My "prediction" was 0.12°C, which is luckier than I expected. Or maybe I'm some kind of genius.
Here are the linear slopes for recent periods. The statistical uncertainties are for the 95% confidence level:
last 10 years: 0.01 ± 0.10 °C/decade
last 15 years: 0.07 ± 0.06 °C/decade
last 20 years: 0.20 ± 0.04 °C/decade
last 30 years: 0.16 ± 0.02 °C/decade
full record: 0.13 ± 0.01 °C/decade
The last 10 years (i.e. 120 months) has been 0.20°C warmer than the previous decade.
John Christy wrote me to say there is a further complication besides satellite drift, overheating, etc: The raw channel values on the Discover website are those of AMSU channel 5, not for the lower troposphere. So they have some influence from the stratosphere which has to be removed. All of these corrections are, of course, in their publications. Figuring all these out is really hard core, nitty gritty science, and I suspect it was a little maddening at times, too.