How to evaluate if a power measurement is good

[edwardtfn]

So, I'm now measuring a couple of Yeelight YLDP02YL lights. I'm not using any dummy load and the results of my measurements are the following (also CSV file attached):

color_temp.csv

Some additional info:
LIGHT_MODEL_ID=YLDP02YL
MODEL_NAME=Yeelight Color Bulb LED YLDP02YL
DUMMY_LOAD=false
MEASURE_DEVICE=Shelly Plus 2PM
MULTIPLE_LIGHTS=true
NUM_LIGHTS=2
COLOR_MODE=color_temp
GENERATE_MODEL_JSON=true
GZIP=true
RESUME=true

Does that looks good or am I doing something wrong?
I'm now measuring with COLOR_MODE=hs and then will do rgb, so I will create a PR with all those files as soon as I finish this.

[bramstroker]

Looking at the graphs measurements look pretty good imo. Slowly increasing watt and higher in the mid mired ranges.

[bramstroker]

Only thing I notice is that there are a lot of similar measurements (for example 10 times the same, and then another value).
With the Shelly Plug S I get a new different measurement each time.
So maybe it would be wise to try increasing the sleep time a bit.

[edwardtfn]

I will double check, but I believe this is more related to the resolution of my power meter. I'm using a Shelly Plus 2 PM, which resolution is 0.1W (and will be calculated to 0.05W as I'm measuring 2 bulbs together).
I will investigate this deeper.