Interesting. A battle between statisticians about how wrong the model of the other is. Happens only twice per day here.
From my basic knowledge of statistics, I fail to see the point of climateaudit.org. He does not present his own model. Statistics are only as correct as the model they are based on. And McIntire is not publishing his model, he later only claimed that the original data of Mann supported his claims.
From the debate, there is one thing known about McIntire's blackbox model: It uses a CE selection (don't ask me what CE is), while Mann used an older RE selection. CE assumes that the mean values of the data don't change over time (stay around the same average for large numbers of samples), while RE allows a change of the mean values over time.
So, the models of McIntyre and Mann differ maybe just in one basic assumption:
Mann: Average temperature can change over time and selection has to follow local averages.
McIntyre: Average temperature stays constant over large numbers of samples and selection has to stay within the variance of the global average.
Or even if I maybe push the interpretation further away:
Mann: Climate is dynamic and can change over shorter periods.
McIntyre: Climate is static and does not change over long periods.
So, the real question is now: which model is closer to the truth?
What do we do about it? Science works very simple. You not only publish as scientist your results, but also how you achieved your results.
McIntyre did once publish his results, but he never released the method he used to get to this data. He also confirmed in 2005, that the data Mann provided in 1998 had been mostly correct (after he had full data sets around), but he still did not tell his own method accuratly. As Mann already released all his method and data to the public for investigation, McIntyre should be the next to do so.
Important IMHO is the question: Is McIntyres dogmatic fighting against the RE selection process reasonable by scientific standards or based on an assumption that should be defended at all costs?
But when somebody mentions CimateAudit.org, let's also add the attacked blog to the list of links...
http://www.realclimate.org/