There are 9 pages of the search term "Vega" and 348 of the term "RSI"

[newwurldmn]

I've been able to work through an answer to my question. So I'm thinking about vega now as basically deltas. If you look at the /ES and /VX futures, a 20 point SPY move is roughly equivalent to a 1 point move in the Vix. So if my portfolio has -10000 vegas, another way to think of it is being long 5000 deltas.

I don't know if I agree with your statement on hedging vega risk. I can structure a synthetic long in Vix to get long vegas. I could also overlay a bunch of SPX calendars. If I wanted to get short vega, I can sell premium. So at the portfolio level, you can definitely impact vega within a range.

So, at least going forward, here is my process: get my theta number where I want. Next, get my theta/vega ratio where I want it. Then convert vegas over to deltas assuming that -2 vegas = +1 delta (all weighted towards the spoos) to get my "vega adjusted" delta number. Finally, hedge total deltas to wherever you want to hedge it.

My only question now is where to hedge overall "vega adjusted" deltas to. Being truly delta neutral is going to be too much of a drag on my portfolio. My gut tells me to pick a delta range based on the total net liquidating value of the portfolio and hedge to within that range.

So what does my ending portfolio look like? Long theta, short vega and short delta.

Vega is correlated with spot price. However, "hedging" that creates basis risk.

In your examples:
1. buying VIX means you will take the view that the roll down in VIX futures will be < your theta/gamma profit. You are betting that if there is a big move in realized vol, your VIX will rally to account for it (this didn't ultimately happen this month).
2. Trading calendars, will create a forward vol position. If views on that forward vol changes, you will make/lose.
3. The delta - vega hedge can work, but needless to say, you can lose when the relationship breaksdown and it does fairly often.

In anyway you can lose in options, you can also earn. So you should recognize the basis risk you are adding by hedging out your nominal vega risk. That was the point of my initial post. And in volatility trading, basis risk generally means you have to trade bigger to achieve the same returns and you have to be more precise in your view and understanding of those risks.

 

[sle]

First, a bit of general trading philosophy. A proprietary trader should avoid trading a spread where one leg of the spread is introduced for risk mitigation (there are a few exceptions to this rule, but it's a longer discussion). If you feel that an change in one of the specific metrics would be too painful, you should think of reducing the position first and think about hedging second.

In case of hedging vega on an short vol position, you are changing the trade. If you are hedging delta and describing it as collecting theta, you are putting on a position based on an expectation of RV under-performing IV . Very straight forward and fairly easy to internalize. (The theta/gamma/vega are conincidental, the Greeks are for risk management rather then alpha comprehension).

Once your hedge your vega with another product, you are introducing a relative value component into the trade. If you bought longer-dated option, you added a forward volatility component, relative term structure etc (forward vol is a secondary risk in a calendar unless you are more or less gamma-flat). If you buy VIX futures, you added forward vol, vol roll-down, skew and a whole bunch of other risks. So now your alpha is offset by a bunch of other risks, with a bunch of noise.

Finally, if on a short gamma position you think that you can't handle the vega risk, you are doing something wrong.

IMHO x2

 

[fyretrading]

<--thats probably what you want

 

[jonny1lot]

Vega is correlated with spot price. However, "hedging" that creates basis risk.

In your examples:
1. buying VIX means you will take the view that the roll down in VIX futures will be < your theta/gamma profit. You are betting that if there is a big move in realized vol, your VIX will rally to account for it (this didn't ultimately happen this month).
2. Trading calendars, will create a forward vol position. If views on that forward vol changes, you will make/lose.
3. The delta - vega hedge can work, but needless to say, you can lose when the relationship breaksdown and it does fairly often.

In anyway you can lose in options, you can also earn. So you should recognize the basis risk you are adding by hedging out your nominal vega risk. That was the point of my initial post. And in volatility trading, basis risk generally means you have to trade bigger to achieve the same returns and you have to be more precise in your view and understanding of those risks.

You nailed the big problem with any long vol strategy perfectly, which is the cost of carry. I have no clue (no one does) when vol will expand and its a losing trade to just buy protection month after month.

The only long term profitable long vol strategy I know of involves either covered calls or a short a put, long a call spread (same thing as a covered call just widened out). Calendars can be profitable if you manage them real early, but they are calendars.

Just looking at what traded today, it looks like you could have sold the Nov 12.5 put for, what, .10, or the 12 puts for .05, there's just not that much juice to purchase a call spread there. I mean, realistically, how much protection is that really going to buy? Maybe I ratio and sell 2 of the 12.5 puts for .20 and purchase the 18/19 call spread? I mean, I don't really love that trade.

Every way I look at vol there just seems to be that persistent drag you talked about. Maybe I'm missing something, but if I'm having to put up with that drag, I might as well just sell ES futures against the position (Occam's razor).

 

[newwurldmn]

You nailed the big problem with any long vol strategy perfectly, which is the cost of carry. I have no clue (no one does) when vol will expand and its a losing trade to just buy protection month after month. The only way I know of profitably getting long Vix is to go into the Vix options and sell a put use that to finance the purchase of a call spread. Just looking at Nov, you can sell a 13.5 put for .35 which means I guess you can buy the 16/17 call spread? I mean, realistically, how much protection is that really going to buy? (Not much unless I am missing something).

Every way I look at vol there just seems to be that persistent drag you talked about. Maybe I'm missing something, but if you want to go long volatility with either /VX or Vix calls (or VXX for that matter), that drag eats into your theta profit. And in reality, if I'm having to put up with that drag, I might as well just sell ES futures against the position (Occam's razor).

I agree with SLE's post. It's very tricky to be short vol and hedge your vega exposure. The whole point of selling vol is that you are selling the tails. Except in a few specific circumstances, I would rather trade smaller than hedge my short vol against a long vol unless that was my explicit view.

Hedging short vol with futures is actually very dangerous because you have a convex risk against a linear risk and the ratios move around. It's easy to lose on the selloff (underhedged) and on the rally (overhedged).

And if you are thinking about buying vix. Don't think 16 or 17. Think 27 or 37. That's what kills volatility sellers.

 

[jonny1lot]

First, a bit of general trading philosophy. A proprietary trader should avoid trading a spread where one leg of the spread is introduced for risk mitigation (there are a few exceptions to this rule, but it's a longer discussion). If you feel that an change in one of the specific metrics would be too painful, you should think of reducing the position first and think about hedging second.

In case of hedging vega on an short vol position, you are changing the trade. If you are hedging delta and describing it as collecting theta, you are putting on a position based on an expectation of RV under-performing IV . Very straight forward and fairly easy to internalize. (The theta/gamma/vega are conincidental, the Greeks are for risk management rather then alpha comprehension).

Once your hedge your vega with another product, you are introducing a relative value component into the trade. If you bought longer-dated option, you added a forward volatility component, relative term structure etc (forward vol is a secondary risk in a calendar unless you are more or less gamma-flat). If you buy VIX futures, you added forward vol, vol roll-down, skew and a whole bunch of other risks. So now your alpha is offset by a bunch of other risks, with a bunch of noise.

Finally, if on a short gamma position you think that you can't handle the vega risk, you are doing something wrong.

IMHO x2

Thanks for the input on this. It's good to have somewhere to bounce ideas off of. I think you are spot on on saying that if a change in vega / delta would be too painful, it is time to start taking off at least part of the position.

 

[jonny1lot]

Hedging short vol with futures is actually very dangerous because you have a convex risk against a linear risk and the ratios move around. It's easy to lose on the selloff (underhedged) and on the rally (overhedged).

I'm not sure I understand? So you'd rather short equities outright to get short deltas?

 

[sle]

I'm not sure I understand? So you'd rather short equities outright to get short deltas?

No, he means that (a) a relief rally will hurt you on both (you lose on your short gamma and you lose on your long vega) and (b) you will find yourself losing money in calm upward market due to vols selling off without the corresponding pick-up from gamma (especially if you are already selling lower vols in the front).

 

[newwurldmn]

I'm not sure I understand? So you'd rather short equities outright to get short deltas?

I'm saying don't short deltas to hedge your vol. a rally will hurt you because you will be over short delta and a selloff will hurt you because you will be under hedged delta.

 

[sle]

Apparently we are saying different things and yet we are both right