Winds, on net, are bad in terms of en-route fuel burn and distance flown. All winds can be broken down into headwind/tailwind and crosswind vectors to determine their net effect, which is bad in both cases. We already know headwind/tailwind combos are bad, per Mir's comment:
|Quoting Mir (Reply 2):|
Consider a 3000nm trip, and the airplane has an average airspeed of 500kts. With no wind, it's six hours each way for a total of 12 hours. If you have a 100kt wind, the trip into the headwind will take 7.5 hours and the trip with the tailwind will take 5, for a total of 12.5 hours. So having the wind present costs you a half hour - this is because the fact that the headwind leg is longer means that it's being adversely affected for longer than the tailwind leg is being benefited.
In addition, crosswinds are bad too: Assume you want to fly 4 miles forward, but there is a strong crosswind that pushes you 3 miles off course to the left in the time it takes you to get there. To correct, you need to point into the wind a bit to track a straight line. In this case, if you know you'll get pushed 3 miles left, you aim three miles to the right. What happens is you fly the hypotenuse of a right triangle, in this case 5 miles, to go 4 miles forward with a square crosswind. More generally, if you need to crab X degrees into a crosswind to fly distance A, then you end up flying A/cos(X).
Between overcoming the crosswind component (flying the hypotenuse) and the headwind/tailwind problem that Mir mentioned, there are few ways to game the system. Most flights are short haul and one thus can't generally fly around the wind, and on long hauls, per Wilco's point, there are few use cases where one can fly around the headwinds. If there's a well-defined jetstream, as over the North Atlantic, there may be times you can game it, but once you add in Europe-Asia, you're still losing.