Calculate possible dark figures

[arildwtv]

I have an idea of how we might attempt to dynamically predict the current dark figures. It's logically, and not scientifically, based, though. What do you guys think:

1. We use live official test data (positive test results / total number of tests) for each country to get a percentage of probable infection (let's call it Pp).

In Norway, Pp is currently 2.625 / 61.251 = 4.2 %.

2. We use data from our infected respondents to calculate the X most common symptoms among the infected (already implemented), or we could use a similar list from WHO (let's call the most common symptoms S).

In Norway, for example, S is { exhaustion, fever, dry_cough, headache }.

3. We calculate the percentage of respondents (excluding infected) with all S symptoms (let's call it Ps).

In Norway, Ps is 5.6 % (with all four most common symptoms), according to our respondents.

4. Now, what if we propose this formula: P = Ps * Pp, where P is the probabilty of infection among respondents with S. (I'm very willing to discuss this formula with you guys).

In Norway, P would be 4.2 % * 5.6 % = 0.235 %. With a population of 5.500.000, this would suggest a total of 12.936 cases. 10.311 of those cases would be unreported. That makes for an unreported case percentage rate of 10.311 / 12.936 = 79.7 %.

I would suggest another parameter in the formula, as well: Whether or not you know that you've been in close contact with an infected. This should increase the chances of being infected. Or conversely, it should decrease the chances if you don't know if you've been in close contact with an infected.

What do you guys think?