CV19 Infection Rate as a %

I don't understand the government figures on infection rates.

Is the 7-day rolling average of new infections calculated from the percentage of those tested, or purely just the number of new infections.

Surely, there is a relationship between the two.

If 100,000 are tested and 5,000 positive, that is clearly 5%.

But if 150,000 are tested and 6,000 positive, that is 4%.

Hence, the percentage rate of infections would be reducing. But the number of infections will have increased.

Hence, what is the 7-day rolling average calculated on?

If it is calculated on purely the number of new infections and not the % rate, then the more we test the more cases we will have and we will never get out of this until everyone in the country is tested!

Or, have I misunderstood something here?

BTW, there was a time in the UK when I would not have to ask this question because the government could be trusted to a/ understand things b/ not spin things.

Not in past 20 years, though.