Trades using greeks from an individual implied volatility model or volatility smile?

[darkshogun]

I don't know about other platforms but in TOS you can change the settings to use a "volatility smile approximation" instead of the default setting of "individual implied volatility" when looking at the greeks of the options you are trading.

I currently have 2 demos running simultaneously, one using the volatility smile approximation to measure the greeks, and the other using the individual implied volatility setting. I'm using the same strategy on both, but the positions are slightly different due to the fact that the deltas and other greek readings are not the same between the 2 settings. I opened each set of positions close to delta neutral. The positions consist of 10 SPY butterflies and 1 SPY call, and 10 IWM butterflies and 1 IWM call.

I've only monitored the positions for a few days so far. What I've observed so far is that the deltas of the positions using volatility smile approximation setting behave far more stably than the deltas of the positions using the individual implied volatility (IIV) setting. The deltas of the latter will make wild swings during trading hours while the deltas of the former react less drastically to changes in price. I've already been forced to adjust the positions that uses the IIV setting by adding debit spreads to balance the deltas, whereas the positions using the volatility smile approximation have remain untouched with suitable delta/theta ratios. These positions are also currently showing more profit.

When trading options, do you trade based on the greeks of a volatility smile or do you trade based on the greeks of an individual implied volatility model?

 

[SIUYA]

It depends on what you are actually trading for.
If I understand you correctly making a few assumptions from your post....

These are all theoretical prices and not market prices otherwise the PL should not really change between variations of the model if you are using market to market prices.
Hence the 'smile approx' accounts for a skew in the OTMs where as the 'individual vol' is a flat line volatility applied across strikes?

If so - off the top of my head, by already using a skew it should probably flatten/smooth the changes in greeks hence smaller changes in results. Whereas a flat implied (non smile) means there is more responsiveness to changes ....hence what to use depends on how you are trading.

Do you want to try and capture small moves OR sit and not worry too much knowing that the prices paid and assumptions made are ideally covering your position?

now take this with a pinch of salt as an old school set of heuristics......

There is one thing that you always have to keep in mind - all options end up either at 0 or parity.....regardless of the model.
You need people to take the other side of your trade and they might be using something completely different, and trading completely differently.
There are plenty of different models, and as a retail guy any close approximation will suffice whereas MM often put in skews either because - they have model that approximates their views of how markets actually/historically move...or it simply is based on supply and demand, the aim of the game is risk control and not saying - my model was right (but just not right this time, or I did not have enough money to prove it over the long term).
Dont get me wrong - some models are poor compared to others but they are all approximations apart from the ones that are just so completely flawed and this is usually able to be shown mathematically.

 

[jamesbp]

I don't know about other platforms but in TOS you can change the settings to use a "volatility smile approximation" instead of the default setting of "individual implied volatility" when looking at the greeks of the options you are trading.

When trading options, do you trade based on the greeks of a volatility smile or do you trade based on the greeks of an individual implied volatility model?

In my opinion, it is simply wrong to calculate the Greeks / Probabilities based on IV% at each strike.

There are various Models that will estimate/extract the Implied Probability Density from the Option Price Chain, and then calculate Probability / Greeks in a consistent manner.

Doesn't mean they will be right ... models are just models ...
But better than being completely wrong using a seperate IV% at each strike!

 

[newwurldmn]

In my opinion, it is simply wrong to calculate the Greeks / Probabilities based on IV% at each strike.

There are various Models that will estimate/extract the Implied Probability Density from the Option Price Chain, and then calculate Probability / Greeks in a consistent manner.

Doesn't mean they will be right ... models are just models ...
But better than being completely wrong using a seperate IV% at each strike!

Absolutely you should be using a different vol for each strike.
How is using the option price chain different from using the implied vols for each strike.

 

[Dolemite]

I don't know about other platforms but in TOS you can change the settings to use a "volatility smile approximation" instead of the default setting of "individual implied volatility" when looking at the greeks of the options you are trading.

I currently have 2 demos running simultaneously, one using the volatility smile approximation to measure the greeks, and the other using the individual implied volatility setting. I'm using the same strategy on both, but the positions are slightly different due to the fact that the deltas and other greek readings are not the same between the 2 settings. I opened each set of positions close to delta neutral. The positions consist of 10 SPY butterflies and 1 SPY call, and 10 IWM butterflies and 1 IWM call.

I've only monitored the positions for a few days so far. What I've observed so far is that the deltas of the positions using volatility smile approximation setting behave far more stably than the deltas of the positions using the individual implied volatility (IIV) setting. The deltas of the latter will make wild swings during trading hours while the deltas of the former react less drastically to changes in price. I've already been forced to adjust the positions that uses the IIV setting by adding debit spreads to balance the deltas, whereas the positions using the volatility smile approximation have remain untouched with suitable delta/theta ratios. These positions are also currently showing more profit.

When trading options, do you trade based on the greeks of a volatility smile or do you trade based on the greeks of an individual implied volatility model?

Not sure if this applies but if your position is a fly with a deep in the money call, the TOS platform for some reason doesn't track your greeks properly. A lot of people use Optionvue to model their trades instead. Not sure why TOS has a hard time with deep in the money options.

 

[jamesbp]

Absolutely you should be using a different vol for each strike.
How is using the option price chain different from using the implied vols for each strike.

Maybe better if we work with real numbers rather than abstract principles.
I have seen, but don't have access, to TOS Option Chains that I understand use IV at each strike to calculate Probabilities / Greeks.

If you do have access to TOS Option Chain, can I ask you post details for

- SPY
- June14 Expiry
- IV / Probability ITM / Greeks covering at least 2 Standard Deviations
- Vertical Spread Prices

Then we can compare Option Prices v Model Outputs ... and see if we can draw some inferences.

 

[newwurldmn]

I don't have TOS so I can't get the info you are requesting, but it shouldn't matter.

Price and ivol are the same thing. I thing what the OP is referring to is the ivol by strike (each option priced individually) vs some parameterized surface (each option still priced individually but then fitted to a curve).

The curve will be more stable because it will dampen supply demand effects and some stale pricing. But it offers more complexity because if the fit isn't good you can get some very wacky results. We had 3 guys working on the parameterizing of the surface and they never got a great model. Too many different types of vol surfaces out there. I doubt TOS is going to be better.

We used to use fitted curves. It makes sense when you are doing a lot of OTC and lite-exotics. It also makes sense if you are making markets in AAPL and have 100 lines on the april expiry. The smoothing effect helps identify kinks which can be taken advantage of.

But for a retail account I think the cost of having to verify the fitted curves outweighs any smoothing benefits.

 

[jamesbp]

Absolutely you should be using a different vol for each strike.
How is using the option price chain different from using the implied vols for each strike.

We maybe at cross purposes, I said that you cannot use IV by strike to calculate Greeks / Probabilities as TOS seems to do.

So, lets play with some real numbers from TOS that I do have

- Stock XYZ , assume no interest / no dividends
- 45 DTE
- Spot Price = 187.00
- ATM straddle = 6.00
- 180/181 Put Spread = 0.18
- 193/194 Call Spread = 0.18

Questions? based on the above data:

- What is the expected 1 standard deviation move?
- What is the Probability of the Put Spread finishing ITM
- What is the Probability of the Call Spread finishing ITM
- Is there any Vol Skew

Broad approximations sufficient for this purpose ..

 

[darkshogun]

Not sure if this applies but if your position is a fly with a deep in the money call, the TOS platform for some reason doesn't track your greeks properly. A lot of people use Optionvue to model their trades instead. Not sure why TOS has a hard time with deep in the money options.

I don't know what's wrong with it. I certainly won't use TOS until I'm satisfied by testing my positions on the platform for several months. After only testing both models for a few days I'm on the verge of completely abandoning any further testing of positions using their Individual Implied Volatility setting, but I will be continuing to test the "Volatility Smile Approximation." However this setting is annoying too, as it doesn't give you an accurate P/L line. So I've been putting positions on using the greeks that I get from the Volatility Smile Approximation but I have to switch over to the IIV setting if I want to see where the current P/L sits in the risk profile. I would think they'd be concerned about fixing this problem.

I'm going to be trying OptionsHouse as well. Does anyone trade with them? If so, do you trust their greeks?

 

[darkshogun]

Maybe better if we work with real numbers rather than abstract principles.
I have seen, but don't have access, to TOS Option Chains that I understand use IV at each strike to calculate Probabilities / Greeks.

If you do have access to TOS Option Chain, can I ask you post details for

- SPY
- June14 Expiry
- IV / Probability ITM / Greeks covering at least 2 Standard Deviations
- Vertical Spread Prices

Then we can compare Option Prices v Model Outputs ... and see if we can draw some inferences.

Here it is. This is info from a demo after the markets are closed. Not sure how helpful it will be: