Calculating cumulative periastron time shift from relativistic orbital decay

[Meerio]

I'm confused about how the predictions of this graph were formed.
I have this formula:

But the change in frequency per second is about : 2x 10^-12 which has a problem because when you multiply this by 10 years you don't even get a change of frequency of 1/1000 of a second and in the graph it says it's around 5 seconds.
Anyone know what I'm doing wrong?

Paper: https://arxiv.org/pdf/1011.0718v1.pdf

 

[mfb]

This is the cumulative change in time, not in frequency. You need the integral or sum.

The first measured orbit was about 7.75 hours, the next was 2.4 * 10^-12 * 7.75h = 67ns shorter (integrated shift: 67 ns), the following one was 2*67 ns shorter than the first one (cumulative shift: 67+2*67 = 201 ns) and so on.

In 10 years, there were 11300 orbits, for a cumulative shift of 1/2 * 67 ns * 11300² = 4.3 seconds, in agreement with the parabola shown there.

 

[Meerio]

Thx so much this was the answer I was looking for !