While I hesitate to question FiveThirtyEight, who's 2012 Senate predictions were only second to Sam Wang's Princeton Election Consortium, I need to point out a false claim when I spot it. While I am not sure who sent out the Tweet below; the individual apparently doesn't understand probability.

The current forecast on their website is 53.3% in favor of the GOP. This is a toss-up. It is not much different than flipping a coin. The claim that 53.3% is only "nearing" a toss-up demonstrates a lack of understanding. 538 says that they run "thousands" of simulations to reach this figure. Let's assume 2,000 simulations. This means in 2,000 simulations, DEMS have the majority 934 times and the GOP would have the majority 1,066 times. If you flip a coin 2,000 times, this is a a possible outcome.

According to a paper, the lead author being Persi Diaconis, the odds of a coin landing "heads" is 50.8%. This is equivalent to heads being up 1,016 and tails being 984. This is a difference of 51/2000 in comparison to a 53.3% probability. So, if you assume Republicans are "heads" and Democrats are "tails": Republicans winning the Senate this year are 2.55% better than flipping a coin and it landing "heads" up. That is roughly 2σ away from the coin toss experiment. This is within the MoE at a 95% CI.

The tendency is for one to read 60% or 55% as 100%. We see a favorite, and we assume the favorite must win. The favorite, loses a certain percentage of the time. If I were to tell you that you can invest money in my company, but I can only guarantee you 53.3% that your investment will make money: you would be well advised to invest elsewhere even though it would be better than another investment offering a 46.7% guarantee. Just because 53.3% is better than 46.7%; this does not mean that 53.3% is necessarily very good.

It also doesn't help when other sources in the media are using misleading terms such as 538 did in this Tweet. The Washington Post's projections today listed the Democrats as "likely" to retain control of the Senate. They list DEM probability at 50%. They have said the same thing with 51%. While this is likely just the individuals running that page are lazy, or careless: it does add to the public misconception.

While I don't believe that 538 actually thinks 53.3% is not a toss-up; it is a point that needs clarification. PEC has had DEM control at 70% for the past few weeks. This is only a 1/5 variation from 50%. Though, his 70% is misleading in that it is Dems+Independent control. He does not factor in the possibility that Orman caucuses with the GOP. This is why on 538, the probability of R's having 51 and the probability of D's having 51 are both higher than the Senate being split 50-50. They work on the assumption that Orman would, more often than not, chose to become member 51 of a caucus rather than 50.

According to a paper, the lead author being Persi Diaconis, the odds of a coin landing "heads" is 50.8%. This is equivalent to heads being up 1,016 and tails being 984. This is a difference of 51/2000 in comparison to a 53.3% probability. So, if you assume Republicans are "heads" and Democrats are "tails": Republicans winning the Senate this year are 2.55% better than flipping a coin and it landing "heads" up. That is roughly 2σ away from the coin toss experiment. This is within the MoE at a 95% CI.

The tendency is for one to read 60% or 55% as 100%. We see a favorite, and we assume the favorite must win. The favorite, loses a certain percentage of the time. If I were to tell you that you can invest money in my company, but I can only guarantee you 53.3% that your investment will make money: you would be well advised to invest elsewhere even though it would be better than another investment offering a 46.7% guarantee. Just because 53.3% is better than 46.7%; this does not mean that 53.3% is necessarily very good.

It also doesn't help when other sources in the media are using misleading terms such as 538 did in this Tweet. The Washington Post's projections today listed the Democrats as "likely" to retain control of the Senate. They list DEM probability at 50%. They have said the same thing with 51%. While this is likely just the individuals running that page are lazy, or careless: it does add to the public misconception.

While I don't believe that 538 actually thinks 53.3% is not a toss-up; it is a point that needs clarification. PEC has had DEM control at 70% for the past few weeks. This is only a 1/5 variation from 50%. Though, his 70% is misleading in that it is Dems+Independent control. He does not factor in the possibility that Orman caucuses with the GOP. This is why on 538, the probability of R's having 51 and the probability of D's having 51 are both higher than the Senate being split 50-50. They work on the assumption that Orman would, more often than not, chose to become member 51 of a caucus rather than 50.