| |
Wednesday, September 29, 2004
The Randomness of Polls
One of the more poorly understood things about polls is how random they really are. For example, the obvious purpose of polling is to estimate as closely as possible what the outcome of the vote is going to be. But suppose we did know the outcome of the vote, because it was the day after the election. Well, then we could look back at the prior day or two's worth of polls and judge them by how well they stacked up against the actual vote.
I set up a simulation of, for example, a Pennsylvania poll with 726 voters. It is assumed for the simulation that if you could poll every likely voter in Pennsylvania the result would be Kerry 46%, Bush 44, 2% others and 8% (of the likely voters) not voting. Then I created a poll of 726 likely voters (the number used in a recent Quinnipiac Poll) and via random numbers compared to those percentages, allocated each voter to a candidate, and compared their polling percentages. Then I ran the simulation 50 times. The results may seem a bit surprising. Kerry's polling varied from a low of 41.6% in year 33 of the simulation to 50.4% in year 39, while the President's numbers varied from 40.1% in year 39 to a high of 48.5 in year 29. In fully 1/4 of the polls the President was leading, when, based on the assumption, Kerry should be leading by 2%.
The polling becomes more accurate in the aggregate; after 50 runs thorugh the 726 likely voters, the overall average was 46.02 for Kerry, 43.98 for Bush. Of course, the problem is that nobody wants to actually poll some 36,000 people, so they use the smaller numbers, which are subject to wild fluctuations. For example, suppose one poll had shown Kerry ahead by a full 10 points, and then the next week showed Kerry losing by almost 7 points; there would be a strong presumption that the campaign was falling apart, and every mistake that Kerry had made would be subjected to scrutiny in search of the cause, and yet it is completely within the realm of possibility that no movement actually occured in the underlying public opinion; both those results came out in the 50 runs I made, when the assumed public opinion was Kerry by two.
Sunday, September 26, 2004
Digital Brownshirts News Is On The Air
Friday, September 24, 2004
Another Dubious Kerry Tale
(Note: This was originally posted at Kerry Haters and is being posted here since I forgot and left this blog in the http section of Captain's Quarters comments. If you'd like to read it over there, here's the link.)
From a New York Times article on Kerry's loss in a congressional race in 1972:
Months later, a sympathetic profile in The Worcester Telegram summed up Mr. Kerry's predicament on a day he had been doing errands, ran out of gas, then headed to the car wash. "All of a sudden, water started washing over my shoes!" Mr. Kerry explained in the article. "The car was filling up with water! Water was coming out of every orifice! I could see the headlines: 'John Kerry, Former Congressional Candidate, Drowned in Car Wash.'"
Okay, we know the pattern of Kerry's tall tales by now. Anybody remember an incident like this in a movie or TV sit-com? Because that's where this came from.
That's one thing the two senators from Massachusetts have in common--they've both been in a car that was filling up with water.
Thursday, September 23, 2004
Oh, Boo-Hoo-Hoo!
Wow, this is truly a sad story:
Relatives of Zeinab Abu Salem had little time to absorb the shock after the 18-year-old blew herself up in a suicide attack in Jerusalem.
They rushed instead to empty the family home in the Palestinian refugee camp of Askar near the West Bank city of Nablus, expecting Israeli bulldozers to soon come to demolish it.
Those mean Israelis!
Hat Tip: Best of the Web Today
Monday, September 20, 2004
Blog Growth
Over at my other blog (which will, I hope, become defunct in about 43 days), I've been eagerly anticipating breaking into the Top 100 blogs on the Truth Laid Bear traffic rating. That hasn't happened yet, in fact we are slipping down every so slightly (currently at 129; we were at 123 about 10 days ago).
This doesn't mean that the traffic is not growing at that blog. In fact, we're up over 20% in the last couple weeks. It's just that the blogosphere is growing just as fast, if not faster. I took a look at what the daily traffic was at various rankings over the last few weeks. The results were somewhat staggering: The traffic at blogs is exploding.
Consider the following. On September 2, the #1 ranked blog under the TTLB system (either Instapundit or Daily Kos) had 249,000 visitors daily. Today, the #1 ranked blog had over 300,000 visitors, a 20% increase in traffic in only 18 days. And that is actually one of the smaller increases over that period. The 50th ranked blog on 9/2 had about 5,300 visitors; today's 50th ranked blog had 8,800 visitors, a 65% increase. On average, it looks to me like the average increase has been about 33% in eighteen days!
That's staggering. If any major newspaper had been able to increase its circulation by 33% over the course of a full year, its staff would be getting million dollar offers from other papers.
Tuesday, September 14, 2004
Why Week One Doesn't Matter--Correction
After I wrote yesterday that the NFL has wised up and given the crappier teams home field advantage I realized that I'd better check that ( check first, then post!--ed). Sure enough, it turns out I was all wet on that point. The teams that had home field advantage in the first week this season were a combined 133-123 in 2003; a little bit better than average, in other words.
The real reason that Week One doesn't matter is that the NFL makes the best teams play the best teams and the worst teams play the worst teams early in the season. There were 14 teams with winning records in 2003; in their first games of 2004, they played teams that went a combined 132-92 last year. There were two teams that were 8-8 last season; they played teams that went a combined 16-16 last year. And there were 16 teams that had losing records in 2003; in their openers of 2004 they played teams that went a combined 108-148 in 2004. Note: This does not apply to every team. Arizona, which went 4-12 in 2003 played the 12-4 Rams, while the 4-12 Giants were carved up by the 12-4 Eagles. But still, on average the crummy teams play the crummy teams while the chalk plays the chalk.
Sunday, September 12, 2004
NFL Week One--Updated!
Week One doesn't matter much at all. I talked about this after week one last year. It's counterintuitive in such a short season, but last year just cemented my opinion. The 16 teams that went 1-0 last year went a combined 111-129 against the rest of their schedule, while the 16 teams that lost to start the season went, by definition, 129-111. Last year's Super Bowl champs, the New England Patriots, lost their 2003 opener 31-0 to the Buffalo Bills, who went 5-10 the rest of the way.
Why does this happen? The main reason is the NFL has wised up and tended to give crappier teams the home field advantage early in the season except for the poor Cardinals who of course never get to open at home given the guaranteed 100+ degree temperatures in Tempe. The Cardinals have been in Tempe for 16 seasons and they have never gotten to open the season at home; for most of those years the NFL never got them home before the third week.
So don't get carried away with this week's results; the history of the last several years says that your team will win more this season if they lost this weekend.
Update: Peter King has a similar take:
Lest you take the first game of the season as some indicator of January success or failure, let me remind you of these scores from opening weekend 2003:
Buffalo 31, New England 0.
Atlanta 27, Dallas 13.
New York Giants 23, St. Louis 13.
Pittsburgh 34, Baltimore 15.
And, on Monday night, Tampa Bay 17, Philadelphia 0.
The Bucs went on to win seven, Buffalo and Pittsburgh six, Houston five and the Giants four. The five losers made the playoffs. The Patriots, you might recall, had a pretty decent year. After such decisive openers, the fortunes of all 10 teams turned radically in the next four months.
Tuesday, September 07, 2004
|
|
