If I were to calculate today's dead vs. total known cases, I would skew the mortality rate toward the low side. The infected cases today likely became infected between 4 and 21 days ago, and the outcome has yet to be determined.
Instead, as a rough guide, comparing the number of dead today vs the number of infected (total) 12 days ago would provide an estimate of those total infected versus the outcome for that group.
When you look at the steep growth curves for number of cases and deaths plotted over time, the distortion becomes evident.
In the current outbreak, there are reporting/recording problems as well, and some estimates put the number of cases/deaths as much as 4 times higher than the numbers which have been confirmed.
For past outbreaks, those which burned out and were considerably smaller, the mortality rate ranged from a low of roughly 40% to a high of 90%
That report addressed Ebola in Guinea, Liberia, and Sierra Leone. At the time of the report, the highest fatality rate was in Guinea, which experienced a 64% fatality rate. Outbreaks were still in process in all three countries.