I have thought about this a great deal. I have always exited early, knowing that the influence of gamma is much greater than either the influence of vega or theta. Of course, by exiting early, I am "leaving money on the table." Even so, the reward doesn't justify the risk of a quick move. Plus, by exiting early, I free margin funds for new IC's. Ironically, if I allowed my IC's to expire, I would have lost money on about 8 over the past year and the losses would have eaten my gains. I would appreciate comments on your personal experiences.
To understand what I'm about to describe, it is first necessary to understand that I do not trade Iron Condors as a unit, but as two credit spreads. Furthermore I only trade index options credit spreads. Some of the explanation my be overly complicated if you are trading Iron Condors on single stocks. I let my credit spreads expire UNLESS the market gap risk exceeds my limits. The first element of market gap risk assessment is to estimate the likelihood that the market can gap the distance between the underlying instrument price at the close of trading for the option until the settlement price is set. For example, the RUT options (Russell 2000) cease trading on Thursday but the price is not settled until Saturday. For me, the probability of the gap closing must be less than 10% or I will exit the spread in the afternoon on the last day of trading. Using this strategy, most of my spreads expire worthless. None have gaped and gone ITM. Most that I exited remained OTM, but I was not disappointed about leaving money on the table. It could have gone against me for significant losses.
Once you get to a week and a half or less, gamma risk really shoots up, and that actually underestimates the risk: there were days in the fall of '08 where the range from low to high was 10%, for instance. More "normal" but volatile times you could see 5% moves, which will easily take a worthless option up to ITM. If you have lots because you sold lots of spreads, that hurts. I use weeklies only to defuse the gamma risk of holding sold spreads: they're cheap, but they have very high gamma, natch. In the short time they've been around, they've saved me twice already from situations that otherwise would have been costly. As I've posted before, I learned the use of front month options for their high gamma from collars I used to create on stocks I owned when earnings were going to be reported: the bought puts were always front month, regardless of whether the sold calls were. The idea was that they go to 100 delta fast on a collapsing stock. Any other use is asking for trouble, IMO. Close 'em at least a week and a half before expiration.
I'm glad to see posts with views different than mine. It makes me think harder and watch more carefully. Thanks.
Even with different underlyings (index futures, gold/oil futures), the greeks' properties always hold true.
I don't have a fixed rule about holding ICs to expiry or not. Each expiration represents a unique situation. However, generally, I prefer to close out a few days prior to expiration (i.e. assuming I haven't closed out prior to expiration week) and move on to the next month.