#Democrats retake the lead in SanityIID U.S. #Senate #polling model

A few days, several new polls, and the Democrats retake the lead. But don't get your hopes up too high. The Democrats lead is so slight that it is practically nonexistent. According to the SanityIID U.S. Senate model, the Democrats have about a 52% chance to retain control of the Senate.

Here is the statistical breakdown:

As you can see, the most probable outcome is for the Democrats to retain 50 seats. Of course, it is important to remember that this model both expects Orman to win Kansas and to caucus with the Democrats. If either of those assumptions prove to be untrue, then the Republicans would hold a substantial, but not insurmountable, lead.

As for the individual races that I am following here at SanityIID:

Rasmussen released a new poll that slightly helps Garner (R) in Colorado. Garner's chances increased from 50.9% to 53.7% This is still a virtual tie, but if this keeps up Garner will move out of the margin of error and have a real lead.

GQR released a new poll for Iowa. This poll helps Braley (D) ever so slightly. Current modeling suggests his lead has increased from 59% to 60.6%. While this latest poll helps us gain confidence in our model, it doesn't significantly change the outcome.

Suffolk/USA Today released a poll for Kansas. This poll favors Orman (I) to win. A note of warning: the SanityIID model is showing a data warning on Kansas. This is caused by their being a lack of polls to gain sufficient confidence in the model. While it is true that the model shows him as the current favorite by a significant margin, the lack of polling makes the margin suspect. I would personally give the lead to Mr. Orman, but it is still so close that anything could happen.

Mellman released a new poll for Kentucky. The poll doesn't significantly change anything as far as the statistics are concerned. The model still shows Mitch McConnell with about the best lead that this model can produce. Mitch McConnell's lead decreased from 94.9% to 93%. The only thing significant about this poll is that it is the first time that Alison Grimes has led Mitch McConnell. This poll by itself isn't significant. Still, if this turns into a trend then there might actually be something to watch over the next month -- if the Democratic Senatorial Campaign Committee doesn't pull the finding. It's going to be a difficult decision for the Democrats to decide whether money spent on Kentucky is a waste or not. Unfortunately, my model says that it is.

The latest Mitchell Research poll for Michigan shows that Peters (D) maintains his lead. The statistics provided by the SanityIID model show that his lead has increased to a point greater than the model's margin of error. The model shows that he has increased from 97% chance of winning to a 98% chance of winning. Take this with a grain of salt, however, since the model's accuracy is only 95%.

And finally New Hampshire. A new ARG poll in New Hampshire shows that Shaheen (D) is still winning. Even though the poll shows her ahead, the new poll pushed some old data out of the bottom of the queue. This, ironically, had the effect of decreasing her overall odds since the poll that was replaced showed her as doing better than the new ARG poll. The model is still showing her about a 76% chance of winning with increased short and medium term outlooks.

With a month and a day until the election, we can expect things to begin heating up from here.

SanityIID #Senate #model for October 1, 2014

Welcome to October. Just a little over a month until the U.S. elections.

The SanityIID Senate model run for this morning shows the Republicans solidifying their lead in the short term and making the long term predictions more stable.

As you can see, the most likely probability is that Democrats will win 49 out of 100 U.S. Senate seats, giving control of the Senate to the Republicans. But remember, this model assumes that Orman will caucus with the Democrats. If he decides to caucus with the Republicans due to their taking control of the Senate, then the most probable outcome will be for the Democrats to win 48 out of 100 seats.

Since my last writing, I have added a few more predictions to the SanityIID model. These predictions use the same polling data, but they force the outcome into a win-lose-tie decision. While not as helpful overall as a probabilistic outcome like the above graph, it does help to see where the polling is going over time.

I have broken the data down into long term, medium term, and short term deterministic predictions. Here are the deterministic predictions from the SanityIID model:

		Short Term
	Democrat	Republican	Tie
Assumed Wins	44	45
Model Predicts	4	7	1
Total Seats Per Party	48	52
		Med. Term
	44	45
	4	6	1
	48	51
		Long Term
	44	45
	4	5	2
	48	50

As you can see, using the long term data, the Republicans win 50 seats, the Democrats 48 seats, and two seats are considered a tie. Medium term data gives one of the tie seats to the Republicans for a 51-48 split with one tie remaining. And finally, with the short term data, the final tie is given to the Republicans showing them at 52-48.

While these aren't necessarily the way I expect the elections to come out (I much prefer the probabilistic model as shown by the graph), it does show that the more recent the polling, the more likely they show the Republicans in the lead. Is this a trend? If so, will the trend continue? We only have a little over a month to wait to find out.

US House Data

I currently don't have a model for the United States House of Representatives, but I am beginning to follow the data. As with all my data at the moment, this data comes directly from the HuffPost Pollster. The only real change is that I have put it in a form that is easier for me to follow.

This chart is for all the polls that have been completed in the past month:

SanityIID Senate model for September 27, 2014

The Republicans are coming! The Republicans are coming!

After updating the SanityIID Senate model with the latest polling data as reported from the HuffPost Pollster, the statistics show that the Republicans have expanded their lead this week.

The chart shows that the most likely scenario is for the next Senate split to be Democrats 49, and Republicans 51. But the next most likely split is Democrats 48, Republicans 52.

What happened? In a word, Alaska.

Just over a week ago, Dan Sullivan (R) took the lead back from Mark Begich (D). Since that time, as the polls have continued to come in, Dan Sullivan has continued to hold the lead all the way through our most recent polls.

All the other states have been relatively stable over the last week.

Small update to the SanityIID Senate model

While this isn't a regular day that I update my model, I made some changes and I wanted to post the results.

I added three new states to the model's statistics: Georgia, Kentucky, and Michigan. While these didn't really have any effect on the outcome, periodically there are political reports that discuss these three states. I guess you could say that I added them to the SanityIID Senate model just to be sure I wasn't missing anything.

Secondly, I made a change to the number of Monte Carlo iterations that the model performs. I originally didn't want to do this since it can give the impression of having more confidence in the model than the underlying numbers would justify; but on the other hand, it does keep the probabilities from swinging too far without any underlying polling behind the number.

Here is the histogram from the last run using the data from September 24, 2014:

Republicans still on track to retake Senate

There's only been a few polls updated since the last run of the SanityIID Senate model. The changes those few polls made only had the effect of making the Republicans slightly more likely to retake the Senate.

As for the details; Alaska, Colorado, and Iowa are a tossup. Their mean (average) polling is so close that statistics aren't really meaningful. They are essentially a coin toss at this point in time. The Monte Carlo distribution I am using shows this in the elevated probabilities of 48, 49, or 50 seats.

As for Kansas, it desperately needs to clarify the issue of exactly which candidates are going to be on the ballot. Once that is done, there needs to be more polling if there is going to be better predictions.

As of this writing, my model is showing Kansas with about a 75% chance of electing Orman, but I also believe these percentages to be overly skewed toward him due to lack of polling. There are only four (4) polls that come from different polling places and are relatively new. Since the state of Kansas is still in flux, expect this to change in the coming weeks and months.

Inaugural run of the SanityIID Senate model

My new Senate model is finished! All the data has been updated and the programming is done. So without further ado, here is the results produced by the model:

As you can imagine, being a Democrat, I don't like the output of the model.

Each vertical line represents the probability that the Democrats will hold that number of seats if the election were held today. But remember, the Democrats need to hold 50 seats to be able to retain control of the Senate. And as this model shows, the most likely outcome is for the Democrats to retain only 49 seats.

New Election Forecasting Model!

I am so happy to announce that the mathematics is finished on my new election model. My old model, which worked pretty well in the previous elections, was a deterministic model. The new model uses the same basic inputs, but the output is now probabilistic.

While the math is done, I still have to program all the equations before I can run it for the first time. I am hoping to be done either today or tomorrow.

If you like #statistics, the #Senate races are the best races in town.

The chances of the House of Representatives switching hands to Democratic control are best described as slim to none, with very little slim involved. Additionally, since this is a midterm election year, there isn't anything to look at concerning the President -- statistically speaking. That leaves us with the Senate to play with, and boy what fun we can have with the Senate.

The Democrats currently control the Senate, but there is a fairly good chance that it will switch hands after the midterm elections. This is the kind of year that the terms 'nail bitter' and 'cliff hanger' were made for. The most optimistic statistics that I can find give the Republicans a 30% chance of taking control of the Senate, while the vast majority of statisticians put the Republican's chance of Senate control after the midterms at 50% or greater.

It all comes down to the decisions of a few people in a few states that might be willing to switch their vote from one candidate to another. There haven't been margins this close (that I am aware of) since Gore vs Bush in the Florida elections of 2000.

So, where do we stand?

I have been limiting the statistics and polls that I have been watching. There is a little danger to this in the sense that something could happen in a state that I'm not paying attention to that would cause the probability of overall control of the Senate to switch without my knowledge. That being said, I am currently expect control of the Senate to rest on three states:

Iowa, Alaska, and Colorado.

Despite the most recent poll, my statistics still put Colorado into Democratic hands. Polling data also places Alaska into Democratic hands, but there are so few polls that my confidence isn't very high concerning Alaska.

But the state that is the most interesting is Iowa. Iowa is a scatter-shot pattern of polls all over the place. Even looking at the statistical median, that will help eliminate outliers, Iowa is a coin flip.

So how do all these different states fit into the overall picture of Senate control? As my statistics break down right now, the Republicans should have 50 seats. The Democrats, and the Independents that caucus with them, should have 50 seats (assuming that Greg Orman caucuses with the Democrats). That will keep the Senate control in Democratic hands since Joe Biden, being the Vice President, is the President of the Senate.

If the coin toss of Iowa goes the other way, then that will give the Democrats and Independents a 51-49 majority in the Senate. So despite the close race in Iowa, I give the Democrats good odds of retaining control of the Senate.

#GamerGate and gaming content

Several years ago, when I played World of Warcraft (WoW) as many hours as I could, there was an interesting phenomenon with respect to the way that your player character dressed. The more armor that was visible on a male character, the better protected they were against the dangers of the game. Conversely; the less armor that was visible on a female character, the better protected she was against the dangers of the game. This phenomenon led to the joke that if you ever saw a female character running toward you wearing only pasties and a g-string, you had better run the other way as fast as you could!

If you assumption of a "gamer" is an adolescent, straight boy, then it could be seen as a reward for your character to become more skimpily dressed the more work you put into them. For others, this isn't the case. Blizzard helped tremendously with this problem by allowing players to change the look of the armor they were wearing to suit their own purpose. Suddenly you could be a female, playing a female character, and not look like a teenage boy's idea of a stripper.

It seems like GamerGate (Wikipedia link) might have started from the dislike of Zoe Quinn (Wikipedia link). It seems like after her game was published, she was accused of sleeping her way to better reviews by her ex-boyfriend. Later, the GamerGate cry expanded to Anita Sarkeesian (Wikipedia link) and her video series Tropes VS Women (feministfrequency YouTube channel) exploring the problems of the way women are treated by game designers. And finally, the GamerGate tags began to point toward game reviewers that weren't publishing reviews that "gamers" liked, such as GameSpot's Dead Rising 3 Review Justifies why we need #GamerGate (article link).

I have spent days reading over the material that people would send me on Twitter from the #GamerGate hash tag, but with few exceptions, it all turns out the same. People that like games don't want to take a risk that they won't be able to laugh at transsexuals or flamboyant gay stereotypes. They won't get to view scantily clad (sometimes dead) women under the guise of a video game. Or, heaven forbid, someone might make games for a broader demographic.

But what they fail to see is that there are more people that like to play games than just them. Some of us are women, or gay, or transsexual, or any other minority group. We want games for us where we aren't forced to look at degrading caricatures of ourselves or those we care about. If the aforementioned gamers really want the zenith of their gaming experience to be shanking a prostitute in Grand Theft Auto, I have no problem with that. But for those of us that want something different from our gaming experience, or simply don't want to feel degraded every time we play a game, we deserve the same consideration, development, and thoughtful reviews as the other gamers.

Everyone might not have been happy with the way that WoW developed over the years. I'm sure some people wanted wall to wall breasts every time they logged into the game. Regardless, Blizzard found a way where, if we couldn't all get along, at least we could all tolerate each other and the game.