When to fade a sharp

Fair warning; I'm bringing some sports betting terms into this post because these were the terms that I felt best articulated the credit downgrade issued by S&P on Friday and Nate Silver's resulting post.

In sports betting a "sharp" is someone who knows what they're doing, they're the smart money. If you've seen the movie "Casino," Robert De Niro's character is a sharp.

Betting with the sharps is a good way to make money in sports betting, but it's not that easy. Some sharps are so well known and so good that when they make a bet the whole market reacts, and a flood of money will come in on the same side as the sharp. Most books will watch these sharps moves as well and adjust their lines accordingly.

To alleviate this, many sharps, because their bets are so influential, will purposely bet a line one way early in the week in an effort to push the line in a direction to their liking and then at the last minute money will pour in on the other side of their initial wager.

Because of these reasons betting with the sharp's isn't always easy and for many sports bettor's, finding a sharp to bet with isn't nearly as profitable as finding a square to bet against.

Who's polling The Minnesota Governors race

Only four pollsters; Rasmussen, SurveyUSA, Princeton Survey Research Associates (PSRA) and Humphrey Institute, have polled the Minnesota Governors race this year and with under a month left before the election it's unlikely that many more will jump in. In 2006 these same four pollsters (StarTribune had in-house polling in 2006, now they contract with PSRA) were joined by Mason-Dixon, St. Cloud St. and Polimetrix, so it's possible that one of them or someone else entirely will drop a poll on us but for now we're just going to look at who's been polling it so far.

A few notes on terms and formatting first.

PIE is Pollster Induced Error, a metric developed by Nate Silver, it's a measure of how much error, above average, a pollster contributes based on an analysis of past polls.

Past election results are listed as follows:
Race Polled: Last polling result (Actual results) Margin Difference
The margin difference is the polling margin minus the actual margin.

Rejected absentee ballots are being counted

You can watch the action live on The Uptake. Meanwhile Nate Silver is up with an analysis of the 953 rejected absentee ballots that will be counted; he's predicting that Franken will net 31 votes, extending his lead to 80. He throws cold water on the idea that Coleman has excluded many more ballots than Franken from the counting.

The apparent edge for Franken comes as each campaign has had essentially a unilateral veto to prevent any particular absentee ballot from being opened and counted. Initially, based on a review of the process in Democratic stronghold St. Louis county, it appeared that the Coleman campaign was being far more aggressive than the Franken campaign in exercising its veto. Information gathered by additional counties by the Star Tribune, however, suggests that this may not be the case, and that the Franken campaign has been at least as aggressive as Coleman in exercising its veto.

Also the Supreme Court is expected to rule on the Coleman campaigns request to put a halt to the whole thing in order for them to change the rules to their liking. Lots of action going on today.

What's ahead for the recount?

Things will start to get interesting again in the Senate recount on Friday. That's when the Canvassing Board will meet to decide on the fate of rejected absentee ballots as well as figure out what to do about the 133 missing ballots in ward 3, precinct 1. Also up in the air, the 12 uncounted absentee ballots from Hennepin county found during the search for those 133 missing ballots. Hey, we're not Florida, we just visit in winter.

What to do with the pile of improperly rejected absentee ballots will prove the most consequential decision. It seems likely, as there is precedent for it, that the election night machine results will be used in place of the 133 missing ballots in Hennepin County. As for the 12 uncounted absentee ballots, if they're counted the most votes either candidate could realistically hope to net from those would be 3 or 4. If the candidates garner the same support as they did in Hennepin County as a whole (using election night results) Franken would gain 6 votes, Coleman 4.3 and Barkley 1.5, for a net Franken gain of 1.7 votes.  

As you'll recall Secretary of State Ritchie asked all of the election officials to sort through rejected absentee ballots and divide them into five piles, with the so called fifth pile representing absentee ballots that were mistakenly rejected. Right now the Star Tribune is reporting that election officials have found 244 improperly rejected absentee ballots, with only 14 of the 87 counties done sorting. However some counties aren't sorting at all. 

Ramsey County Elections Manager Joe Mansky had this to say about why his county had decided against sorting absentee ballots:

"(Our) advice is, we probably need some direction from a court."

Only 56 counties are listed on the sorting schedule on the Secretary of State's website. Five, it appears, have decided not to sort their rejected absentee ballots at all, being listed as "will not participate." That leaves 16 counties that may or may not sort their ballots. With all that said the majority of counties are sorting their rejected absentee ballots and 14 counties have completed this process representing 15% of the pre-recount vote total. From this information we can extrapolate how many ballots will be in the final total under a few different scenarios.

First we need to figure out the rate of improperly rejected absentee ballots, or IRAB's. To do this I will only use data from counties that have completed sorting their rejected absentee ballots, as those counties in the process of sorting don't provide data about how many ballots they've sorted through, only how many IRAB's they've found. In the counties that have completed their sorting 150 IRAB's have been found, out of 45,241 total absentee ballots, for an IRAB rate of .0033.

Scenario 1 is that all improperly rejected absentee ballots are counted in all counties. There were a total of 292,535 absentee ballots counted in the original count, apply our IRAB rate and we get an additional 970 ballots being included in the total under this scenario, which is close to the Franken campaigns estimates that there were 1,000 such ballots. Scenario 1, however, looks to be the least likely scenario barring court action, though under those circumstances it might be the most likely outcome.

Scenario 2 occurs if we accept that only the counties listed on the sorting schedule provided by the Secretary of State will actually sort their ballots, and that those sorted ballots will be counted. Under this scenario 676 rejected absentee ballots will get added to the ballot totals. 

Scenario 3 is that the Canvassing board decides not to count the IRAB's at all. This is the most likely scenario, mainly because the other two scenarios are unlikely. Since not all the counties have sorted their rejected absentee ballots scenario 1 is physically impossible and I don't think the ad-hoc approach of scenario 2 will fly with members of the Canvassing board. For those reasons they'll probably decide that it's better to leave it to the courts to sort this matter out.

In an analysis of this issue a few weeks ago Nate Sliver predicted that if there were 1,000 IRAB's, Al Franken would gain 25-100 votes. The current IRAB rate suggests that will be the case, meaning this will only grow larger as an issue no matter what happens. If the Canvassing Board decides to count these ballots expect the Coleman campaign to go apeshit.

So my predictions are; the election night totals get used in ward 3, precinct 1, the 12 uncounted absentee ballots will be counted and what to do about the IRAB's will be left up to the courts. If all of that is correct then we're back to waiting for the challenged ballots to be adjudicated to learn who the winner is of the pre-trial phase.