cpdn allocates credit based on the number of trickle_up files received.
These are multiplied by a number that's decided on during the testing phase.
The number is based on the amount of cpu work involved in creating a model.
It varies, as the length of time is dependent on the size of the area being modelled at increased resolution for the regional models, (which involves more calculations), and by the application type. e.g. the RAPID-RAPIT models use a different program to the regional models. And some of the models under test use a new "surface use" scheme, which increases the time needed for calculations. i.e. they're more detailed.
And then of course, you have the situation where a partly completed model will not have as much credit as a completed model.
This is because the trickle-to-credit programs only run once per day.
And they recalculate EVERY trickle_up that's been received, for all of the various totals.
If you want to know how much for each model, look at the page for one.
This is one of your EU models.
At the bottom is a list of the trickles, which you can count.
Just above the Perturbed Parameters list, is the amount of credit that you received for that model.
Divide amount by number, and you have the amount of credit for each trickle, for that model type.