Now that the NFL schedule is out, people are going through and making game-by-game predictions for each team. That is an impossible task: No one can tell you with any degree of certainty today who’s going to win a game six months from now.

But if you’re going to try it at all, you ought to at least start from the premise that every game that one team wins, another team has to lose. Which means that, assuming there are no ties, NFL teams will go a combined 256-256 this season.

So when ESPN’s team of NFL writers compiled their predictions, it was amusing to see that they picked a total of 289 wins and only 223 losses. Yes, league-wide the Worldwide Leader thinks NFL teams are going to go 289-223, a .564 winning percentage in a world where the only certainty is a .500 winning percentage.

The problem isn’t that ESPN’s staffers went overboard with the best teams. In fact, the best record assigned to any team was 12-4, for the Rams. If anything the folks at ESPN might have understated the number of teams that will have particularly strong records.

No, the problem is that none of ESPN’s writers want to admit that the teams they cover stink. The worst record assigned to any team was 5-11 for the Browns. The next-worst was 6-10 for the Cardinals. And four teams are projected by ESPN to go 7-9. Other than that, every team is predicted by ESPN to go 8-8 or better. That is obviously way too rosy an outlook. Last year there was an 0-16 team, a 3-13 team, two 4-12 teams, four 5-11 teams and three 6-10 teams.

Overall, ESPN’s writers predict that the NFL standings will finish with one 12-4 team, seven 11-5 teams, six 10-6 teams, five 9-7 teams, seven 8-8 teams, four 7-9 teams, one 6-10 team and one 5-11 team. That’s not just overly optimistic, it’s mathematically impossible.