2008 NFL Undefeated Odds

BetUS has granted my previous request and posted odds on each team to go undefeated during the 2008 NFL regular season. Using the winning percentages derived from each team's over/under, I've determined the actual odds of each team going undefeated for comparison. The list is quite long, so I've only posted the odds for the five most likely teams below, but the odds for each team are here.

If you believe the over/unders--and it would be silly not to--then the Patriots are by far the best bet of the 30 teams. That's not to say they're a good bet; not even close. As you can see above, they're expected to go undefeated about 1.3% of the time, while it needs to happen at least 6.3% of the time for it to be a good bet. But compared to the rest of these, it's a bargain.

The reason for this is that the relation between W% and undefeated percentage is not linear. A .700 team does not have twice as good a chance of going undefeated as a .350 team; they are actually about 65,000 times more likely. This is quite easy to see if you do the simple calculation (.700^16), but most people obviously do not. So as you go down the list of teams, the chances of them going undefeated decrease dramatically, while the odds change only slightly.

Which brings us to what may be the single worst prop I've ever seen:

I defy you to find a worse bet than this, of any type, at any book. You will not. If you were to bet $10,000, the expected value of your bet would be -$9,999.97.

They've also posted more detailed odds on the Pats, regarding their exact regular season record. Again, I'm using the .761 W% derived from their over/under odds, to come up with the "true" odds of each record:

I have no idea how they came up with these.

The Pats' over/under is 12.5. I think it would logically follow that their two most likely win totals are 12 and 13. Yet BetUS has installed 15 as the favorite, closely followed by 14. This makes no sense.

Maybe the thinking is that they went 16-0 last year, but that was obviously somewhat fluky, so this year they'll go 15-1? I honestly have no idea.

This has the effect of putting value in each number between 9 and 13. That is pretty amazing, since all of these add up to 124%. The best bet here is 10-6; it's not often that you can get 15-1 odds on something that happens nearly 10% of the time.

2008 NFL Wins Over/Unders, Part 2

Last week, when I posted Sportsbook's win totals for the upcoming NFL seasons, a few people (including myself) wondered how we should factor in the juice. For example, New England's total was 12, but the over was -130, and the under was EV, so we'd expect them to go over more often than they went under. With the help of RMMB statistics guru Adam, I figured out how to account for this.

First, I want to note some changes- exactly half of the lines have moved since I posted them. This should be self explanatory, except for the last column. Red means people have been taking the under; blue, the over.The Buffalo line has really moved a ton- they must've read my comment that the Buffalo under might be a good play. Also interesting to note that people have been taking the overs on the two highest totals on the board, New England (12) and Indianapolis (11).

Below are the expected wins for each team, after factoring in the juice. I also adjusted for the fact that after doing this, the average is 8.05 wins. These average to 8. The next column is simply their expected winning percentage. The final column is how often they'd be expected to win 10 or more games. This was initially intended to create very rough playoff odds for each team, but they only add up to 8.77. Oh well.

With this, we can also take another look at the strength of each division:

The West and South really hold the NFC down- the conference's overall average is 7.86, while the AFC's is (predictably) 8.14.

Finally, the odds that the top five teams in the league to 16-0:

Some site should really offer odds on these, or at least on the Pats going undefeated. My guess on the listed odds of that would be 15:1. I'm sure that would get a good deal of action.

30,000:1?

I got an entertaining e-mail this morning regarding a favorite subject of this blog:

"theres a rumor around town here that some guy bet 200 bucks to win 6million before the season started if the Pats went undefeated.On SB sunday Vegas called offering him a million bucks and a trip to Vegas to cancel the bet,2 other people have similiar storys,anyone heard of this or is it a urban legend?"

Obviously I would like to believe this, but there's just no way. This means the bettor in question got 30,000:1 odds on the Patriots going 19-0. That is pretty clearly too high. But what would the odds have actually been?

The over/under for New England regular season wins was 11.5 prior to the season. That is a .719 W%. There is a 0.5% chance that a team with a true W% of .719 wins 16 games in a row. So we're at 200:1 for the Pats to finish the regular season undefeated.

From there, let's figure that a 16-0 team has a 80% chance of winning their first playoff game, 70% for their second, and 75% to win the Super Bowl. That's a 42% shot at sweeping their playoff games.

Combining these two, we arrive at 0.213%, or 1 in 470. So I think it's safe to say that 30,000:1 is a bit high.

The other problem with this is that the proposed buyout makes no sense. The Pats were about 80% to win the SB; offering $1 million against a possible $6 million win is completely unreasonable.

So, nice little story, but there's no way that actually happened.

What Are The Odds: 31-0

This year's Memphis team really seems like the perfect storm. They were pretty solid last year, reaching the Elite 8 before losing to a team that only had two eventual top-four picks.

Obviously, that doesn't make them unique. The difference between Memphis and those other seven teams is that they returned nearly their entire team.

Below are the key losses of each of last year's final eight teams:
Ohio State: Conley, Oden, Cook
UNC: Wright
Georgetown: Green
Kansas: Wright (the other one)
UCLA: Afflalo
Florida: Ha
Oregon: Brooks
Memphis: Jeremy Hunt

Now, this is not to diminish Jeremy Hunt's contributions to last year's team. He did average 24 minutes and 14 points. But his loss clearly didn't hurt Memphis as much as the other Elite 8 teams (with the possible exception of Kansas).

Not only that, but Memphis added the third best freshman in the country this season. Combine that with a 13-0 start and the fact that they play in Conference USA...I think you see where I'm going with this.

So, how likely is it that they finish 31-0, ? Moreso than you probably would have guessed.

Here's their schedule the rest of the way:
1/9 East Carolina
1/12 @Marshall
1/16 @Rice
1/19 Southern Mississippi
1/23 @Tulsa
1/26 Gonzaga
1/30 @Houston
2/2 UTEP
2/6 Southern Methodist
2/9 Central Florida
2/13 Houston
2/16 @UAB
2/20 @Tulane
2/23 Tennessee
2/27 Tulsa
3/1 @Southern Mississippi
3/5 Southern Methodist
3/8 UAB

Projections for each of these games can be found at Ken Pomeroy's site. Needless to say, they are going to be significant favorites in each. They've been favored in every C-USA game since 04-05 (when their conference schedule looked a little bit different), and their two non-conference games (Gonzaga, Tennessee) are both at home.

Pomeroy gives them a 46.9% chance of running the table in C-USA, and since I'm not about to try to handicap games like Memphis @ Rice, I'm going with that for these purposes. He predicts Memphis to beat Gonzaga by 12 (with an 87% chance of winning), which seems about right- I would guess they'll be favored by about that number when they play later this month.

In their other non-conference game, Pomeroy has Memphis beating Tennessee by 11, with a 83% chance or winning. Both of these numbers seem high to me; Tennessee is damn good. On 12/22, the Volunteers won at Xavier by seven. Check out what Xavier has done since that loss:

12/28 Delaware State W 65-33
12/31 Kansas State W 103-77
1/3 Virginia W 108-70
1/6 @Auburn W 80-57

Not bad. I'd guess Memphis will be favored by about 7.5 against Xavier, which gives them a 75% chance of winning. Combine that with 87% in the Gonzaga game, and 46.9% to go 16-0 in C-USA, and they're at a 30.6% to go 31-0.

If there was a line on this (there's not...there should be), here's what it would look like:

Memphis to go undefeated, +215
Memphis to NOT go undefeated, -255

(I understand that there are five other undefeated teams. It's really too early to look at the odds that any of them to undefeated, though. Pomeroy has Kansas rated #1, and they are by far the highest, at almost 12%. But UNC is at only 1.02%, because they're still rated below Duke, who they play twice. And Vandy, Wash St. and Mississippi are all under 0.1%. If one of them still hasn't lost a few weeks from now, then it'll be worth looking at.)

Photo: InsideHoops.