Not a counts table, a cell with the minimum number of records before you use the average so you don't get a small number of results messing up the data.
i agree, without enforcing a minimum cell count, you will end up with erroneous spikes in your data from cells with only one or two records that averaging hasn't been able to smooth out.

you should also provide a percentage against target AFR data, if any is available, people find that really handy.

im not entirely sure how to do this stuff with a spreadsheet. this is the kind of stuff that always drove me towards programmed tools for analysis.

really wish tunerpro had filtering in its history tables.....