You make money off the decay on the front month but wouldn't the back month where you are long decay also? Example: You sell front month for $1. You buy back month for $2. Front month expires worthless, back month is now worth $1. In my head your net profit is $0. But I don't think this is the case... I assume the back month would decay less than the front month due the options having exponential decay on expiration. But is it really that much more? A few cents maybe? What's the risk?
Volatility Changes over time. The bet on the difference in decay rate. Specific events, time of year for specific commodities, you get it. http://www.theoptionsguide.com/calendar-spread.aspx
RGLD: I don't trade nor do I like Calendars! To me personally their behavior is odd. The basic idea behind calendars is related to the time-decay of premiums. The time decay (assuming price is not changing) accelerates as you approach Expiration. Therefore the Front month should erode faster than the back month. So, selling the front month, and buying the back month, is an attempt to capture the differential in the time decay. In practice, however, the price of the underlying is likely to move, making this messy. Also, volatility spikes can ruin your day, as these can result in backwardation, wherein the front month may increase more than the back-month (harming your position in an additional way). Others that like them, may provide other view points.
The options decay at rate of the square root of time. Here is a rough guide: Our ATM strike for stock XYZ shows the following 1st month = 1 x 2^1/2 = $1.00 2nd month = 2 x 2^1/2 = $ 1.41 3rd month = 3 x 2 ^1/2 = $1.73 As the 1st month goes to $0, we expect the 2nd month to decay to $1.00. Our net profit is: 1st month prem. - (2nd month decay) or $1.00 -($1.41 -$1.00) = .59c
Everything depends on your view of future realized volatility and relative pricing of vol between the months. But, also grasp the concept that Bob mention about directionality. That adds the concept of skew behavior and favorable b/a spreads. Before he drove that concept in my head, all my calendars were done at ATM which may or may not have the most favorable pricing.
I find time spreads very directional. You can create your spread and be long/short and have some degree of vega risk, which is the point. However, if the underlying moves far away or toward the strikes, it will often determine if you make or lose money, unless it precedes an event where the vols make a big change like after earnings. I like them, but you have to have some plan and expectations. Like other option spreads, you should not do them randomly.