The way you describe your problem sure makes it difficult to tackle ...
Here's what I think you should do:
Use an average on the Col2 of the dataset, instead of an average on the other value that is in your grid.
The average that you use, is pointing to the datapoint for the specified group, and therefor will evaluate only the Col2 value for one x-axis point, as opposed to the average of the whole dataset (as i think you want to have).
Peter Rijs
BI Consultant, The Netherlands