Equation error for linear trendline with dates

[davemcwish]

I have the following data that I've plotted and trying to forecast when 'Value' reaches 8


[TABLE="width: 193"]
<colgroup><col><col></colgroup><tbody>[TR]
[TD][TABLE="width: 500"]
<tbody>[TR]
[TD]Date[/TD]
[TD]Value[/TD]
[/TR]
[TR]
[TD]01 Feb 2015[/TD]
[TD]6.9330[/TD]
[/TR]
[TR]
[TD]04 Mar 2015[/TD]
[TD]6.9588[/TD]
[/TR]
[TR]
[TD]04 April 2015[/TD]
[TD]6.9847[/TD]
[/TR]
[TR]
[TD]04 May 2015[/TD]
[TD]7.0697[/TD]
[/TR]
[TR]
[TD]04 June 2015[/TD]
[TD]7.1576[/TD]
[/TR]
</tbody>[/TABLE]

[/TD]
[TD][/TD]
[/TR]
</tbody>[/TABLE]
Using a linear forecast trendline, I get the equation

y=0.0019x-71.545

which gives a value of x as 41865, for y=8

This doesn't seem right as the first value I have is for 01 Feb 2015 (42036)


Interestingly, if I switch to a Logarithmic trendline I get

y=78.558ln(x)-829.44

which gives a reasonable value of 42622 (9 Sep 2015)


Anyone got an answer to why the linear trendline equation gives a seemingly incorrect value ?

 

[pgc01]

Hi

The way you display the coefficients, it seems out of proportion. You have 2 significant digits in the x coefficient but 5 significant digits in the x values and in the other coefficient.

Try for ex. 12 decimals. (in the chart, right-click on the equation label, choose Format Trendline Label->Number->Category: Number->Decimal Places: 12)