o.k., now I understand.
That is the same thing as to say, that from 60 throws of a dice we expect that 10 of them will be a “6″. Are you sure? It might also be 7 or 15 “6″. Or even zero.
From 100 statements, it might be 3,11,20 or whatever (each with a certain probability) that will turn out to be wrong. That’s the theory of probabilities. We just don’t know. And we have no idea, which of these statements it might be.

So, what has this to do with the discussion about correction of the global temperature record? The IPCC says, that with 90% confidence (or probability) the trend over the last 50 years (i.e. 1956-2005) is in the range of 0.10 – 0.16K per decade (numbers in sqare brackets represent the 90% confidence interval in the SPM). Do you really expect that after the correction the 1956-2005 trend will be out of that range? I am pretty sure that not.

I simply mean that if the IPCC provides (for example) 100 different assertions in its report, each one expressed at 90% confidence, then we should expect that 10 of those statements will ultimately be in error …

If the IPCC has accurately calibrated its uncertainties.

And if the statements are independent.

Similarly, half of all statements expressed at 50% confidence will be wrong.

you misunderstood.

The IPCC says with 90% confindence:
“Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.”

You wrote:
“When the IPCC says something with 90% confidence, it would not be a surprise to later learn that 10% of those statements were in error, would it?”

So can you please explain what it means that the above statement is “10% in error”? How can the statement, that GHGs are the main reason, be “10% in error”? How do you calculate that?

“Even if the trend would be reduced by 50% (which will not be the case, as you could see in your calculation, it’s likely less than 20%), this assertion is still true,…”

No, it’s not “still true,” EXCEPT if it is made clear that the TREND has been lowered by 50% (assuming 50% error…which I agree is unlikely).

If the error is 50% or more, the original statement (which was made about the ORIGINAL, pre-correction trend) is false. If the error is 50% or more, then the majority of the observed trend was measurement error.

Another strawman I see. Where did I ever say that “the anthropogenic part is at least 90%”? (Hint: I did not.)

I remain puzzled by your repeated mischaracterizations of my statements, followed by a failure to correct them when pointed out.

We’ll have to see if increasing the GHG influence to 70% (or more) of the mid-century trend will not cause some questions for existing attribution studies. Maybe, maybe not, but it is a bit concerning when a scientist tell me the answer before the work is done.

The 90% confidence relates to the assertion that “most of the trend is related to GHG’s”, not to the amount of the trend, and it does not mean that the anthropogenic part is at least 90%. O.k.?

Even if the trend would be reduced by 50% (which will not be the case, as you could see in your calculation, it’s likely less than 20%), this assertion is still true, because the influence of other known factors on the trend is very low (volcanoes negative, sun likely less than 15%). The GHG influence would still be more than 70%, which corresponds to “most”.

The first graph I posted (from Steve McItyre) was also a “first guess” and so everything that you say about the Independent “first guess” applies there as well.

I am unclear why you wish to present McIntyre’s first guess as being “mine” and The Independent’s as being yours and Gavin’s. They are both just guesses. I don’t know which one will look more like whatever future adjustments are made. That is why I discussed them both. In either case they are significant (as in large), with likely implications for a number of studies (e.g., for instance huricanes vs. ssts).

The relevant IPCC statement from the SPM:

“Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.”

I discussed why a reduction of 20/30/50% might be problematic for this statement in my first post, so no need to discuss here. When the IPCC says something with 90% confidence, it would not be a surprise to later learn that 10% of those statements were in error, would it?

Where did I blame the IPCC for anything? Here you raise another strawman.

I never said that the newspaper graph had a better adjustment, I only said that your interpretation was questionnable (suggesting about a factor of two), and your reanalysis has confirmed this.

I agree that the notion of the 5y average in the newspaper is wrong. And everyone knows that there will be some difference before 1945 in a smoothed graph. The graph is not perfect, and I don’t defend the newspaper, because it is irrelevant and we are not talking about that. I, and Gavin too, were talking of a “first guess” and didn’t say anything else than that a correction might look similar to what was presented in the graph, irrespective of a mistake in the caption and mathematical inaccuracies. A first guess is not intended to represent all mathematical details, especially not in a newspaper graph. If you are happy to prove that there has been a mistake in a newspaper graph, feel free. I am not interested in that.

Why do you think that the trend since 1950 is a “central conclusion” of the IPCC? Where exactly does the IPCC say something about a 90% probability of a trend since 1950, and of what trend exactly?
The IPCC SPM says: “The linear warming trend over the last 50 years (0.13°C [0.10°C to 0.16°C] per decade) is almost twice that for the last 100 years.” They speak about the trend 1956-2006. Do you really think that after the correction this trend will be outside the given uncertainty range?

That’s what I meant, BTW: after a correction at one particular point of one particular data set, you are very quick to suppose that “one of the central conclusions of the IPCC might be incorrect”. There is a long way from a correction of the 50y trend, which is most probably way in the given uncertainty range, to an incorrectness of central IPCC conclusions, which are based on much more than a simple trend.
I never said, nor did RC, that there would be no implications for the trends, and that the last word has been spoken on the global temperaure record. There will be a lot more corrections and improvements in the future. But before questionning “central IPCC conclusions” the relevance of the corrections for these conclusions has to be taken into account.

You blame e.g. the IPCC for not having done these corrections for a long time. That’s what I meant with “that not all the work has been done yet“.

Out of curiosity, I clicked the link to the IPCC discussion of the complications involved, we find what we have always known: The complications are discussed, but the text in no way suggests the series of adjustments might be off by as much as 0.3C during any period of time.

So what we find in the supplemental material for the IPCC AR4 is a document that discusses the many issues involved in determining (or estimating ) the SST, but overall conveys the impression the current estimate in the AR4 are more or less good. Or more specifically, there is nothing to suggest the SST data could be off by as much as 0.3C. (There is, in fact, no quantitative statement made about possible errors during WWII.)

If errors as large as 0.3C in the SST record were suspected, they were not disclosed in the AR4. The alternative is they were not suspected to exist — or at least errors of this sort were not considered likely by the authors of the AR4.