I'm using a "custom" distribution by pasting in my own data from a spreadsheet. In one case I sorted the values (smallest to largest) before pasting in to Guesstimate; in the second case I randomized the order of the values. Note the differences in the sensitivity analysis graph. It seems like differences like this should not result from randomizing the order of the custom data, or else the documentation should warn users that the order of their data matters a lot.
In other cases I get differences in the reported r2 of the sensitivity analysis, and in the reported median of the dataset, just by sorting the values. I think what's happening in those other cases is that I pasted in more than 1000 values, and Guesstimate truncates to just the first 1000 before some analyses.
I'm using a "custom" distribution by pasting in my own data from a spreadsheet. In one case I sorted the values (smallest to largest) before pasting in to Guesstimate; in the second case I randomized the order of the values. Note the differences in the sensitivity analysis graph. It seems like differences like this should not result from randomizing the order of the custom data, or else the documentation should warn users that the order of their data matters a lot.
In other cases I get differences in the reported r2 of the sensitivity analysis, and in the reported median of the dataset, just by sorting the values. I think what's happening in those other cases is that I pasted in more than 1000 values, and Guesstimate truncates to just the first 1000 before some analyses.