This is a ridiculous critique. We have thermometer data for the last 100 years. We don't need to look at rocks for that part. Chopping off the hockey stick would be disingenuous, ignoring accurate data.
This correlation of data acquisition appears to have been done until the conversion of recording stations to MMTS stations. The correlation stop, at that point, and we are now entirely dependent on thermometer data. Coincidentally this is is also the 'hockeystick blade' time frame.