2011 NFL Survior: Second Round

Well, the Texans-Broncos combo worked out well, which was fortunate. In my particular pool Houston actually ended up being more popular than New Orleans, which must have been people wanting to save NOLA for future rounds. That made Saints-Broncos the best combo from an round one EV standpoint, although it's obviously not as attractive when you consider future value.

Anyway, on to the next round. This week I only need to make one pick. I think it makes more sense to look at this from the standpoint of picking a path the rest of the way, rather than just picking one team this week.

For example, if I pick Baltimore this week, that basically means I need to pick the DEN-NE winner next week, since if that team doesn't beat Baltimore I'm screwed anyway. And then in the Super Bowl I'd be left with the NFC winner, since I wouldn't have the AFC champ available.

With that in mind, here are my estimates for surviving through each round based on who you pick this week, which I figured out using various Pinny markets:

Worth noting that there's a negative correlation between the R2 and CC columns. For example, if you pick DEN this week and they win, then you get to pick either Baltimore or Houston hosting the Broncos next week. On the other hand, if you pick New England and advance, you're forced to take the Ravens or Texans on the road in Foxboro, which is less than ideal.

The numbers in this chart didn't surprise me much. Baltimore is the obvious choice, since if BAL advances you can (likely) take New England next week, and then the (likely favored) NFC in the Super Bowl.

I don't think they're the best pick though, as popularity has a huge effect here:

Realistically, I have no idea how many people are going to pick the underdogs; it hasn't happened in either of the last two years in this pool, but there are a lot more people left at this point than there have been in the past.

I don't think it really matters though. As long as you're in a decent sized pool, I would go with Denver. It's an extremely ugly pick this week -- which is the best part, really -- but it puts you in a great position in future weeks, and gives you a chance to be the only person to remain undefeated through the Super Bowl.

2011 NFL Survivor: Wild Card Weekend

Quick format summary: Same idea as regular season Survivor: except pick one team on Wild Card Saturday, one on WC Sunday, one in second round, one on CC Sunday, one in Super Bowl. Can't use any team twice, lose and you're out.

As noted on Twitter the other day, this year's first round doesn't set up quite as favorably as it has in the past.

Last season there was a huge favorite on Saturday (New Orleans) which was extremely popular, as you'd expect. But on Sunday, the majority of entrants picked the Ravens, who were the day's biggest favorite but only ~61% to win. This essentially meant that any of the other three Sunday options were strong picks, particularly the Chiefs (who got blown out, of course).

This year both days set up pretty much the same as that Saturday, with the Saints at ~84% to beat the Lions, and Pittsburgh about 78% to advance against Denver. The one positive is that taking those two is the very obvious pick, and I'd expect over half the entries to go that route. In fact, I have made up consensus numbers for each team:

Saints 70%
Texans 17%
Bengals 10%
Lions 3%

Steelers 68%
Giants 17%
Falcons 8%
Broncos 7%

These numbers, along with the MLs from Pinny, give us almost everything we need. Just a few more things to consider:

• If someone picks New Orleans on Saturday, they're more likely to take Pittsburgh on Sunday than the average entry. In math terms, the % of entries that take both NO and PIT will be higher than the straight calculation of (.70)*(.68). This is taken into account in the EV table below.

• Future value isn't normally too much of an issue in the first round, since the teams that advance will all be playing road games next weekend. However, this season the Saints are the second favorite to win the NFC (+298/-341 at Pinny), so that must be considered. Pittsburgh also has a worthwhile amount of future value (+528/-642), if not quite as much. This is not taken into account in the EV table below.

• Beyond a specific team's future value, it's worth considering how you set yourself up for future weeks. For example, I think it's worthwhile to try to avoid picking the Lions if possible. That way, if the Saints win (likely) and I advance (maybe), I'll have the option of picking New Orleans over SF next week, which will not be a popular option since a) most survivors will have already used them, and b) they're on the road.
  

And finally, the EV table below (assuming $1 entry fee):

Unsurprisingly, the really obvious combo doesn't do very well in this analysis, as even if you advance by picking those two teams, it's guaranteed that at least half of your competitors will still be alive as well.

Such a popular pick being -EV allows 14 of the other 15 possibilities to be at least breakeven. Since I'm assuming only a 17/10 split in popularity between HOU/CIN, despite the Texans being 62% to advance, they come out looking like a pretty good option to pair with any non-Steelers partner.

I think the EVs in the chart above give us a good general idea of the best picks, but it's tough to be exact due to the uncertainty and importance of the consensus numbers. I will stand by Texans/Broncos as the best combo though, as it has a decent lead on the field.

Of course, there's a 87% chance that combo doesn't survive the weekend; all I'm really trying to do here is predict which option will have the highest EV when I repeat this analysis with the "real" consensus numbers next week.

2011 NFL Survivor: Week 17

Last week:

CAR: W, 48-16
PIT: W, 27-0
BAL: W, 20-14
HOU: L, 19-16
TEN: W, 23-17
WAS: L, 33-26
GB: W, 35-21

Teams used: Houston, Detroit, Pittsburgh, Tampa Bay, New England, Cincinnati, New Orleans,Tennessee, Philadelphia, Green Bay, Dallas, NY Jets, San Francisco, Baltimore, Atlanta, Carolina.

This week's consensus numbers from OFP and Yahoo!:

Impressively, I've already used each of the seven biggest favorites this week. That leaves me picking from some pretty marginal teams, but the best option left is Jacksonville, so they're the final pick of the regular season. Coming next week: Playoff Survivor!

2011 NFL Survivor: Week 16

Last week:

TEN: L, 27-13
ATL: W, 41-14
NO: W, 42-20
GB: L, 19-14
ARI: W, 20-17
NYG: L, 23-10
CIN: W, 20-13

Teams used: Houston, Detroit, Pittsburgh, Tampa Bay, New England, Cincinnati, New Orleans, Tennessee, Philadelphia, Green Bay, Dallas, NY Jets, San Francisco, Baltimore, Atlanta.

This week's consensus numbers from OFP and Yahoo!:

I'm pleasantly surprised that I have two decent options available this week, in the Panthers and Redskins. Carolina is a bit more likely to win and not overly popular, so they are the pick for Week 16.

2011 NFL Survivor: Week 15

Last week:

BAL: W, 24-10
PIT: W, 14-3
SEA: W, 30-13
NYJ: W, 37-10
DET: W, 34-28
GB: W, 46-16

Teams used: Houston, Detroit, Pittsburgh, Tampa Bay, New England, Cincinnati, New Orleans, Tennessee, Philadelphia, Green Bay, Dallas, NY Jets, San Francisco, Baltimore.

This week's consensus numbers from OFP and Yahoo!:

The Falcons are easily my best option this week, so they are the pick.

2011 NFL Survivor: Week 14

Last week:

SF: W, 26-0
NE: W, 31-24
CHI: L, 10-3
BAL: W, 24-10
NO: W, 31-17
DAL: L, 19-13

Teams used: Houston, Detroit, Pittsburgh, Tampa Bay, New England, Cincinnati, New Orleans, Tennessee, Philadelphia, Green Bay, Dallas, NY Jets, San Francisco.

This week's consensus numbers from OFP and Yahoo!:

Picking the most popular team again, but I really have no choice but to go with Baltimore this week.

2011 NFL Survivor: Week 13

Last week:

ATL: W, 24-14
CIN: W, 23-20
PIT: W, 13-9
DAL: W, 20-19
CAR: W, 27-19
NYJ: W, 28-24
HOU: W, 20-13

Well that is boring.

Teams used: Houston, Detroit, Pittsburgh, Tampa Bay, New England, Cincinnati, New Orleans, Tennessee, Philadelphia, Green Bay, Dallas, NY Jets.

This week's consensus numbers from OFP and Yahoo!:

The Patriots are a ridiculously good option this week, at ~95% to win and only being used by about 1 in 4 entries. Sadly I have already used them.

Beyond that, the 49ers are by far the biggest favorite of the teams I have left. They're more popular than I'd prefer, but they really don't have much future value, so San Francisco is the pick this week.