Great old threads about Gamma scalping
- Thread starter IV_Trader
- Start date
Surely model frequency impacts on the replication error as per Ursa. In this instance you want a positive replication error > execution costs in addition to crystalizing the stat/implied differential. What am I missing?
There is actually two more reasons.
Quote from riskarb:
No there isn't dumas; there is one.
Nope, I can think of two more, although both are insignificant for the most part. No cookie for you!
Yes, so can I, the Dividend!!!!!!!!!!
There is another reason.
Quote from MajorUrsa:
I'm not so sure about that: if you adjust every "X" point you'll lock in 20 cts every time. Reaching point "Z" will have locked you in 60 cts, whereas only using point "Z" gives you 80 cts. The point is, every time you "reset" your delta you give away opportunity in the direction of the trend.
Also, gamma is at it's greatest at the center of the straddle but the zero-delta point is moved around with every adjustment, thus it will take longer to acquire new delta once you're farther out.
In theory you could adjust as much as possible keeping the delta at 0 all the time. This will cost you money because the hedge instrument will be bought high and sold low. But, in theory, the amount lost will be equal to the amount of time value of the straddle at the beginning; if IV is a correct estimate of the actual established SV, you will not earn any money. In fact you will lose because of transaction- and carrying costs.
So, you can only earn from this method if you do not adjust too often. You have to profit from the delta being non-zero, letting gamma do it's work. The profits are represented by the holes between the original delta-curve and the broken line formed by the adjusted position.
Of course this takes discretionary choices by the trader. How wide should the openings be, should I be greedy of cautious? Riskarb's picturing of the psychological profiles is very accurate and familiar .
In my opinion it is only possible to do positive gamma-scalping on straddles when you are able to predict the movements of the underlyer. But if you are able to do that, you are maybe better off just trading the underlyer proper.
The attractive property of long straddles seems to be that they will gain in both directions when delta is 0, unlike the underlyer which has downward risk.
But when delta is not 0 you have similar problems as with the underlyer proper, but now you also have to compensate the fact that no movement means a loss as well.
Ursa..
it's very easy to back test both strategies and find out which one is better :
1.Let the initial position (straddle/strangle) expire.
2. Use stock for adjustments , always close intraday and keep separate PnL ( for stock only)
3. If stock's PnL is negative at exp , the "adjustments free " strategy is better.
Agree ?