General Discussion on Option Modelling

Marcas

Member
The goal was simply to see if the formula works by entering the IV and the result being the option price no more than that I am not planing on using it for any modeling or anything like that
OK, this is fine. I don't know how Bruno designed spreadsheet but BSM formula is used to calculate IV from given price or to calculate price and greeks when fed with IV number. Now, when you tested this, try to think about purpose of all this jazz. Why one wants to take the trouble at all?

if you have strong command over Greeks, you're in a better position to manage your positions well. That includes understanding interactions or cross derivatives like for instance how IV also impacts Delta.
I agree. The only thing is I'd use word 'understanding" instead of "command", or even "feel" would be better. This is what options trading is about (imo), "to be better prepared for the unknown". Plus you have to add arsenal of trades to take advantage of IV at hand and of greeks dynamics. I prefer that over "rules".

Not sure if you read an article on Aeromir's main page by Joanna White about being part of trading group. There are so may directions trader can take that there is no way one person can sufficiently cover everything alone. Traders need to find his own trading way, his style. It's much easier to sort it out by hanging in proper environment. I agree with Bruno's conclusion about trading, but I took different path. Other way was chosen by Kevin Lee (machine learning) I also think this is not the right path and don't want to follow, but I can be wrong and want to hear from Kevin, from Bruno, from Ted, and other traders about their experience and how they go about improving trading results. Unfortunately Kevin disappeared from my radar.

That article says about building group in fashion of Mr Hill's Mastermind idea. Of course there should be group of people of similar interest. In our case this will be about trading options. Mr Hill, if I remember correctly, suggested that members of Mastermind group should contribute and also that each member should benefit from participation, not only via output of the group as a whole but on personal level ( I think Mr. Napoleon Hill suggested monetary incentives for each person).

I looked up few communities, admit - not to many, and most of them are in 'passive mode'. That means most of participants are in listening mode trying to learn as much as possible and are not willing to participate (by fear of asking stupid questions or not having enough time or not being interested). This is fine but I also understand that in this environment traders who can contribute or share their 'trading discoveries' are more and more reluctant in doing so when they get no responses, positive or negative, at all. As I remember old CD Forum was pretty good once, with a lot of knowledge pouring in from different directions. Now Aeromir with TG1 is, in my opinion, one of the best alternatives out there as a platform to exchange thoughts but still it is not horribly active.

I remember once there was 'Eagles' group out there. Wonder if it is still on? How it is doing... But this was private one constructed in Mastermind fashion (I suppose). Did time of public forums passed? Do you, gents, know of any other options trading forum which is more active, that is not guru dominated or focused on single strategy? Maybe private groups, a la Hill's Mastermind, is the way to go? Or, maybe, it is all about pure commerce now.

That said, I say 'thanks' to all active participants of TG1 and to Bruno and Tim and Himanshu add others who contribute to trading knowledge by sacrificing their effort and time.

Let's say that you have that vol surface equation worked out, or you have a very comprehensive data set available in which case you only need local approximation to fill the gaps, that continuous non linear function IV(X,t) tells you what IV to use in volatility surface mode.
Bruno, you lost me ;)
 

status1

Active member
You can check out elitetrader
They seem to be quite active although it has additional chit-chat that is not relevant
 

garyw

Active member
Bruno:
A penny for your thoughts for small "improvement" on future volatility "assumptions".
I have been using "sticky-strike" for baseline implied volatility estimates for a few years, which is similar to what TOS does with iiv, and OV with EIOIO. This basically is an acknowledgement that we have NO crystal ball on what the iv will do in the future, so we assume it will remain about what it currently is. -- I have been comfortable with this, as we know what we don't know!
However, for those of us with access to the data (I have all SPX chains) we can make a slight improvement to compensate for term structure changes (again, the assumption is the term structure shape may continue), we can merely "lookup" the expected future implied volatility of each strike from the available chains. With the proliferation of chains in SPX, we can observe the "desired" new projected time +/- 1 day resolution (so interpolation to precise time will be accurate)! No fancy math, but just observe what the market is stating the IV will be if current expectations persist!
This seems to be a nice simplification to working thru moneyness projections.
While I do not think this approach makes major improvements, I do think is is a bit more precise than the standard EIEIO and TOS iiv as it takes into account Contango of the time frame being targeted.
 

Marcas

Member
Gary, I'm not Bruno, but I use this method in my sandbox although I don't use it on regular basis. This is still static and still unaccurate method.
Yes, you see t0 like that is a bit more realistic but 'handle bar' method still need to be used. How much more 'realistic' it is? Hard to tell, probalbly it's accuracy depends somehow on present state of IV and other factors. I like to apply 'correction' to call side (or put) when IV hoover around edges (very high or very low)
Math involved, as you said, is simple. The main reason I don't use this method regularly is that with 'standard' method it is easier to compare your graphs with TOS/ONE/OV/*. I call it "sticky iv" method.
 

garyw

Active member
Gary, I'm not Bruno, but I use this method in my sandbox although I don't use it on regular basis. This is still static and still unaccurate method.
Yes, you see t0 like that is a bit more realistic but 'handle bar' method still need to be used. How much more 'realistic' it is? Hard to tell, probalbly it's accuracy depends somehow on present state of IV and other factors. I like to apply 'correction' to call side (or put) when IV hoover around edges (very high or very low)
Math involved, as you said, is simple. The main reason I don't use this method regularly is that with 'standard' method it is easier to compare your graphs with TOS/ONE/OV/*. I call it "sticky iv" method.
I think you must be referring to a different topic, since you make reference to "T0" being improved. T0 represent no change to time (ie Now). My reference was for "volatility estimates", which infer at a time in the future, as present time value is known (can be simply derived) {Simply may not be the best choice of words here}
Consider my statements as relating only to the best value to use for implied volatility at each T+N (unique values for each N) for each option at the present spot price only! -- The "normal" sticky strike method would replicate the T+0 iiv per option at the T+N points! -- Consider only the single spot price for this exercise.
 

Marcas

Member
Indeed. I was talking "along" t0 line, but,if I understand you correctly, the same method can be used "along" horizontal (t+N lines). Instead of using IVs from your expiration cycle to calculate t+N, you use IV set from expiration +N. This method include steepening of the IV skew with time which usually occurs. You include in your t+N graph what market predicts. If only those predictions were accurate...
Must say I don't work with t+N in my sandbox only t0 and expiration - source of my mistake in quick response above. I also look at various expiration cycles - not to draw t+N lines but to see what market thinks 'now' (which, in same cases, can be the same thing :) ). I'll wait for Bruno and his opinion.
 

garyw

Active member
@Marcas :
I think we are in agreement, that "... along the T-n lines" (which includes along the t-0 line) introduces many unknowns (which is beyond the scope of what I wish to address currently). My point is to merely zero-in a bit more on what we do know. Below I include two plots of a current BWB position I have on in SPX, with one using my normal "EIOIO" style, and the other "improving" on it slightly, by looking up the T+n IV's from the IV Surface.
1577045022789.png
While it is almost the same, observe the proximity of the T+48 lines to the Expiration line differences, as well as the slight differences in the STATIC ASSESSMENT values for the Expiry-15 and Expiry-30 for the two cases. I remain confident that the IV Surface lookup is a better estimate than the assumption that the T+0 IIV will be constant over T+n.
A simple way to view this in TOS is to configure your position (Entry Trade), then merely alter the Target Expiration to one that is shorter by the # days you wish to consider as your n! -- voila. you implemented a crude Volatility surface lookup! (to simplify, just look at the price/cost --no greeks/IV calculations or models actually required, as the focus is at the current spot price!)
1577045200690.png

For the BWB trades I do on SPX, this change produces a "less optimistic" outcome for my trades than the normal method, which I believe to be slightly more accurate. --- Those trading BWB's using closer to the money strikes, may observe the opposite effect!
 

Marcas

Member
Gary, yes, we are talking the same thing. "Sticky IV" method produces different graphs. In many cases they are "less optimistic" and "more realistic" but not always. This is simply graphic display of market sentiment at given data-point. For environment we have today, 'sticky iv' method do produce more probable graphs. Still, even those need 'handle bar' adjustment, especially for bigger moves.

One can benefit from switching entirely to 'sticky iv' method but one should know what he's doing especially if he trades by graphs (by delta for example) because it can get 'messy' if you use TOS and improved method in your code together - the same way if you use TOS greeks and ONE greeks interchangeably.

You can put both t lines ('standard' and 'improved') on single graph to see the difference clearer.
Waiting for Bruno as I know he is convinced that more accurate graphs are achievable.

I'm not sure what 'unknowns' you are referring to at the beginning of your post.
Good job on producing 'sticky iv' graph with TOS. In my opinion TOS is good for study, for 'understanding things' but to cumbersome for practical use.
 

garyw

Active member
Regarding: "I'm not sure what 'unknowns' you are referring to at the beginning of your post. "... The unknown, I intended to refer to is the unknown Volatility at each "underlying" price along every T+n line! -- We are using a fixed value with EIOIO, Sticky IV, Sticky Strike, Sticky Delta, and Sticky Moneyness per T+n N value! As a Russian friend of mine once clearly stated (as a humorous point) "There is much we do not know about the unknown!" -- The more you ponder the statement, the more profound it becomes!
Perhaps the IV to be used for these would be a ever widening "range" with further excursions from the Spot price, but if you think about this, that also has serious flaws! (so we seek some sanity with understanding the inability to know future volatility, and make a compromise that seems viable to each of us)
 

Marcas

Member
This is true what you said.
There was time I was thinking about t-lines (predictions of iv) as an example of Heisenberg's uncertainty principle, where you can tell future value of iv but not time-point or vice versa :D, but even this approach failed. Now I simply don't bother. With U I'm trading I don't need to spend time on this. I rather stress test positions or, better to say, try to get familiar with variety of possible outcomes position can produce (and impossibles to some degree), betting on 'typical behaviour'. Still watching work of others in this field, yours, Bruno's and Kevin's mostly. Thanks for sharing.
 
Top