If you operate a long backroom schedule and roster it accordingly, you can see how many flights you can crew and avoid selling flights you can't. You can also see that 12 into 9 won't work before you start selling it. If you sell 6 months out but operate with a rolling 6 week roster, one day you will have to tell your boss that you 5 weeks out can't make the following week work since you don't have enough crew to cover the already sold flights. And at that stage it will be to late to recruit and train people to make up the shortfall.
I remember when the published roster was only 1 week out. This worked when they where a small airline and the following cancellations didn't affect hundreds of thousands of people that got huge press and gained political interest. You could disappoint only a small percentage of your total customers with a simple refund. They've just discovered that being the big-boy also have some disadvantages.
Problem is you are never going to roster far enough out using just more manpower. It will be a case of too many cooks and too litle oversight. And what about running scenarios. We are talking making up maybe 3 or more alternative full year rosters with different parameters. Luckily others have already had the problem, solved it, and FR just have to utilize more of the resulting tools with some adjustments/localization.