Friday, March 30, 2012

Why I love Patrick O'Brian:
Reason #... damn, I've lost count.

I finished the 20th Patrick O'Brian novel some time back, and instead of reading the unfinished 21st book - O'Brian died before he got the chance to bring it to conclusion - I decided to start again at the beginning, Master and Commander.  A lot of the characters we know and love are introduced in the first book. Of course there are Aubrey and Maturin, but we also meet Bonden, Pullings and Killick, not yet Preserved Killick, his lovably irascible self from the later parts of the series.
The second novel Post Captain begins on board the Charwell, chasing a French ship in the English Channel, commanded by Captain Griffiths. The French ship is larger by far with more guns, but the Charwell is not alone, though its companion does not sail quite as quickly.  We are several pages into the chase before we get this paragraph, an absolute corker I wish I wrote, but only in my dreams. There are several guests aboard the Charwell, and here is how O'Brian describes three of them.


Captain Aubrey was standing by the aftermost larboard carronade, with a completely abstracted, non-committal look upon his face.  From that place, being tall, he could see the whole situation, the rapidly, smoothly changing triangle of three ships; and close beside him stood two shorter figures, the one Dr. Maturin, formerly his surgeon in the Sophie, and the other a man in black - black clothes, black hat and a streaming black coat - who might have had the word intelligence agent written on his narrow forehead.  Or just the word spy, there being so little room. They were talking in a language thought by some to be Latin. They were talking eagerly, and Jack Aubrey, intercepting a furious glance across the deck, leant down to whisper in his friend's ear, 'Stephen, will you not go below? They will be wanting you in the cockpit any moment now.'


You do not have to read Aubrey and Maturin books if you do not want to.  They are not to everyone's taste.  But know that if you do not, I will be silently judging you.

That is all.

Monday, March 26, 2012

What's wrong with the margin of error?

In media reports of polls, the margin of error is always given, usually an afterthought. Very few news organizations bother telling their readers what it actually means – in the U.S., only the New York Times makes an honest effort about this on a regular basis – and even if every “i" is dotted and “t” crossed, the number can be rendered meaningless by two common factors, time and apathy.

If Candidate A has 42% of the vote in a poll and the “margin of error” is reported as ±4.0%, what this means is that we are 95% confident the true number for Candidate A will be between 38% and 46%. (Why 95% confident, you ask?  It’s the industry standard and it has been forever.)

Since we “catch the fish” about 95% of the time, we expect to be wrong about 5% of the time, almost equally divided between too high and too low.

Here’s the thing. In the polls that have been held since this the Republican nomination race turned into a four person field, the freshest polls for these contests cycle have been awful.

We look at the data from 43 polls from 11 different contests, none taken more than a week before the election date, we see the 95% confidence interval is more like a poorly observed guideline than it is a solid mathematical concept based on the Binomial Theorem. Everyone’s numbers should be captured in the margin of error around 41 of 43 times.  Instead, the most easily predicted candidate Mitt Romney ends up in the 95% interval only 62% of the time and Ron Paul only gets captured in the interval 51% of the time.

What gives?

Mathematicians are loath to scrap the Binomial Theorem, (really, it’s as pretty as a hummingbird, it can't be wrong), so we blame the messy real world for not meeting the standards our lovely calculations promise.  “Data collection bias” is the culprit, the First Great Lesson of the poll taking business, the one that made George Gallup famous and helped kill The Literary Digest way back in the FDR era.

In 1936, The Literary Digest conducted a survey involving 10,000,000 questionnaires, a number no polling company today would dream of trying due to the expense.  They got 2,400,000 replies, still a stunning sample size. Their prediction said Alf Landon would win in a landslide. The margin of error should have been microscopic. Instead, Roosevelt crushed Landon, who carried only Vermont and Maine.

There were two problems. One was the people who were sent the questionnaires in the first place.  The mailing lists were made up of the Digest’s readers, automobile owners and people with telephones.  Cars and phones may not seem like big ticket items now, but in 1936 they were and the group skewed towards the wealthy end.

The other reason was how few people returned the questionnaire, only 24%. Polling companies today have a similar problem, though not on as grand a scale.

George Gallup polled much smaller numbers, predicting Roosevelt’s victory. For good measure, he came up with a sample that would mimic The Literary Digest’s snooty demographic and got numbers within a percent of theirs.  Gallup’s name was made and the publishing company Funk & Wagnalls closed the doors on The Literary Digest in 1938.

How does this relate to polling today?  The important question is “Who actually answers the phone and stays on the line for an opinion poll?”  How many people with caller ID refuse to respond to strange numbers? How many answer then hang up when they realize it’s an opinion poll?  The difference in the numbers for Rick Santorum and Ron Paul is illuminating.

Ron Paul’s numbers get overestimated a lot, the green column above his name.  Rick Santorum has not been markedly overestimated by a single poll yet, but as you can see by the blue column above his name, he has been underestimated a stunning 44% of the time.

It’s as though Ron Paul voters are waiting by the phone, wishing to be asked for their opinion the way a Tennessee Williams heroine prays for gentleman callers.  On the other hand, Rick Santorum voters might as well be at prayer meetings 24/7, politely making sure their cell phones were off.

The freshest polls in the 2008 general election did much better.  If we sift away the undecided from the polls and assume 100% of the vote for the two main candidates (not actually true, but let’s go with it for this conjecture) 94 of 102 results were in the 95% confidence interval. This amounts to only 92%, well within expected values.

Have polls become this much worse in just four years or is it a function of primary vs. general election? My educated guess is that these primaries in particular are tough to gauge because of their nature, so many voters not paying attention until the last minute. While the general election feels like a lingering illness for many voters, polling companies do a better job when public opinion has a longer time to fall into place.

UPDATE: A look at 2008 primaries vs. 2012 primaries.  Super Tuesday was much bigger and much earlier four years ago, and there were four candidates, three of them capable of winning contests and a fourth guy named Ron Paul.  Quite the coincidence!

Sunday, March 25, 2012

Results from Louisiana

When I saw the relative positions of my numbers and Nate Silver's in Louisiana, I had a good feeling for the first time since the primary season started. We were tied exactly on Rick Santorum, we were just a tick apart on both Mitt Romney and Ron Paul, but I was 0.8% on the low side of the Newt Gingrich number and 0.8% to the high side when measuring None of the Above.  Gingrich has been underperforming recently - there's really no chance of him winning and this isn't a crusade for him like it is for Ron Paul - and Buddy Roemer's name on the ballot seemed a good guarantee that None of the Above would be closer to 2% instead of 1%.

My good feeling was correct and my algorithm beat Nate Silver's handily, which gives me a record of 8 wins and 6 losses in 14 contests where we both put up predictions.  Here are the numbers.

Santorum 49.0% (we both were low by plenty, but no gain for either)
Romney 29.5%  (both of us high, I gain 0.1%)
Gingrich 15.9% (both of us high, I gain 0.8%
Paul 6.1%(both of us high, I lose 0.1%)
None of the Above 2.3%(both low, I gain 0.8%)

Hubbard 89.4%, Silver 87.8%

No primaries this Tuesday, but April 3 has winner take all contests in Wisconsin, Maryland and Washington D.C. When it comes to the delegate count, Romney is way ahead with 55% of the allotted delegates so far, with Santorum a distant second with 25%. Mitt has garnered about 40% of the vote so far, a very respectable plurality in four way race. With wise use of his huge advantage in money and organization, he has won eight contests where he got all the delegates, Florida, Arizona, Massachusetts, Idaho, Guam, Northern Marianas, American Samoa and Puerto Rico. Though we are barely short of the halfway point in delegates, Romney's nomination is a statistical lock right now. If Santorum can't break his streak in the winner take all races, there will be little reason to continue.

Of course, there has been very little reason in this race for a very long time.  It's a certainty Ron Paul is in until the bitter end.  We'll see if Santorum and Gingrich also have the stomach for enduring defeat after defeat.

For exciting horse races, you are best served by following the Hubbard vs. Silver match.  Though I currently hold and 8-6 lead, that could vanish if the numbers go bad for me a week from Tuesday.

Stay tuned.

Saturday, March 24, 2012

Two firsts on the same day.

Gallup tracking poll numbers from January 1 through March 23
Even though the national numbers are meaningless in terms of garnering delegates and Gallup asks registered voters instead of likely voters, I am fascinated with the numbers generated by these almost daily updates of the opinions of the Republican electorate.  As you can see, Mitt R-Money (graph in green for cold hard cash) has been the leading candidate for about a month now and Sick Rantorum has rallied and then fallen back.  On the most recent roll that concluded yesterday, Mitt became the first GOP candidate to break the 40% support barrier, garnering 42% of the vote.  Just as significantly in my view is None of the Above (marked in black) went under 10% for the first time ever, coming in at 9%.
This year's Republican contest is dragging on much longer than it did in 2008, when Super Tuesday was a big win for John McCain in early February and everyone but diehard Dr. Ron Paul has folded their tents by early March.  Rick Santorum is expected to win easily in Louisiana today, but because the delegates are handed out proportionally,  Romney may very well pick up enough delegates in Louisiana to get over the halfway mark to 1,144, the number needed to get the nomination.  
It took long enough, but Bill Clinton's line about "Democrats fall in love and Republicans fall in line" is looking more true than it has in a while.

Friday, March 23, 2012

Prediction for March 24:
The Louisiana Primary UPDATE

Yet another primary this week, this time a Saturday affair in Louisiana.  As has happened nearly every time in the past, Nate Silver's prediction model and mine agree on the basic shape of the race (who is in first, second, third and fourth), but we disagree on the precise numbers. Of course, our "agreement" is not a promise, just the best guesses we have based on polls. The polls have missed important events in the past, including wins for Santorum when the polls said he'd be in third and Newt Gingrich's horrible showing last Tuesday in Illinois.
Here are our percentages for the four candidates and the None of the Above tally, which actually has a meaning in Louisiana, because native son Buddy Roemer is on the ballot. My numbers are first and Nate's second in parentheses.

Santorum 44.0% (44.0%)
Romney 29.5%  (29.6%)
Gingrich 18.0% (18.8%
Paul 6.5%(6.4%)
None of the Above 2.0%(1.2%)

Wow, Nate and I have never been closer on our predictions.  Complete agreement on Santorum, one tenth of a percent disagreement on Romney (Nate has the high side) and Paul (I have the high side).  I'm on the low side with Gingrich by 0.8% and on the high side by the same distance with None of the Above, a.k.a. Buddy Roemer+Spoiled Ballots.

I've never said this before the polls closed.  I like my chances this time.

I'll be back Saturday night or Sunday morning to report if I was right.

Wednesday, March 21, 2012

Results from Illinois.

Illinois produces a victory for Mitt R-Money over the field and a win for Nate Silver's prediction model over mine.  He was bullish on Mitt and that made most of the difference. Here are the final numbers.

Romney 46.7% (both low, Nate gains 2.0%)
Santorum 35.0% (both low, I gain 1.1%)
Gingrich 8.0% (both way too high, Nate gains 0.4%)
Paul 9.3% (both too low but we agreed, so no gain for either)
None of the Above 1.0% (I hit it exactly and gain 0.5%
Silver 88.8% Hubbard 88.0% (I lead 7-6 in the first 13 match-ups)

The big story in terms of the polling data is how far the opinion polls misjudged Gingrich's performance.  Every poll had him beating Ron Paul comfortably and getting about 12% to 15% of the vote. Instead he finishes a truly pathetic fourth place.  If Gingrich had a lick of pride, he'd quit now. More than his bad showing, he's had a hard time raising money.  Remember that the Super PACs can buy TV ads, but they can't put gas in the planes or buses or pay the staff.

Later this week, I'll have a post about the margin of error, more correctly called the confidence interval, and how tough it is to get it right in a four person race.  Some companies do better than others, but mathematically they are supposed to be getting it right 95% of the time and it isn't even close.

The next contest is a Saturday primary in Louisiana, another proportional primary.  Polling so far has been sparse, but if there are two more polls reported between now and Saturday, and I expect PPP and Rasmussen will both be on the case, I'll be back with a prediction.

Tuesday, March 20, 2012

Prediction for March 20:
The Illinois primary

Only one primary this evening, Illinois' proportional contest for 69 delegates.  Nate Silver and I do not disagree on the basic shape of the race, but we are a few percentage points off on some of the candidates. Here are the numbers, my prediction first and Silver's in parentheses.

Romney 43.5% (45.5%)
Santorum 32.5% (31.4%)
Gingrich 14.0% (13.6%)
Paul 9.0% (9.0%)
None of the Above 1.0% (0.5%
As you can see, we differ sharply on Romney with Silver being bullish by two percentage points. The less votes we think a candidate is going to get, the less we disagree. These predictions are based on only three polls, two from the modern workhorses PPP and Rasmussen, one from the flaky We Ask America.  (More on the reliability of the different companies this week.)
I'll be back with the results tomorrow. In our first twelve contests, I'm ahead 7-5.

Monday, March 19, 2012

First weekend of March Madness

My friend Art put together a bracket league on Yahoo! this year. Any player is allowed to make three brackets, so he and I made brackets of our own and two brackets based on picks made by people who have made their brackets public.  Art's extra brackets are based on the picks of basketball expert Ken Pomeroy. I used the prognostications of Nate Silver and... 

A well known Islamo-fascist socialist usurper, Barry Soetoro.

Right now, the usurper is kicking everybody's ass hard. He has 35 of 48 picks right, a remarkably good score this year, given how many big upsets there have been. Nate and I are tied for correct picks at 31 of 48.  You get more points for picking correctly in the second round than in the first, so by that way of counting, Silver is ahead of me 40 to 39. But our league also has bonus points for picking upsets and Silver's best guesses lean very much on higher seeds, so with the bonuses the top of the table looks like this.

Barry Soetoro's Halal Bracket: 70 points
Grown Ass Man's Bracket (mine): 65 points
The Pom-Poms of Doom (based on Pomeroy): 52 points
Nate Silver Eats Chalk: 51 points

Art, who is usually pretty good at this, is taking a pounding this year, but there are still four rounds to go and the scoring increases each round.  I'll update the standings next week after the Final Four has been determined.

Friday, March 16, 2012

March 16, 2012 - the day of two fifteens

Quick question. What is the team name of Lehigh University?

I'll give you a moment.

Did you guess Mountain Hawks or did you know?  I'm wagering you didn't guess.

They beat Duke in basketball this evening.

You know, the Blue Devils.

That's a 15 seed beating a 2 seed.  That hasn't happened in NCAA history in... oh, let me look it up... nearly four hours.

I think this now can be counted as the maddest day in March Madness history.

It's Friday afternoon of the first weekend and my bracket is crushed like an itty bitty bug.

This was the scene at Norfolk State on Sunday when they found out they would be in the NCAA Division I tournament for the first time in their school history, a 15th seed against #2 Missouri, a great team who has played great on the road this year. As the cliche goes, they were just happy to be going to The Big Dance. But now at just before 4:00 p.m. Pacific, they have pulled off one of the biggest upsets in history, only the fifth 15th seed to beat a 2nd seed ever.

I had Mizzou winning the whole thing.  My bracket is complete crap now.

But you've got to be happy for these kids.  This is one of those games that will be remembered for a long time, like Michigan losing to Appalachian State in football or more locally, #16 seed Harvard beating #1 Stanford in women's basketball, even more humiliating because it was played on the Stanford campus at Maples Pavilion in 1998.

Whether they make it to the Sweet 16 or not, this has to be the upset of the tournament, one of the biggest of all time.

UPDATE: Okay, this is the biggest upset of the early afternoon.  Got one just as big if not bigger this evening.  See the other post labeled "Two major upsets".

Wednesday, March 14, 2012

No silver mask?
Then it's not Santo.

 This guy is running for president.  Not of the senior class, president of the United States.

A lot of people don't like using his last name.  Since you found this blog on the Internet, you probably know why.  On Twitter, several people are shortening his name to Santo.

I hope this trend will stop.

Santo is a hero.  Santo has a silver mask.  Santo fights the Vampire Women, not The Women Who Want To Go To College or The Women Who Don't Want To Get Pregnant.

The name is taken, please use something else.

Thank you for your kind attention.

Results from Alabama and Mississippi

The polls were all wrong in the Deep South, so neither Nate Silver or I did very well, turning in our worst performances since Iowa, the other huge upset win for Sick Rantorum. That said, in this contest our algorithms are competing against each other, and I pulled off two low scoring wins.  Here are the results.

Santorum 32.9% (I gain 0.6%)
Gingrich 31.3% (I gain 1.1%)
Romney 32.9% (I lose 0.8%)
Paul 4.4% (I gain 0.3%)
None of the Above 1.1% (break even)
Hubbard 85.0%, 83.8%

Santorum 34.5% (I gain 0.4%) 
Gingrich 29.3% (I gain 1%)
Romney 29.0% (I gain 1%)
Paul 5.0% (I lose 1.3%)
None of the Above 2.2% (I gain 0.3%)
Hubbard 85.6%, Silver 84.2%

Simply put, we overrated everybody in the field except Senator Rick.  We also underrated None Of The Above.

The news will be about these two victories and ignore the fact that R-Money picked up more delegates last night than the other candidates by winning big in Hawaii and American Samoa.

Delegates won on 3/13
R-Money 40
Rantorum 35
Old Cheerful 25
Dr. Ron 0
The boa constrictor continues to squeeze. Newt has said many times he won't quit, but even a guy as cheerful as he is has to be getting a little glum by now. This next week has a non-binding caucus in Missouri, a winner take all caucus in Puerto Rico and a direct election next Tuesday in Illinois.  I expect the polling companies to give a lot of data in Illinois and next to none in the other two contests. If that's the case, I'll be back to report on the prognostications a week from now.

Tuesday, March 13, 2012

More fun with the tracking poll.

Gallup tracking poll - Beginning of January through middle of March
The Gallup tracking poll produces almost no information that will tell us who will be the Republican nominee, but because it is updated nearly every day, it gives a great snapshot of the mood of the GOP voters and trends over time.  Let's look at the patterns.
R-Money (green): He rose, he fell, he rose again and fell again, and we have seen his third rise, which may be now entering his third fall.  Even so, he has been the leader in the tracking poll for about 70% of the time and has never been worse than second. Notice that the highest grid line any candidate has passed this year is 35% and R-Money has done it at least once in every month.  It has been really tough for anyone to stay over that level of support for even as long as a week at a stretch.

Sick Rantorum (greyish brown): In the beginning of January, he was a sad joke.  By the beginning of February, he has become R-Money's main competition, even taking the lead for about a week in mid-February.  It's too early to tell, but his numbers are currently rising as Mitt's are falling, so we may see a fifth lead change within a few days. The only thing that might prevent it is two third place finishes tonight in the Southern primaries, which is the general prediction from the polls but close enough that there might be a surprise.

Cheerful Old Newt (grey): December isn't shown on this, when Newt went from being a prohibitive favorite to second place. He peaked again in late January, but since then he has faded to a steady and unimpressive third place. Who knows if GOP votes nationwide will give him a boost based on what should be strong finishes this evening in Alabama and Mississippi.

Dr. Ron Paul (GOLD!): Very steady and consistently in the 10% to 15% bracket, with one small dip in early February. Can't gain traction, won't die out.
None of the above (black): At the beginning of the year, NOTA was up around 25%, but over time it has shrunk and for the past month it has been hovering around 15%. The GOP voter may not be in love, but they are becoming accustomed to the idea that one of these guys is getting the nomination and no white knight is coming to the rescue.

You will see updates of this graph in the future, because it's my blog and I love numbers.

Predictions for March 13:
Primaries in Mississippi and Alabama

There are four contests today, primaries in Alabama and Mississippi and caucuses in Hawaii and American Samoa.  Polling companies only showed interest in the primaries and once again, my model for prediction goes head to head with Nate Silver's, this time in the two Southern states.  Here are the numbers, with my predicted percentage first and Silver's in parentheses second.

Romney 34.5% (33.7%)
Gingrich 32.0% (33.2%)
Santorum 26.5% (25.9%)
Paul 7.0% (7.3%)
None of the Above 0.0% (-0.1%)

Gingrich 31.5% (32.2%)
Romney 31.0% (31.3%)
Santorum 28.5% (27.9%)
Paul 8.0% (7.3%)
None of the Above 1.0% (1.3%)
We agree on the basics, predicting victory for R-Money is Mississippi and Cheerful Old Newt in Alabama, Sick Rantorum in third in both races and Dr. Ron Paul a distant fourth in both.  As the numbers shake out, I'm bullish on Mitt and Rick in Ole Miss with the advantage that None of the Above can't be negative - it's like have a tiebreaker in pocket, while in Alabama, my model is bullish on Santorum and Paul.  Both of us are predicting close battles for first place based on the polling data, and it would not be truly startling if Gingrich or R-Money swept the South tonight.
I'll be back tomorrow with the results.  Currently, Silver and I are tied at 5-5 in the ten contests where we both made predictions. If we take the average of our ten predictions so far, Nate is ahead 89.6% to 89.3%.

Thursday, March 8, 2012

So how is the election really going, Matty Boy?

Well, hypothetical question asker, here's the long and short of it. 

Mitt R-Money will be the Republican nominee on the first ballot, barring some stunning disaster.

After Super Tuesday, pundits were saying it wasn't impressive and he can't seal the deal because he didn't run the board.

That's just silly.

So far in essentially a four person race since things really started to matter, R-Money has garnered 40.5% of the popular vote and much more importantly, 55% of the delegates. He needs about 700 of 1500 remaining delegates to get the nomination, way less than half. Sick Rantorum would have to get on a massive hot streak and win 1,000 of the remaining delegates to pull out the upset, which means two out of every three from now on.

Ain't gonna happen.

And Cheerful Old Newt and Dr. Ron Paul have a combined chance of... wait, let me check the numbers... epsilon to win.

For non-mathematicians, epsilon means a number ridiculously close to zero that can't be said to be exactly zero.  All they need is toe tags, because both of them are dead meat.

Ron Paul doesn't quit because this is a crusade for him, not an election.

(photo by Vincent Yu/AP)

Newt doesn't quit beacuse of this odd looking grease weasel, billionaire casino owner Sheldon Adelson. Maybe it's just because I've grown up online reading Princess Sparkle Pony, but being a billionaire should mean getting better skin and hair care than this.

Adelson is Newt's Super PAC sugar daddy, so getting ads on the air is no problem. But R-Money has a Super PAC, too, and those ads effectively killed Cheerful Old Newt's chances of bouncing back. (Some observers say the moon base was Newt's true undoing, and that is certainly a major contributing factor.) For whatever reason, Adelson has taken a disliking to Sick Rantorum and promises to throw his support behind R-Money once Cheerful Old is out of the picture. In other words, instead of a knockout punch, the R-Money campaign is more like a boa constrictor, slowly squeezing the air out of its victims.

When it comes to the general election, this primary season hasn't been much help to R-Money's chances, but without a third party right wing candidate I think Mitt has a chance, though not even money at present. For exact numerical stuff, this is way too far away from November to give numbers that have any meaning at all.  Heck, I'm not confident about the predicting the primaries next Tuesday yet. But checking all the background information, from Gallup tracking down to Rush Limbaugh's woes - he has been a serious foe of Mitt's candidacy - R-Money is the Republican standard bearer whether they like him or not. Right now, he has to prove he's a better candidate than Michael Dukakis, let alone President Obama.

It's going to be a long year and I don't just mean 366 days.

Maru levels... Dangerously low!

Maru + Enclosed spaces = fun for the whole family.

Wednesday, March 7, 2012

Results of Super Tuesday:
Predictions in Ohio, Georgia and Tennessee.

The final numbers are in for all ten races from last night, but there were only three states where there was enough data for both my model and Nate Silver's to predict the results candidate by candidate.  Without further ado, here are the numbers.


Romney 38.0% (both low, Silver gains 1.1%)
Santorum 37.1% (both low, I gain 0.2%)
Paul 9.2% (both high, Silver gains 1.2%)
Gingrich 14.6% (both high, Silver gains 0.4%)
NOTA 1.1% (I'm low, Nate's high, he gains 0.5%)
Final: Silver 91.6%, Hubbard 88.6%

Romney 28.1% (both high, I gain 1.1%)
Santorum 37.2% (both low, I gain 2.4%)
Paul 9.0% (I'm low, Silver's high, Silver 0.4% closer)
Gingrich 23.9% (I'm low, Silver's high, I'm 0.5% closer)
NOTA 1.7% (I'm high, Silver's low, I gain 0.5%)
Final: Hubbard 91.8%, Silver 87.4%

Romney 25.9% (both low, Silver gains 1.0%)
Santorum 19.6% (both low, I gain 0.3%)
Paul 6.6% (both high, I gain 0.7%)
Gingrich 47.2% (both low, I gain 0.6%)
NOTA 0.7% (both high, Silver gains 0.8%) 
Final: Silver 94.8%, Hubbard 94.6%

Long story short: His model whipped mine but good in Ohio, mine whipped his in Tennessee and he beats me by the narrowest possible margin in Georgia.  (The tenths place is always even, because any tenth of a percent a prediction is too high for one candidate must be balanced by a tenth of a percent low someplace else.)  This brings the overall prediction battle record to five wins and five losses for each, a flat-footed tie.

No day from now on is going to have ten contests again and I don't know how avid the polling companies will be about covering the next contests. There are four contests this Saturday and four more next Tuesday and so far I have seen zero polls concerning any of them.  There might be some polling in Mississippi and Alabama, possibly in Kansas - though it's a Saturday caucus, which is kind of a double whammy - but I don't expect much data from the far-flung contests in Guam, Northern Marianas, Virgin Islands, American Samoa or even in Hawaii.

I will let you know the next time both Nate and I put up predictions head to head.

Tuesday, March 6, 2012

Predictions on Super Tuesday:
Georgia, Ohio and Tennessee

The crystal ball is back and we have three key races where lots of polling companies gave us lots of data, much of it on this last weekend before the primaries.  There is enough data in Georgia, Ohio and Tennessee for my prediction model to go head to head with Nate Silver's in those states. The seven other races - Alaska, Oklahoma, Virginia, Vermont, Massachusetts, North Dakota and Idaho - either had no polling or too little for my tastes.

Here are the predictions in the three states, my number given first and Nate Silver's in parentheses.
Romney 35.5% (36.6%)
Santorum 34.5% (34.3%)
Gingrich 17.0% (16.6%)
Paul 12.5% (11.3%)
NOTA 0.5% (1.2%)
Silver and I do not disagree on position of the four candidates in any state. We also agree that Ohio is relatively close, but late polling makes R-Money a slim favorite over Sick Rantorum. Nate's position makes him bullish on Mitt and None of the Above, and there are large gaps in our predictions of Mitt and Ron Paul.
Santorum 35.0% (32.6%)
Romney 30.0% (31.1%)  
Gingrich 22.0% (26.5%)
Paul 10.0% (9.7%)
NOTA 3.0% (0.1%)
Another close race, but Silver's model makes it even closer than mine does. We have large disagreements on everybody but Ron Paul, so the result of this one is likely to be a landslide one way or the other in terms of prediction numbers.

Gingrich 46.5% (45.9%)
Romney 24.5% (25.5%)
Santorum 19.0% (18.7%)
Paul 8.0% (8.7%)
NOTA 2.0% (1.2%)

Again, my prediction model and Nate Silver's agree on the basic order of things and there are big gaps between the candidates, so any change to Newt > Mitt > Rick > Ron will be a big surprise.  I'm bullish on Cheerful Old Newt, Sick Rantorum and None of the Above.  The gaps in our predictions are relatively small, so this could be a close race between Nate and me, though it should be a landslide for Gingrich in the state he used to call home.

I'll be back tomorrow to report on results.  After seven matches, I lead 4-3.

Monday, March 5, 2012

Okay, Matty Boy, what was your problem with the Master and Commander movie?

Hypothetical question asker! So good to have you back, it's been far too long.

The easy guess would be Russell Crowe, but honestly, I think he is just about the perfect Jack Aubrey. Brave, honest, clever as a tactician but not in the use of the English language, the movie got Jack down to a tee.

The problem was Stephen Maturin.  We see him as a physician and a naturalist, but not of him as an intelligence agent.  I understand cutting that part out, it could easily take up too much time, though I missed it dearly. The problem for me was Paul Bettany, far too tall and ten times over too pretty.

I was trying to think of better casting and the best I could come up with was David Bamber.  My favorite roles of his are as the obsequious Mr. Collins in the BBC mini-series version of Pride and Prejudice and his turn as Cicero in HBO's Rome, who stands against Caesar on principle but is not soldier enough to stand against him on the field of battle. Mr. Collins puts him in almost the right costume, and his Cicero shows him very capable of playing a beautiful wordsmith but not a beautiful man.  This haircut does an excellent job of hiding his ears, which stick out like the open doors on a 1957 Buick Special.

Of course, there is nothing like reading Patrick O'Brian, so it's just as well they made the one film to take a not particularly accurate snapshot from the twenty novel story of Aubrey and Maturin, and leave the rest of it to be enjoyed as it truly should, on the printed page.

Two at the same time? UPDATE

Okay, thanks, I was already awake.

About ten minutes ago there was a quake I felt here in Oakland.  There was that little start like someone unloading something heavy from the elevator down the hall, and then there were some serious left to right shakes.  After about three, I decided to stand in a doorway.  It was over soon enough.

Of course, I googled "seismic activity" and up came a 4.3 centered in Richmond, about 13 miles north of my house.  That was about five minutes after the shock, but a few minutes later, there was a revision.  There were two quakes, one in Richmond and one in Tiburon, over in Marin County, and both were measured at 4.0.

A little math help.  Two 4.0 quakes do NOT add up to an 8.0 quake, thank Odin, Vishnu and the little baby Jebus. The Richter scale is logarithmic and twice as strong only increases the reading by the log of 2, which is about 0.30103... That's why two 4.0 quakes would read the same as one 4.3 quake.

I don't ever remember two noticeable quakes centered so close to one another so close to simultaneous, though I will admit I don't keep close records.  With as many faults and fissures as there are in the Bay Area, I'm a little surprised that movement in one fault line wouldn't nudge another into a quake more often than this, if that is in fact what happened.

UPDATE: In the comments, Susan S. said it had been changed to one quake, so I re-Google seismic activity and as of noon, it's two quakes, eight seconds apart, the first one 3.5 and the second 4.0. Both are centered in Richmond, so not one fault triggering another.

Sunday, March 4, 2012

No prediction based on a single poll.

Quick update: I didn't publish a prediction in Washington state and neither did Nate Silver. Had I used my model with only the PPP poll, I would have been about 77.1% correct, a grade of C instead of the B+ to A- grades I expect when more data is available.  The PPP poll had Sick Rantorum beating Dr. Ron Paul - the only candidate with a key to the TARDIS - handily. Instead, it was Paul barely out-polling Sen. Rick and Gingrich completely out of the running, his usual outcome.

There will be enough data for some predictions before Super Tuesday.

Saturday, March 3, 2012

No forecast in Washington caucuses

Washington state has a Republican caucus today, but I am making no prediction, and neither is Nate Silver.  Only PPP did any polling there, one poll this week and one two weeks ago.  They disagree radically, the early poll at the height of Santorum's national popularity showed he was ahead while the poll this week said Romney was the leader.

It's not that I don't trust PPP, but more that I don't trust a single fresh poll, especially when it concerns a caucus with a miniscule turn-out. Wyoming had a caucus on Wednesday, February 29, but I didn't see a single poll published anywhere. There are ten contests this Tuesday but several are still with no polling data whatsoever. North Dakota, for example, has a non-binding caucus.  There is no reason to keep track of what they do at all.

I am not yet sure of how many of Tuesday's contests will have enough data for me to make a prediction.  It all depends on how many polls get published in each state between now and Tuesday morning. The two big prizes are proportional primaries in Georgia and Ohio.  Recent polls show Gingrich comfortably ahead in the state he used to represent and Santorum is ahead in all the recent polls out of Ohio. The next biggest contest is Tennessee, which has only one poll so far this week.

Let me set a threshold of data for my system.  I need at least three polls from three companies in the last week and at least one of those has to be from the last three days before the election.  That means for Tuesday contests, I want a poll that finished up on Saturday, Sunday or Monday and at least two more polls that concluded Wednesday, Thursday or Friday. There are ten contests on Tuesday.  I'll be surprised if I have a prediction in more than five.

Stay tuned.

Thursday, March 1, 2012

If you have nothing nice to say, come sit next to me.

Andrew Breitbart is dead at the age of 43.  That is stunningly young.  From what I have seen of his life, he earned it.

There is a custom that you are not supposed to speak ill of the dead.  This is not a custom I follow.  One of my first stops on the very old World Wide Web was a newsgroup called alt.obituaries, where people from all English speaking nations shared the stories of the recently dead, some you knew and some you didn't. Sometimes we praised, sometimes we buried.  Sometimes we dug the corpse up and pissed on it for a while, then re-buried it.

All this was permitted.  It was not permitted to harp on the living, unless there was solid evidence that they might not be living for long.

Several stories about Breitbart written today say he was fearless.  This is an obvious fucking lie.  He was shameless and there is a world of difference.  When you lie about someone and that person loses their job and the lie is proven a dozen times over to be a lie, holding to your version of events is not brave, it is obscene. Shirley Sherrod still has a court suit against him.  Whether she pursues it now that he is dead and she is taking money from his heirs is completely up to her in my book.  I will not think less of her one way or the other.

Breitbart lived on booze and vitriol.  I rather like the former but I am doing my best to cut back on the latter. In some ways, I look at the modern conservative movement and the Republican Party/Tea Party as I would look on an infectious disease, interested in its progress (or lack of same) but doing my best not to turn to anger as I watch it eat everything in its path, including its own fellow travelers.

For me, Breitbart is like Christopher Hitchens without the talent. I won't miss either of them very much. I am much heartened by how many people assumed that the first reports were a hoax, some attempt to goad his political opponents into action his political allies could find objectionable.

If one ally of the happily now dead scumbag Breitbart shows up here, I challenge them to find an untrue sentence in what I wrote.  It may not be true for his immediate family, but for the vast majority of humanity, it is a blessing he was erased so quickly.

Here endeth the lesson.

A last look at February

Gallup national tracking for February
It's March now and winter is starting in the Bay Area.  Across the country, Republicans are still not completely sure who they want as their standard bearer.  As of the 29th, R-Money is now at the 35% mark while Sick Rantorum is under 25%. You can see that as Senator Rick has fallen since the middle of the month, the one who has benefited most  Governor Mitt.

There's a passel of primaries and caucuses on Tuesday, but the next really interesting thing will be Newt or Rick throwing in the towel. I expect R-Money will fall behind once it becomes a two and a half person race, but I also expect him to come back.

It's still a crazy crap shoot.