id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
cord-127109-jdizyzbl	Bertschinger, Nils	Visual explanation of country specific differences in Covid-19 dynamics	2020-04-15		.txt	text/plain	3246	214	59	Indeed, I show here that SIR type models -and others exhibiting similarly flexible growth dynamics -are non-identified with respect to the CFR and the fraction of observed infections. Figure 4 shows the country specific estimates of reporting delay, CFR and fraction of observed cases (assuming a true CFR of 1%) obtained in this fashion. In turn, Figure 5 shows the implied relative case counts when shifted by the estimated delays and scaled to reflect the unobserved fraction of cases for each country. A suitable reporting delay τ c can be estimated by visual inspection of the data, but again the fraction of observed cases α c and CFR cfr are not jointly identifiable if there exist sets of parameters such that a t−τ = αa t , as is the case for dynamic SIR type models. Assuming that only a fraction α of cases is observed, the model is estimated with the following Relative days since two death per mill.	./cache/cord-127109-jdizyzbl.txt	./txt/cord-127109-jdizyzbl.txt