First of all, I want to thank the developers for improving the charts by allowing the option to manually adjust the Y-axis scaling.
However, there is still a fundamental flaw in the design by creating separate scales on Combo Charts.
There is no reason to ever have different scales for the 2 Metrics on a Combo Chart.
The primary reason to use a Combo Chart is to visually compare 2 Metrics' data sets, which becomes meaningless if they are on different scales since it effectively creates ''an apples to oranges comparison''.
For example, when comparing income (bar) to budget (line), you can see visually if the income (bar) is below or above what was budgeted (line). If there are separate scales for income and budget, then it's possible that the income (bar) could be visually BELOW the budget (line), even though in reality the income number is HIGHER than the budget number. That would be very misleading and confusing to anyone viewing that chart.
There may be times when one of the income (bar) values are much higher than the corresponding budget (line) value. Currently, if that ratio is above a certain threshold, it will automatically create separate scales for the Bar and Line graphs given the priority to make the graph more visually appealing. I think this is a mistake and suggest that the priority should be the TRUTH of the 2 Metrics' relationship, otherwise we end up in a situation noted above with a bar looking like it is below a line, but with a larger number than the corresponding line.
If there does happen to be a big gap between the bar and line data sets, it doesn't "fix" that issue by creating two separate scales. It may "look better", but its misleading and effectively meaningless comparing ''apples to oranges''. In fact, the big gap may be the point of comparing the two data sets in the first place. For example, if we are comparing our monthly Spend Rate (bar) against our Cash on Hand (line), having a large gap would visually demonstrate healthy cash management. By making those two graphs closer together using separate scales, it minimizes and distorts that visual impact.
I appreciate the improvement that the new Y-scaling "Off "option now allows one to turn off the separate Bar/Line Metric scales in order to create only one scale. However, there is no value in having separate scales in the first place, given the visual confusion mentioned above. I recommend removing that function to avoid this confusion.
The related issue is that since creating 2 scales is currently still possible, choosing the "Auto" or "Manual" scaling option can create separate scales and making those options unhelpful. It is helpful that the "Auto" option will automatically create the bar max value slightly above the highest value in the displayed bar data set, but it can become very confusing if it also creates a separate scale for the line creating the same ''apples to oranges issue'', and should set the max value slightly above the highest data point in either the bar or line data set.
Similarly, when choosing the "Manual" Y-scaling option, it should only be for setting the max value for the ONE scale used for both, the bar and line graphs.
Please keep the Combo Chart with ONE scale, so we can have a truthful visual representation of the two data sets, regardless if it seems visually appealing or not.