Lets make the default the leftmost data point.

Expected behavior:

1. Default behavior will now correspond to "Begin trend at first data point" in UI.  Image.
2. Legacy behavior will now correspond to  "Begin trend at year zero" in UI.  Image.
3. There should be no visible change in the chart when switching between the two options.  Only the numerical values of slope and intercept (or other fit coefficients) should change.
4. Attached are workbooks for testing the various fits.  Since we also have rserve library we used when testing curves, we can compare those values to the new trend values.  For example, we can compare linFit results with rServe linFit results and the trend result. Image.
5. In general a good check is whether the new y-intercept makes sense, i.e. does the value correspond to the chart values.

Attached workbooks:

1. dateGroupings.vizx has tabs for primary continuous, secondary continuous etc cases.
2. newTrendTests.vizx has tabs for comparison of lin, log, exp, and polyFits
3. refFits.vizx, has log, exp, and polyFits with reference functions. I.e. calculated fields which are polynomials, logs, etc to verify that the fits are capturing them correctly.  These are very easy to check as the fit and raw data should exactly coincide.

Verified in ADS 18.0.14-3

a) New Time Series options are added in Edit TrendLine dialog.
- Begin trend at first data point
- Begin trend at year zero

b) By default "Begin trend at first data point" option is selected.

c) In case of continuous date dimensions,
- If "Begin trend at first data point" is selected
> Trend line starts from the leftmost point of the dimension axis.
> Trend calculates according to the first available position for a chart data point.

- If "Begin trend at year zero" is selected
> The legacy behaviour of trend line corresponds to this option.
> The trend is calculated for zero date (Jan 1, 1970)  rather than the leftmost date visible on the chart.

Please refer attached Verified_15258.png