Eric Coffin nails it.... Here's the conclusion:
It could drop the average grade by several grams/ton, or more, for the higher grade areas. Not good at all. And, to be fair, cutting core is a shitty job and they are not hiring PhDs to do it. Entirely possible the sampler was doing what they thought was right and there was inadequate supervision perhaps. Not an insurmountable problem, but it will take a bunch of work to get an accurate grade estimate and it is likely to be lower. Sucks.
Then Eric's whole post, from CEO.CA:
@HRA-Coffin Assay discrepancy issues are always a (pain) as there is always a big "he said, she said" statistical fight involved and it is hard to get to the bottom of a lot of the time. That said, the more I think about it, the more I think the problem with $NFG may rest with sample selection in the core shack. You log core, then split it and send half to the lab while retaining the other half. If the samplers were, consciously or unconsciously usually tossing the half with "more shiny stuff" in the sample bag you introduce a huge amount of bias when dealing with a nuggety project with lots of VG. That is not the way you are "supposed" to do it. You are supposed to always put the same side (left or right side of the diamond saw) of the cut core in the sample bag, regardless of appearance. Better yet (and this is what $GBR does) is to always drill oriented core which is marked by the core barrel so you always know where the "top" is and it is consistent (core can get jerked around during the drilling process). You then use the orientation line as the diamond saw cutting line, always, and always choose the same side to go to the lab. I know they do it this way because I had an extensive conversation with Bob Singh about this one time - he is a stickler for this stuff. The fact that the high grade $NFG samples seemed to almost always be significantly lower on the umpire samples (i.e., the half of the core the original sampler chose NOT to put in the sample bag) certainly implies a major potential issue on QA/QC side, at the sampling, not the assaying level. If that turns out to be the case its a big problem short-term at least, though taking the samples from both sides of the core sample and averaging them will hopefully give you a fairly accurate picture. It could drop the average grade by several grams/ton, or more, for the higher grade areas. Not good at all. And, to be fair, cutting core is a shitty job and they are not hiring PhDs to do it. Entirely possible the sampler was doing what they thought was right and there was inadequate supervision perhaps. Not an insurmountable problem, but it will take a bunch of work to get an accurate grade estimate and it is likely to be lower. Sucks. $NFG