One additional thing about enroute winds: there is both seasonal and daily variation. We had a good handle on seasonal variation (winds are normally out of the west with stronger winds in winter) and divided the year into five seasons to account for the seasonal variation. Domestically, we had lots of schedule latitude. Internationally, due to how most airports outside the U.S. manage slots, we were usually tied to a fixed arrival and departure time for the entire Summer or Winter industry season so could only vary times on the U.S. ends. (the industry seasons are Summer from the last Sunday in March to the last Saturday in October with the rest being Winter - those dates correspond to when Europe goes to and from DST).
But while we knew what the seasonal variation would be, daily variation dwarfs it. If the history and wind model said a market should expect a 30 knot headwind in summer and a 50 knot head wind in winter, a given day in summer might see anything from a slight tail wind to 100 knot head wind. On a six hour flight, that 100 knot wind, 70 knots over plan, is going to add an hour to the flight (but daily variation that extreme was pretty infrequent).
There was one international market where the seasonal variation was so extreme (about 100 knots) that we did monthly block times for it. Lots of manual work on my part as much of the automation we used to support making our forecasts did not go that level. The westbound longest block time in winter was about 25% longer than the shortest time in summer. And even then, in April when the wind really started dying, it would not be unusual to have no westbound on-time arrivals for the month by today (April 9) and yet I could state with reasonable confidence that we'd get to 50% by the end of the month as once that wind change kicked in, we'd start arriving early.