Survivor: Week 3 - Taking the Chalk

There's nothing worse than losing in Week 1 - except losing in Weeks 1 and 2. And I came close after Matt Schaub threw a pick-six to put Tennessee up eight with six minutes to go. Of course, a couple minutes before that, the Texans were up six with a toothless Tennessee offense backed up on its own one-yard line, so this was a game of extreme probability swings. And I won't even get into the end-of-regulation icing-the-kicker fiascos.

But that's how this season has mostly gone. The Pats - the board's biggest favorite - were life and death with the Jets on Thursday night, the Bears won at the last second, and the Eagles actually lost. Even most of the teams that won more comfortably - Cincinnati, Atlanta and Baltimore - were hardly in the clear until fairly late. Other than Green Bay, there were no stress-free blowouts. We should get more in Week 3.

One further note (and if you don't care about how this all works and simply want to read the picks skip down to the main chart):

A reader brought to my attention yet another error I've been making, namely comparing teams head to head and assuming the rest of the pool (other than those two teams) remains intact. I don't know why I assumed if Team A is a better pot-odds play than Team B heads up, then that must be the case regardless of what C, D and E do. Clearly, if Team A has 50 percent of the pool on them, and Team B has 10, it makes a huge difference what happens to the other 40 percent. We can illustrate this by looking at two possible scenarios:

So by assuming everyone else survives, I was undercounting the full effect of the pot odds. Of course, it's much more likely everyone survives than everyone gets wiped out - after all, people are almost universally picking favorites, but we do need to account for the probability that a certain number of them do lose, and that will boost the pot odds angle.

So, in prior weeks I had undercounted the disparity in win probability by simply comparing winning percentages (or losing ones) and not accounting for both, and I've also been undercounting the disparity in pot odds by assuming all the other teams survived. Maybe those errors cancelled each other out at times in the past, but I'm pretty sure that wasn't always the case, and I apologize for that. When I first wrote this article, I did it based on gut feel, but over the last couple years I've been attempting to quantity what I knew in my experience to be true: that every time I've won or come close to winning it's because I avoided a major upset, simply because the favorite in that game wasn't available to me. And so I've tried to build a strategy around that premise, though I'm realizing now it wasn't as precise as I thought.

The solution is when comparing Team A and Team B, to figure out the expected number of people who will lose taking Teams C, D, E, etc. and subtracting them from the total in both scenarios to calculate the pot odds. This should give a more accurate sense of the payout you get by taking a team with fewer people on it.

Okay - that out of the way, let's turn our attention to Week 3.

First off, note the two insane moneylines - you almost never see ones as big as Denver's, and the Seattle one is just off the charts. In fact, Vegas is so scared of SEA-JAX, that it's -4400/+2200 - an insane spread, meaning an enormous vig should you bet either side. Averaging the two it's still a 33:1 game or better than 97 percent.

But 53 percent of your pool is on them, so they're not an obvious slam dunk, especially given Denver is a whopping 92 percent favorite and has only 17 percent on them. Let's see which is the better bet.

In our hypothetical $10 buy-in with 100 entries, if Denver wins and Seattle loses there are 47 people left minus the other expected losses. The Vikings are at 71 percent and have 11 people on them, so they have an expected loss of three. The Saints are roughly two, the Niners one, the Pats about a half. There are also a some unlisted teams with fewer than one percent of the entries. So let's estimate the non-Broncos/Seahawks losses at seven.

So it's 100 - 53 = 47. And then take off another seven = 40. So if Denver wins and Seattle loses, your pool equity would go from $10 to $1000/40 = $25.

If Denver loses and Seattle wins, there would be 83 people left, minus the seven expected losses from the other teams, making it 76. $1000/76 = $13.16. So the payout from taking Denver over Seattle is 25/13.16 or 1.9 to 1.

But how much more likely is the Seattle-win/Denver-loss than the opposite? The Seahawks are 97 percent to win, and Denver is eight percent to lose, making 7.8 percent. A Denver win/Seahawks loss is three percent of 92 = 2.8. The ratio of 7.8 to 2.8 is 2.8. In other words, the payout for taking Denver over Seattle is a good deal less than the added risk it entails.

And if Denver at 17 percent taken and 92 percent to win can't beat Seattle, there's no way the other teams who are far bigger risks with only slightly better payouts will either.

My Picks

1. Seattle Seahawks

I don't need to tell you why they're a huge favorite at home against the Jaguars. I did take Jacksonville +19.5 because I'm pretty sure the Seahawks will get drunk on Saturday night while the Jags might just take the game seriously, but even so I give the Seahawks a 92 percent chance to win this game.

2. Denver Broncos

The Broncos should roll at home against the Raiders, but with Ryan Clady, the team's best offensive lineman out for the year, and Von Miller still out, Oakland could hang around if it's able to get to Peyton Manning. Still, I give the Broncos an 87 percent chance to win this game. (Incidentally if you do the above math with my numbers instead of Vegas', it's much closer between Denver and Seattle, but the Seahawks still win out).

Notable omissions:

San Francisco 49ers - Trent Richardson moved Indy's Super Bowl odds from 75:1 to 50:1, but it didn't move the line, and I'd expect San Francisco's offense to bounce back at home against a soft Indy defense. Still, Andrew Luck has a knack for late-game comebacks, and the Niners secondary is vulnerable.

New Orleans Saints

The Saints defense has actually been good early on, but I don't trust them to keep it up, and Arizona has some balance now.

Minnesota Vikings - The Browns are starting a backup QB and anyone's guess at RB, but the defense isn't bad, and the Vikings still have Christian Ponder at quarterback.

New England Patriots - Still no Rob Gronkowski, no Danny Amendola and no Shane Vereen. Aaron Hernandez is in jail, Wes Welker is in Denver and Randy Moss has retired. This team was life and death with the Jets and Bills. I'd expect more of the same with the Bucs.