It's not.

Firstly, how would I even record this? There are 256 games across 17 weeks (every NFL team has one bye week), so such a system could possibly make the participants in the prediction league end up with a different amount of total games being recorded (some guys have their worst weekly score during a 16 games week while others have their worst record during a 14 games week). I could count this suggested bye week as 0-X ("X" being the number of games in a given week) to achieve a common total, but this defeats its purpose. Why would anyone want to go 0-X when they got 5-11 or 4-10?

Secondly, it's poised to cause an imbalance. There might be players who will never go below 7-9 while others end up with 3-13 at one point. This suggested bye week would penalize better predictors, because they would get more correct picks removed from their overall total. There's no way I am going to implement a "the harder you suck, the bigger your reward" rule.

Thirdly, what exactly is the harm of having a bad week anyway? It's likely to happen to everyone at some point, albeit to different degrees. If someone can't handle losing, then maybe they shouldn't play a game in the first place (this league is a game).

The rules aren't set in stone (for example, the playoff setup possibly needs to be reduced to less than 16 players eventually), so I am open for suggestions. But as it is, I see the majority of the laid out rules to be as good as they can get, so there aren't going to be many changes.

Your idea is not good, because it's going to punish people who do consistently well and show up each and every week of the season. If anything, these should be the people who get rewarded, not the ones who get penalized for their skill and effort. In fact, this is why I set up the rules for the playoffs in favor of these people. They will deserve it, because they will have earned it over the course of 17 weeks.

In order to balance your suggestion, I could remove both the worst and the best week of everyone at the end of the season, but this too would affect different people differently and thus distort the results. I'll use the example from my previous post again to show you why your suggestion is a bad idea. After the season is over, one guy's worst week is 7-9 while another guy's worst week is 3-13. You propose to remove these weeks, so one guy loses seven correct picks while the other guy only loses three. The former player may lose a position or two in the overall standings, because the guys who were previously right behind him get less than seven correct picks removed from their total, thus putting them ahead. Imagine how pissed that player would be; he did well all season long, but then loses a couple of spots in the final standings, because of a strange rule that affects players differently.

Such a rule is not going to be implemented anyway. I have given this prediction league a lot of thought during the off-season. How it needs to be organized, what the tie-breakers should be and so on. If removing the worst week would be a good idea, it would have been a rule in the first place. The one thing I am unsure about is how the PP will add up over the course of the season and if I should have handed them out to the top 5 players every week, but I think it will be fine the way it is. And the biggest question mark is how many people will keep predicting NFL games each week, because last week had already five guys go missing (outlawauron is excused for missing most games, because he didn't post in this thread until Sunday).

Tie-breaker in the regular season; part of the home field advantage in the playoffs.

I am only worried about how it's going to look in the playoffs, specifically in the matchup between the highest and the lowest seed. The home field advantage could very well exceed 50 points (difference in overall record + difference in PP between the two players), but then again it's not unreasonable that the lower seed should face an uphill battle in case the tie-breaker is needed in the playoffs.