NFL Survivor Pool Strategy: Week 1 Advice

September 8, 2010 - by Matt

Fantasy, pick’ems, survivor … oh my! It’s good to be back in football season. This year, we’re launching a recurring series of posts specifically about picks and strategy for NFL Survivor Pools. We’ll post our thoughts based on what the TeamRankings.com models are telling us each week, along with other relevant information.

Also known by the less politically correct moniker “suicide pools,” Survivor pools require contestants to pick one NFL team each week. If your pick loses its game, you’re out of the contest. The catch, of course, is that you can’t pick the same team twice.

In math geek terms, these contests are fascinating optimization problems where the underlying data changes every week. Your optimal strategy likely depends on a number of factors including the teams you have left to pick, the number of contestants still alive in your contest, public opinion on the current week’s matchups, the state of the NFL playoffs races. Fun stuff!

Two basic guidelines to get us started:

1) We’re in it to win it. We’re going to be making our picks with the goal of maximizing our odds to win an NFL Survivor pool this season. That means we won’t always recommend choosing the most likely team to win based on the projections of our models for the current week.  (If we don’t, we’ll tell you why not.)

In other words, it’s critical to understand that simply picking the most likely team to win each week, based on the teams still available to you, is not typically the best strategy for winning an NFL Survivor pool. There are 17 weeks in the NFL regular season, and going the distance means keeping both the short and long term in mind.

2) Know the rules. There are lots of ways to do survivor pools. We’re going to assume the standard “your pick needs to win its game, and you can only pick each team once” setup. (We will not be addressing picks against the point spread in this series.) If you win, you continue on in the contest; if you lose, you are out.

If the rules for your pool are different, fire away in the comments and we’ll do our best to respond. We’ll also keep track of teams we’ve already picked and let you know if any of them would be a particularly strong pick if you’ve ignored our advice previously. Without further ado…

Week 1 Survivor Strategy

Teams already picked: None.

This early in the season, it’s not easy to introduce any long term strategy into our pick making process. It’s possible to use Vegas futures (bets on season win totals) to establish some sense of which teams will likely be good this year, and evaluate teams’ 2010-11 schedules based on that information. You could also use our data-driven NFL power rankings from the end of last season, and assume nothing changed in the offseason.

Neither approach is rock solid, though, as there’s no substitute for actual performance this season. Consequently, in Week 1 we typically just look for lopsided matchups.

Unfortunately, we can’t remember a Week 1 with such an exciting and apparently well matched slate of games. Minnesota at New Orleans. Green Bay at Philadelphia. Baltimore at NY Jets. Even the teams that are supposedly going to suck all seem to be playing one another! You want to choose between Cleveland and Tampa Bay?

We’ve still got to make a pick though, so let’s look at our TeamRankings.com NFL Week 1 Predictions and the Vegas lines to identify the brightest possibilities. Here are three games that sit atop our list of potential picks:

1) Detroit at Chicago
2) Arizona at St. Louis
3) Carolina at NY Giants

Detroit at Chicago (Team Rankings Pick: Chicago)

In Week 1, Chicago has the highest win odds in two out of our three models that make NFL game predictions. However, our Week 1 algorithmic projections are still partly based on data from last year. Is Detroit the same team it’s  been the past few years? Or is this the year the Lions finally start to turn the corner? How is the Bears’ new offense going to fare? At this point we don’t have any solid data to tell us one way or another. With a spread of -7, the most lopsided on the odds board, Chicago doesn’t seem like a bad pick. But can we do better?

Arizona at St. Louis (Team Rankings Pick: Arizona)

Last year the Cardinals had Kurt Warner at QB and Anquan Boldin running opposite Larry Fitzgerald, and until the Cardinals prove otherwise, our models assume Arizona’s performance for Week 1 2010 will largely be the same as at the end of 2009-10. That seems like a fairly risky assumption, based on what we know about Arizona. We’re not sure St. Louis is going to be vastly improved, or even improved enough to take down Arizona, but we’re a little nervous about this one. The spread, at Arizona -4.5, is much tighter than it would have been last year.

Carolina at NY Giants (Team Rankings Pick: NY Giants)

Our Similar Games model is giving the Giants a 67.4% chance to win the game. Those numbers incorporate what we feel is a somewhat undeserved bump in Carolina’s power ratings, on account of how the Panthers finished last season. They beat Minnesota, the Giants (41 to 9) and New Orleans in largely unimportant Week 15-17 games. The Giants head into the season with no major reasons, as of yet, to expect a big dropoff from their high level of play a year ago (before the injury bug struck), and will play their first home game in a brand new stadium. Vegas currently has New York as 6.5 point favorites.

What The Public Thinks

It’s always informative from a strategy perspective to review which teams the general public seems to favor. Here are the top user picks for Week 1 from ESPN’s Eliminator Challenge game, as of Wednesday evening:

1) Tennessee (home) over Oakland (26.8% of users)
2) San Diego (away) over Kansas City (17.1%)
3) Chicago (home) over Detroit (11.9%)
4) Miami (away) over Buffalo (11.3%)

No other pick garnered more than 6.6% of the public vote.

Tennessee is one of our models’ higher rated picks this week, but looking at the public numbers, the Titans look extremely overpicked. They’re a strong favorite, sure, but their odds to win don’t appear any better than several other teams.

San Diego is a perennial public darling, but faces some holdout issues that are concerning to us. Plus, the public seems to have a much more favorable view of the Chargers and the Dolphins than both the Vegas lines and our models. Sure, their opponents may have been horrible last year, but that’s not a reason in itself to make a pick.

Our Week 1 Survivor Pool Pick: NY Giants over Carolina

There’s nothing like starting off your Survivor pool with a relative nail biter of a week. We’re picking the New York Giants in Week 1, knowing we could have about a 1 in 3 chance of not making it to Week 2. Cue the Gloria Gaynor soundtrack!

The Giants seem to have win odds that are even a little better than the Titans this week, but only 4.8% of ESPN players picked them, compared to 26.8% who picked Tennessee. That’s a little crazy. Even if the New York’s win odds were slightly worse than the Titans, it still may be worth avoiding Tennessee from a competitive strategy standpoint.

If the Giants win and the Titans lose, we’ll be off to a great start. A lot of these Week 1 games could be close, so if New York ends up winning and the Chargers and Titans both lose, you will still be alive in your pool, while almost 45% of your competitors will have been eliminated. According to our projections there is a little less than a 10% chance of that scenario playing out. That’s pretty appealing to stat geeks like us.

If you completely disagree with this pick, our honorable mention pick goes to Chicago at home over Detroit. There’s a decent amount of uncertainty in that game, but Detroit is still the team with more to prove. Third place from a competitive strategy standpoint is probably Dallas over Washington, since almost nobody is picking the Cowboys and they are currently favored by more than a field goal, but there are a lot of question marks in that game.