Picking the correct Smallworld team defense each week might be the most critical determinant of success in this game. The range of actual weekly points for defense this year spans more than 1000 SWP, from +73 to -935 (through week 5). No other roster slot presents so much potential volatility. So, what's the best way to analyze the choices?

Certainly, we can look at the weekly averages for each team defense and draw some inferences from those values, but historical averages for each team are highly dependent on the actual opponents. In fact, the weekly opposition matchup is probably the most important factor to consider. (OK, maybe weather is really the most important factor, but that's pretty hard to predict, as we all know.) While it's fairly easy to look at historical results for any team defense, identifying the results for the opposing defenses of any offense is problematic. Until now - Guru to the rescue!

The following table shows the per-game point averages for each team defense, and for the opponents of each team. The table on the left sorts the averages for team defenses, while the right hand table sorts in order of each team's opponents' defensive averages.

No matter which way you look at it, the average for all teams is the same, at -418 SWP/game. That's an interesting number relative to the 325 SWP deduction for a bye-week defense, suggesting that having a bye-week defense produces an above average result. Even more interesting, though, is to look at how individual teams compare to that -325 SWP threshold. Look at team defenses in the conventional way, only three teams are averaging better than -325 per week. But look at the opponents averages. No less than nine different teams have an opponents' average better than -325 SWP through five week. One of those teams, Indianapolis, has never stuck the opposing defense with a result worse than -325 SWP.

This suggests to me that the opposing offensive team may be a more important factor than the defense itself in choosing the best defense for the coming week. A consequence of following this strategy, however, is that optimal management of defenses may require frequent trading, so the cost of the extra trades must be weighed against the potential benefits.

In the end, I guess the magic answer is to find a strong defensive team that plays multiple weak offenses in the next few weeks, where the games are to be played in locations subject to frequent bad weather. Not too much to ask for, is it?