Sounds like you're approaching it the right way then. As for answering your question - I did the best I could here: http://www.elitetrader.com/vb/showthread.php?s=&postid=2008933#post2008933 I wish I could do better but it's a big subject and impossible to cover in a few words. I'll try to keep giving examples as they come up. In the meantime, get yourself a good analyzer/simulator like Hoadley and just keep playing "what if" games with it - especially simulating what will happen if the implied volatility of different strikes change.
Gamma scalping is really about looking for short periods (days rather than weeks) of choppy up/down moves in the underlying. This is mainly done with a) front month options (<30 days) as they have the greatest gamma and b) at-the-money options because they are generally cheaper than other strikes. Generally you are looking for periods when the implied volatility remains static (of goes up if you are fortunate) and time decay is not too punative.......avoid trading too close to expiry (can't be gamma-hedged efficiently in last few hours). A good place to start backtesting would be to look at where realised volatility has moved ahead of implied volatility (graphs on bloomberg or ivolatility.com) and what has been the reasons behind this (earnings , financial figures etc). Technical analysis of the underlying itself can help with looking for areas where realised volatility might be about to increase. I hope this helps.
I would have NEVER entered scalping position based on RV/IV relationship. Two reasons: 1.RV is not reliable , it can go down at any time and drag IV down 2.RV calcs methodology is very tricky: high levels can be attributed to ONE big move that occurred 25 days ago. In 5d from now (when this day will be out of calcs range) RV will drastically collapse. Ever saw these ugly graphs (vertical free fall) on Ivol ? One needs to find a rent free ( or closed to it) position , based on potential IV ramp before known event. And even if you are right here, you still need to nail the magnitude (speed) PS : keep the thread going .
In his book "Options as a Srategic Investment", McMillan argues that volatility often trade within a well defined range and his therefore more easily predictable than prices. I would like to have your opinion regarding the above statement. Do you believe it is true/false? Can you mathematicly demonstrate whether the statement is true/false?
It's very true that implied volatility - the VIX for example - is a completely different animal than price, and behaves quite differently. Implied volatility (and real volatility) represents emotion - complacency when it's low, panic when it's high. There are 4 times in its 18-year history that the VIX was up above 47%, and it just stayed there for a moment, intraday, and corresponded to the minute with an important low in the underlying. Humans simply cannot maintain that intense level of emotion for long. So it will never happen IMHO that the VIX will go steadily over a period of weeks or months from 40, to 50, to 60, to 70, to 80, etc. There will always be a natural tendency for it to revert back to a "neutral" level, somewhere around 20% or so. Price, of course, is quite different - it can just keep climbing and climbing ad infinitum.
I looked at the VIX chart: http://finance.yahoo.com/echarts?s=^VIX#chart3:symbol=^vix;range=my;indicator=volume;charttype=line;crosshair=on;ohlcvalues=0;logscale=on;source=undefined -If we look at the VIX over its full history, it seems indeed to be mean reverting (with a mean around 18). -If we look at the VIX over the last year, it seems also to be mean reverting (with a mean around 22). Now, over other periods, it seems to be trending: -from 92 to 94, the trend is down -from 96 to 98, it is up -from 03 to 07 ,it is down The point I want to make is that looking at a 1 year history we might question Mc Millan's statement that Implied Volatility is mean reversing. Do you think that when considering 1 year historical data on stocks' IV it is reasonable to assume that it is mean reverting?
Don't forget that the Gamma scalping is a dynamic, not static strategy and could only be successful if the position is constantly adjusted to neutralize its Vega and Theta. Gamma will only manufacture Deltas in your favor if you stabilize Vega and Theta.
I am on my jury duty, Can't really type. We had a lunch break, so I checked the ET. Get your self a copy of Cottle's book , "Coulda, Shoulda, Woulda"; it has a pretty good chapter on the Gamma scalping. If you want, you can PM me your e-mail address so I can send you an electronic copy of this book in .pdf format.