You don't have provided any rational argument for you criticism, when you were proved wrong anon when the time series started, you changed the goalposts.
As I'm very tired of deniers like you, have a nice day.
Funny how your criticism of me consists of 2 things :
1) I have a predetermined conclusion
2) and it's the wrong one (obviously you have a predetermined conclusion as well and it doesn't seem to tolerate people of different opinions)
My big argument is this : why don't we look at the main point ? Obviously that last year of data doesn't fit the data series. Obviously it's an outlier (just freaking look at it). That means it should be treated as an outlier : don't base conclusions on it, and figure out what has changed.
And it's an outlier so let me tell you: the effect seen here is NOT (directly) related to a very slow change that started in 1850. It is something else.
This also means, don't use statistics to analyse it. It won't work.
Also, you can certainly study outliers with statistics. In fact, most scientific discoveries (these days) all start out as outliers or statistical anomalies. The Higgs boson, Top quark, J/Psi,... every single particle described by the Standard Model (the most accurate scientific theory by far) all started out as anomalies and outliers. The fact that this years sea ice measurements are outliers doesn't mean they are necessarily a large statistical fluctuation, it could be that there is a new climate phenomenon taking place (look up the runaway greenhouse effect). I'm not saying that is what is happening here but we have no idea a priori which is the truth at the moment: it's a statistical fluctuation or it's a sign of a new climate phenomenon. Clearly all we can do is continue to study it using, among other tools, statistics.
[0] - http://earthobservatory.nasa.gov/Features/CarbonCycle/page2....
But pedantically the point stands. The planck satellite studies a phenomenon that exists independently of the satellite or the humans studying it. This satellite was launched to study climate change by people who claim they're causing it. There's a clear difference.
> And exaclty why the results are disatrously?
Because using data about something you're doing yourself to study that phenomenon invalidates the basic axioms of statistics. If you do that, all your conclusions about the data become wrong. Why ? Because you yourself are now a factor in the phenomenon under study. At that point, and I've seen this many times, there is no longer any real hope of being correct, as there either a big gaping hole in the model, or circular reasoning. Both obviously invalidate the model.