I have collected the "To Win Division" odds from Carib, BetUS*, SportsBetting, VIP, and The Greek. To break these down, I figured it'd be useful to have a single metric to look at. The PFP projections are great, but there's no reason to use them exclusively. Since the Vegas over/under lines are nearly as good a predictor of future performance, they should be considered as well. So I threw both PFP and Vegas' predictions into a regression, with the dependent variable being how many games each team ended up winning. Doing so brings us this equation:
Projected wins = 0.789 + .458 * PFP + .443 * VegasThe two sets of projections end up being weighted almost equally. The fact that both have coefficients of less than 0.5, and there is an intercept value which is the same for each team, incorporates some regression to the mean. PFP obviously does this, but I don't know that the Vegas lines regress as much as they should. Anyway, both the PFP and Vegas projections will be included in the tables, but this number is, as far as I'm concerned, more accurate than either of those taken by themselves.
AFC East
The Greek has the field--the three non-New England teams--at +600. The PFP projection for the Jets does not factor in the Favre addition. You can probably bump them up between a half and full win from that 7.5.
It doesn't matter. Regardless of which column you look at, New England is the heavy favorite. It is possible that they win this thing over 85% of the time, putting some value in the VIP line, but you certainly wouldn't want to have your money tied up for four months to find out. None of these heavy favorites are even close to being viable bets.
I really don't think there's anything here. What, are you going to bet on the Jets and take advantage of the market inefficiency that has caused Brett Favre to be vastly underrated? Let's just move on.
AFC North
That's more like it. I think this one should be pretty obvious. Take it away, Aaron:
Jim (I City): Based on the PFP expected wins, Baltimore should be the favorite to win the AFC North. Do you feel comfortable with that prediction? (This is a team that Miami beat last season).There you go. I don't really have anything to add here. Buy PFP, I guess.
Aaron Schatz: Yes. Very, very comfortable. There are a lot of strong trends here. Just a few:
1) Baltimore's offense recovered only 5 of 25 fumbles last season.
2) You know the concept of "bend but don't break" defense? Baltimore last year had the league's worst "break but don't bend" defense. As Ned Macey and Jim Armstrong explored in an article in PFP 2007, teams that give up lots more points per drive than you would expect from the rest of their performance tend to improve the next year.
3) VERY high injury rate in secondary in 2007.
4) Our projections see Pittsburgh and Cleveland both declining. Cleveland in particular -- remember the Bill James "plexiglass principle?"
5) The teams of the AFC North play a very difficult schedule, with only two games different between each team. Pittsburgh gets New England and San Diego. Baltimore gets Miami and Oakland. That's the division title, right there.
You can bet on either the Texans or Titans winning the division at +485 at The Greek.I think there is some value in Houston here. In PFP, they noted that a lot of the time when a team spends multiple high draft picks on defense, the entire unit comes together at the same time. In theory, that is the case in Houston, with Amobi Okoye, DeMeco Ryans, and Mario Williams. Houston was .500 last year with a defense that finished just 30th in DVOA; with even an average defense, they could contend. In a situation like this, you want as much variance as possible from that projection of 8.2 wins. If they win 12 games half the time, and four the other half, their mean comes out to even less than 8.2, but you'd obviously want to take them at +1500. That is obviously a very extreme example, but I do think it's worth taking Houston at that price.
Oakland at +800 is pretty funny. Beyond the Raiders, the pecking order seems to be pretty clear: San Diego, then Denver, then Kansas City. The only thing that looks somewhat out of place is that +3000 at Carib for Kansas City. The Chiefs obviously aren't very good, but Oakland certainly isn't three times more likely to win the West than they are.*I did this before yesterday's incident, and don't feel like taking them out of the tables.
7 comments:
I believe you are comparing apples to oranges when an apples to apples comparison is available.
we can look at the Vegas win totals as the spread and the divisional odds as the moneyline. The PFP win totals can advise of of what teams to bet but we have to compare the Vegas offerings to decide how to bet them.
Presume for a second that we feel team a should be an 11 point favorite over team B. They are 7.5 point favorites with a moneyline of -250. We clearly have an interest in betting the favorite but just because 11 > 7.5 doesn't mean I'd lay the points in that scenario.
Full disclosure: I have Pitt u9.5 -135 675-500, clev u 8 +110 500-550, and Balt o6 +100 1000-1000.
I don't get your example at all.
I also don't really see how this is comparing apples to oranges either. We are basically regressing PFP's projections to common sense/public perception. In the past, this has made for more accurate projections. I don't see why this wouldn't continue to be true in the future.
I would agree that this post would be essentially the same if I just used the PFP totals, if that's what you're getting at.
I'm willing to grant your assumption that a regression of PFP+Vegas will produce a more accurate prediction than either alone.
The issue I have is whether or not betting the teams to win the division outright is better than betting the win totals themselves. I think the answer is dependent upon the Vegas win odds relative the divisional odds.
Simply put: The regression suggests whom you wish to back, a comparison of the various Vegas/offshore offerings suggests how you should back them.
I can elaborate on the math involved if you wish.
Got it now.
Since I just spent a few moments doing a quick math version to estimate the divisional win odds I might as well post the odds for the AFC North:
1.03 Pitt
14.12 Balt
2.55 Clev
5.23 Cinc
This is assuming 9 wins on average for Pitt, 6.1 for Balt, 8.1 for Clev, and 7.3 for Cinc.
+125 at the Greek on Pitt and +600 at BetUS on Cinc would be the suggested bets IF we believed the Vegas win totals.
We don't.
Now I'll run the regressed numbers:
1.76 Pitt
3.11 Balt
3.73 Clev
4.47 Cinc
Clev and Pitt are now suckers bets while Cinc looks better and Balt is the nuts.
Now, here is the real test: What are the respective EV's for the Baltimore bets presuming the regressed numbers?
I'm going to use over 6 -110 as the team total.
A bet of 1100 to win 1000 on over 6 would have an EV of $483.63.
A bet of $1000 to win $10000 on Balt to win the division would have an EV of $1674.83.
Now, my question to you is this:
Does refusing to believe your own BS cause cognitive dissonance?
The reason I ask is that my style of wagering is such that I use my own methods to find the sides I like but I bet them in a fashion that assumes the main lines to be accurate. In other words, if I find a 3-1 dog that I feel should be a 3-1 favorite (in the NFL), I'll still take +12 points at 11 to win 10 instead of +300 even though the +300 would be higher ev if my analysis of the game were accurate.
What's the best way to deposit money into the Greek from the U.S.? I've tried using CC, E-Check, and InstaDebit and none of them work...
The best way to deposit money at the Greek is to get a Visa all-access card. Go to netspend.com and look up a location in your area that sells said card. Get your card, activate your card and call customer service and in seconds you'll have $ in your account.
I know I sound like a commercial, but it was rather easy and this is what the Greek recommends as the fastest way to fund your account.
Post a Comment