Mike's Nature Trick

In response to comments on this article at the guardian:

Mike's Nature Trick, was the statistical blending of tree ring proxy data with the modern surface temperature record in a graphic for the WMO in 1999, this was done because the proxies differed from that record (the divergence problem) in the modern era.
This practice was unorthodox because:
* The blended lines were contiguous with multiple proxy records (i.e. they gave the impression the proxies matched modern temperatures when they actually diverged).
* The blended lines were not labelled.
* The mechanism for the divergence is not known (and still isn't - though there is some generic speculation). Therefore the divergence is extremely noteworthy and should be highlighted not hidden.
It amazes me that people continue to defend the indefensible on this. The WMO graphic invites you to draw conclusions (that proxy records show the modern warming is unprecedented), that are not supported by the proxy data they are displaying.
Also, just because a phenomenum is named (the divergence problem) doesn't "magically" get rid of the issues it raises. The divergence problem undermines the whole practice of using tree rings as proxies, and the resulting paleo reconstructions are a lot less on message as a result (see Moberg as an example).

I think the point here, is that it was a post-hoc redaction of data, and blending with the instrumental record, for the purpose of hiding the decline on the proxy records, and to strengthen the argument that modern era warming is unprecedented.