Help with plot returning strange values for trendline

[TroyBarnes17]

Hi folks,

I am trying to do a calibration curve and have strange results from the trendline. It appears to be giving me a trend that beggars belief!

Just looking at the X and Y axes, it seems apparent that the relationship given is not likely at all.

Can you advise on where I'm going wrong please?

Thanks in advance

 

[DRSteele]

Welcome to the forum.

You have only three data points. The linear trendline is doing the best it can.

You can reconstruct the line using functions INTERECEPT and SLOPE, and the goodness of fit with function RSQ.

 

[TroyBarnes17]

Welcome to the forum.

You have only three data points. The linear trendline is doing the best it can.

You can reconstruct the line using functions INTERECEPT and SLOPE, and the goodness of fit with function RSQ.

Thank you!

I'm not sure this is a problem related to the number of data points. My first X point is close to 5 million, my first Y point is just over 1. So the trendline is predicting a value of 10million or so for the first Y value...so it's a 6 powers of 10 off!

 

[TroyBarnes17]

Thank you!

I'm not sure this is a problem related to the number of data points. My first X point is close to 5 million, my first Y point is just over 1. So the trendline is predicting a value of 10million or so for the first Y value...so it's a 6 powers of 10 off!

I meant 7 powers of 10!

 

[DRSteele]

The Chart Trendline is not predicting anything here. It is simply a line of best fit. If you use INTERCEPT and SLOPE on that data you will see the coefficients are identical to those of the linear Chart Trendline.

 

[TroyBarnes17]

The Chart Trendline is not predicting anything here. It is simply a line of best fit. If you use INTERCEPT and SLOPE on that data you will see the coefficients are identical to those of the linear Chart Trendline.

I have used those functions and their outputs are different than those of the trendline, by orders of magnitude. To address your first sentence, I know the trendline isn't forcasting anything but it's clearly saying that the boundary values it seems to have are vastly different from those in the data.

Like I said, if you plug the first X value on that graph into the trendline equation, you get a result that is nowhere near the first Y value of either the graph or the trendline. I was hoping someone could point me in the direction of why that was

 

[DRSteele]

What are your three Xs and Ys?

 

[TroyBarnes17]

What are your three Xs and Ys?

Hi, they're as follows. I get the intercept 15.15 and the slope = 2.7E-06 for them, which is different to the standard linear trendline where I get y = 3.135x - 2.2967

1.24​	4920178​
3.17​	4709716​
7.51​	2807479​