How far off are the deltas?
Disturbing delta discrepancy between TOS and OptionsVue
- Thread starter darkshogun
- Start date
What exactly are you trying to do? If you are just selling premium you can throw your deltas out the window, they are useless. Your delta estimate is dependent on your vol estimate. If TOS is using a different vol skew then their deltas will reflect that. From memory I believe Optionvue uses a proprietary vol skew that they model their deltas off of. Most retail traders do not need to know their delta as they execute very plain vanilla strategies.
If you go into more detail about what you are trying to do, we can respond with what risks you need to be able to quantify.
Try comparing with the CME's deltas in their daily bulletin, e.g.
http://www.cmegroup.com/daily_bulletin/Section49_S_And_P_500_Call_Options_2013135.pdf
from
http://www.cmegroup.com/tools-information/build-a-report.html
However, as has been pointed out, there are no real true values because many of the models do not apply, strictly speaking.
For example, the Black model assumes constant volatility vertically and to expiration.
This is already violated at all point in time by skews. So strictly speaking you cannot apply most of the models. In an ad hoc manner, people take a "weighted average" volatility or an at-the-money one etc., to get out some values, but these assumptions, technically speaking, invalidate the use of simple models. Most of the values have to be "wrong" if you believe the model.
There are more sophisticated models than Black that do allow for skews. Some may be using these models, some not.
You hear people saying "whose delta are you referring to?" as there are many.
http://www.cmegroup.com/daily_bulletin/Section49_S_And_P_500_Call_Options_2013135.pdf
from
http://www.cmegroup.com/tools-information/build-a-report.html
However, as has been pointed out, there are no real true values because many of the models do not apply, strictly speaking.
For example, the Black model assumes constant volatility vertically and to expiration.
This is already violated at all point in time by skews. So strictly speaking you cannot apply most of the models. In an ad hoc manner, people take a "weighted average" volatility or an at-the-money one etc., to get out some values, but these assumptions, technically speaking, invalidate the use of simple models. Most of the values have to be "wrong" if you believe the model.
There are more sophisticated models than Black that do allow for skews. Some may be using these models, some not.
You hear people saying "whose delta are you referring to?" as there are many.
By the way, it appears that OptionVue's model may be one that has a deficiency relative to others in that, until now, they have used a single composite horizontal (month to month) implied volatility, instead of tracking separate implied volatilities for each contract month.
They are about to remedy this in a new version though.....
They are about to remedy this in a new version though.....
No, I believe they track each month separately. And they have an option where you can actually customize how you want to track it. They actually do a decent job of modeling volatility. However, for 99% of the guys on ET who simply sell teeny options or put on flys, modeling vol is useless.
That's for vertical skew, not horizontal skew.
Their horizontal data is a always a composite 60 days out (which, by the way, can also distort prices, although it is convenient)..
Some other vendors have data per contract per day with separate volatilities.
Now, they do also track individual prices of each strike on their servers only for simulation purposes. Maybe that is what you are referring to.
And I believe they will be adding horizontal volatility skew modeling in their new version just coming out.
One place where you obviously need to track each month separately is time spreads.
No, they do that for equities and cash I'm positive. For futures options I think you are right. I should have been more clear. I'm only referring to options on cash and equity, not futures. I think for futures the problem always was the lack of data. You didn't have that problem for equities and cash.
I was an optionvue user for a decade and I made a huge bulk of my option gains on time spreads and modeled them very accurately. Mostly biotech stuff. They had both the vertical and horizontal skew. And they allowed you to scan for both vertical and horizontal skew. That is how I found most my trades thank God.
That's interesting to know.
It does appear they are adding new enhancements to horizontal skew modeling, for futures at least, in version 7.3:
http://www.optionvue.info/videos/watch.php?vid=cbaa75ec2
I have not looked at the video yet though.
It is a fine product overall for sure
They completely ignored futures options for two decades. Probably because most of their customers didn't understand them and there is basis risk in most futures options that most people don't have the faintest clue about so it probably was a good idea. The irony of course is the futures options market was really there for the taking and they should have gone after it. But on the equity side, they had that stuff down well. Probably the best thing they had going was their customized scans. I think at one point I had over 60 scripts I wrote for the software.