I tend to adjust my trades by the Greeks. If you don't know much about the Greeks of options trading, let me put that another way. I tend to adjust my trade to keep the profit-and-loss (PnL) line or "T + 0" line relatively flat within a one-standard deviation move from the then-current price. On the chart below, the one-standard deviation move to either side is marked by the lighter-blue column.
My Live Trade as of the Close, February 27, 2013:
The lighter shading on either side of the then-current price of 909.92 marks a one-standard-deviation move either side of that price. The white T+0 line marks the current profit-or-loss, a loss in this instance. This was a trade for which I set aside $52,000 in margin, although the margin requirement at this point was closer to $30,000, so the small loss was just that at the time: a relatively small loss.
Theoretically, it would stay small across a one-standard deviation move. Adjustments would be required if the RUT moved too far beyond a standard deviation in either direction and that T+0 line curved down more sharply. However, over one standard deviation, it looked fine.
As an additional test at the end of each day, I often roll implied volatilities down and look at the shape of the upside curve (because implied volatilities tend to come down when prices rise)and roll them up and look at the shape of the downside curve (because implied volatilities usually rise, sometimes rapidly, as prices drop). Moreover, because my brokerage's platform and most others allow me to do so, I roll the date forward a day to add in time-related decay.
I understand, as you should, too, that the tests I'm running return only theoretical results. Real-world results may be quite different, but this practice allows me to determine how I want to adjust my trade. For example, minutes before this chart was snapped, I had sold one of the butterfly contracts that compose this butterfly. That was done to raise my deltas since they had become too negative. So close to expiration, I tend to choose adjustments that lower my overall investment in the trade rather than raise it.
These previous paragraphs describe what I'm doing on a weekday to test the next day's movement and decide what adjustments I should make at the end of the day, if any. What about on a Friday? Simple, right? We roll the date forward three days and not one.
Not so fast. If it were that simple to accumulate three days of time decay from Friday afternoon to Monday morning, everyone and her uncle, to employ an idiom, would be selling options on Friday afternoon and buying them back cheaper on Monday morning, absent three days' worth of time decay. The market makers would be buying them from us on Friday afternoon, taking the other side of our trade, and having to sell them Monday morning for a cheaper price, taking a loss. They're just not going to make a habit of that, or else they wouldn't stay in business long.
Through the years, I've heard several market makers and former market makers talk about what they do. One caveat is that this presumes relatively quiet trading conditions and no big announcement due over the weekend. Starting mid-morning Friday, they use their nifty programs that calculate theoretical options prices to re-price the options. They re-price them to incorporate two days' worth of decay. The only decay the option trader can typically expect from Friday afternoon to Monday morning is about the same amount that is typically occurring overnight during a non-Friday weekday.
Since we all know there are a limited amount of inputs into the pricing of options, how do market makers re-price the options so that, by Friday's close, about two days' worth of time decay has already occurred? They roll down the implied volatilities.
To show you what happens, I've rolled down the implied volatilities for the trade I showed in the chart, reducing them by three percent.
Chart of Live Position with Implied Volatilities Rolled Back
This chart is quite a bit different than the previous one. The white T+0 line has curved up into the profit zone, just as it would if time decay had occurred with no other changes being made. However, the curve steepens quite a bit and isn't so flat. That new theoretical profit can seemingly all be lost by a one-standard-deviation move to the upside, at least.
For those familiar with the Greeks, you'll notice that delta has changed quite a bit, too, and gamma has grown more negative. That combination comprises the reason that the curve slides down so fast to the upside. Delta has changed from -40.77 to -85.37. It's much more negative. Theoretically, more than twice as much money is lost for each point the RUT moves to the upside.
Several Fridays in late January and early February, when fiscal cliff discussions had been kicked down the road and Europe was quiet, I noticed this happening. Beginning about late morning on Friday, my position deltas started growing more and more negative. As the afternoon progressed they grew rapidly and alarming negative, even though the RUT's price might not have changed much. Suddenly, adjustments were needed, and maybe rather big adjustments. I knew this was the weekend effect in action and commented on it to some of my trading partners, but it was difficult to ascertain how much credence I should give what I was seeing.
Why should I question how much credence to give it? On Monday mornings, depending on what's happening, those implied volatilities may be rolled right back to their rightful level, whatever that might be.
If market makers suspect that there's risk of an adverse event over the weekend, they're not going to roll those implied volatilities back. Whether or not we're seeing the weekend effect and how much import it's having on our trades may be difficult to ascertain for those of us who have only a general working knowledge of what's going on. I would include myself in that group.
What does that mean in live trading? That means that on Friday, especially on Fridays when we don't know of much risk of adverse effects over the weekend, we're likely to see this effect. Not all complex positions will react just as mine has in January and early February, but I was forced to make adjustments when my delta-based adjustment levels were reached. However, I elected to make as conservative of adjustments as I could and still stay at least on the edge of my own preferred guidelines.
What we shouldn't do is wait until late in the trading day Friday afternoon and then roll the dates forward three days and expect three days of decay from that point on. We're likely to have seen two of those days' decay already factored into the price of our options. It bears noting that my two charts illustrated a case in which the underlying's price was inside the tent shape that is the profit zone for the butterfly. If price had been outside that tent shape on a Friday, particularly a Friday close to expiration week, the weekend effect might have escalated the losses seen that day, moving the today line closer to the loss that would be seen at expiration. Perhaps, then, if price has moved outside the profit tent of your butterfly, calendar, or iron condor, you might want to adjust sooner than the end-of-the-day, even if you're a once-a-day adjuster and your adjustment time is late in the day. Of course, there are always risks when moving away from your trading plan. I can't verify this is a good idea for all trades, but it is something to consider when prices go outside the tent on a Friday morning, stay there after the first hour, and look likely to stay there throughout the day.
If I wanted to know more about the weekend effect, I'd likely search through an index of one of Jeff Augen's books to learn more. I know that his collection, Options Trading Strategies offers several references to what happens with options over the weekend, and that may be specifically in his Volatility Edge in Options Trading, part of that collection. A quick search of the CBOE's free educational webinars with the word "weekend" in the search box turned up two presentations by a former market maker.
For most of us, however, we should know enough about the weekend effect to know when we're likely to see it, how it's put into effect, and what not to expect as far as time decay as we head into the weekend.