if it is pretty darn important.