The way I did it was simple: I first calculated the year-to-date average (January - July) and then calculated what the August — December average would have to be to keep the 2015 average temperature at or below that of 2014. I then calculated the August — December average for each year since 1970, fitted a trend, and calculated the standard deviation of the residuals. Last, I calculated the expected August - December average for 2015 given the trend and the difference between the expected August - December average and what that average would have to be to keep 2015 from setting a new record. I then used z-scores to calculate the probability that the remainder of 2015 would fall to that level or below.

Annual global temperature according to NASA GISS since 1970 |

Year to date, 2015 sits at +0.82ºC above the 1951-1980 baseline. The average for 2014 was "only" +0.75ºC above the baseline. Keeping 2015 at or below the standard set by 2014 would require an average temperature of at most 0.652ºC for the remainder of the year. So, how likely is that average temperature for the August - December period?

Average August - December temperatures since 1970. |

There is one important caveat that means that I overstate the chance that 2015 will not break the record: I did not account for El Niño years. That was done deliberately. I wanted to be conservative with my estimate. With a strong and strengthening El Niño event in the Pacific that might rival the 1983 and 1998 El Niños, I personally believe that 2015 has a nearly 100% chance of smashing 2014's record baring a major volcanic eruption.