Next try:
I used a different laptop than the computer I used the web app on again only as a power source (because the laptop is too old for the web app). The computer I use the web app on needs a power supply, but my wall plug destroys the signal.
My room was pretty dark, because of my blackout curtains.
I did 4 tries and always reset the data in between.
They were all the same basically:
ratio: was very stable from 0,59 - 0,62. Sometimes it was the same for about 10 seconds or so. From what Joshua wrote so far, it should be pretty stable and from 0.5-2.
score: around 0.10
score line: was very flat
I think these values don't necessarily mean I've done something wrong, but: The circle only changed a very very small bit, if at all. Almost not visible. I'm pretty sure that I wasn't able to influence it. The very small movements seemed random.
The band was pretty tight/tense. I think the LEDs were pressed into my forehead too much. See the picture. 😂 I also felt my pulse, so I think the band was too narrow. Can this be a problem?
I wore it like this to prevent light going from the LEDs to the sensor directly.
But for the last attempt I made the headband about 5cm wider and still got the same results.
During one attempt I pressed the LEDs more into my head and the ratio and the graph went down. But during a later attempt nothing happened at first, but after it the ratio remained a bit higher.

Or could it be possible that there's nothing wrong and that it just takes a few sessions to be able to change the circle?
But unfortunately I don't that's the case, because people can use e.g. the mendi device successfully from the first session on.
Sounds like it's working. Low fluctuations are a good sign. It doesn't need to be so tight it 's squeezing your head, just tight enough not to come off your forehead and screw up the signal. Expect slow changes, it takes a few seconds just for your body to process a single breath, but I have noticed that some people start out less responsive to the breathing exercises.