There's a pretty easy way to do this.

Let's face it: Vegas knows better than you do what the true lines are. So use them. In this game, there's no need to try to beat the casino. Take the chance it tells you each underdog has of winning (also known as the "moneyline", I can help you convert that into a percentage if you need it), then multiply it by the number of points you'll win if you do, and voila! You know what the best and worst picks are for the week. Here, I'll do Week 4 for you:

1. 11.5 WASHINGTON at Dallas 19.4% 2.231
2. 9 OAKLAND vs. San Diego 24.7% 2.223
3. 7 BALTIMORE at PIttsburgh 31.3% 2.191
4. 10 KANSAS CITY vs. Denver 20.4% 2.04
5. 6 SAN FRANCISCO at New Orleans 33.3% 2.0
6. 8 ST. LOUIS vs. Buffalo 24.1% 1.928
7. 7 ATLANTA at Carolina 26.0% 1.82
8. 7.5 HOUSTON at Jacksonville 23.8% 1.785
9. 3.5 CLEVELAND at Cincinnati 35.8% 1.253
10. 3 MINNESOTA at Tennessee 39.2% 1.176
11. 3 CHICAGO vs. Philadelphia 37.3% 1.119
12. 2 ARIZONA at NY Jets 46.5% 0.93
13. 1 GREEN BAY at Tampa Bay 47.8% 0.478

The first number after the matchup being the chance the underdog has of winning, and the second number the average number of points you'll get by making that your underdog of the week. Washington, Oakland, and Baltimore are the best choices. Why long shots? If you think about it, it makes sense: Why go for a 50% chance of getting one point when you could go for a 20% chance of getting a dozen? It's not at all risky to go for the 20% shots, it's simply correct.

Trust me, there are people in your group who know this sort of stuff (probably the Miami over NE guy). They'll be the ones with the highest point totals when it's all said and done. If you want to win, you've got to be aware of it.

O.k., I'll bite -- Where do you get the "moneyline," and how do you convert it to a percentage?

Any decent sportsbook really. Pinnaclesports.com is an easy one. Just click on NFL.

The "moneyline" can be a little confusing before you get familiar with it. Take the Denver/KC game. It says -410 for Denver, +370 for KC. That means you have to bet $410 on Denver winning the game to win $100 if they do, and you have to bet $100 on KC to win $370 if they do.

This is saying that the book thinks the true price of the underdog/fav is +390/-390 or thereabouts (and it's making you pay $20 vig on the bet whichever side you take). To convert from +390 to percent chance of winning, just do 100/(390+100). Maybe it makes more sense if you think about it in terms of how often you'd have to win at those odds to break even.

For example, +200 is a 33% chance of winning, +300 25%, +400 20% and so on. Calculating that will really help you. I'll do it for next week if you're interested.

If you want much more simple advice, never pick a team that's less than a six point dog, it's very very rarely going to be an optimal play. Pick from the long shots. It works out in the long run.

Now that the numbers are up, I am flipping a coin between BAL and SF. Sorry, I just use my gut.

Statistically, it looks as though your system will net about 2 points a week. But last year, it took more than points a week to win the thing. So, is your system doomed to a mediocre finish?

That's not how math works.

Yes, you have to get lucky with the optimal ones you pick. Like most contests, this one is luck + skill. You might as well have some skill to go along with it.

At the least, do yourself a favor and never pick small underdogs, even if you "just can't see" one of the longshots winning.

Among the underdogs that will yield the most points, I like Baltimore over Pittsburgh in a low scoring affair. They are playing good defense as usual, and Pittsburgh's starting RB Willie Parker is out this week, and Pittsburgh depends on its running game to set up its passing game. Ben Roethlesberger is beat up too. Baltimore's offense is suspect, but among the big underdogs I think they have the best shot.

Good luck.