Abstract There is big concern for estimating the lethality and the extent of undetected infections associated with the novel coronavirus SARS-CoV2 outbreak. While detailed epidemiological models are certainly needed, I suggest here an orthogonal approach based on a minimum number of parameters robustly fitted from the cumulative data easily accessible for all countries at the John Hopkins University database that became the worldwide reference for the pandemics. I show that, after few days from the beginning of the outbreak, the apparent death rate can be extrapolated to infinite time through regularized regression such as rescaled ridge regression. The variation from country to country of these extrapolated death rates appears to depend almost only ( r 2 = 0.91) on the ratio between performed tests and detected cases even when the apparent instantaneous lethality rates are as different as 9% in Italy and 0.4% in Germany. Extrapolating to the limit of infinite number of tests, I obtain a death rate of 0.012 ± 0.012, in agreement with other estimates. The inverse relationship between the extrapolated death rate and the intensity tests allows estimating that more than 50% of cases were undetected in most countries, with more than 90% undetected cases in countries severely hit by the epidemics such as Italy. Finally, I propose to adopt the ratio between the cumulative number of recovered and deceased persons as an indicator that can anticipate the halting of the epidemics.