For some background, it is essential to know that Ebola is actually a group (genus) of ebolaviruses, each with different fatality rates. There are five known species of Ebola, and four are known to cause disease humans (WHO: Ebola virus disease; wikipedia).
The known species of Ebola includes:
- Zaire ebolavirus (or just ebolavirus)
- Sudan ebolavirus
- Bundibugyo ebolavirus
- Tai ebolavirus (only one known human case)
- Reston ebolavirus (can infect but does not seem to cause disease in humans)
Zaire ebolavirus is the deadliest strain (with Sudan second), and also the one that is causing the current outbreak. A table of all known ebola outbreaks can be found in the WHO report (WHO: Ebola virus disease), which shows that the historical fatality of the Zaire strain has been between 44% and above 90%.
From my limited understanding, when dealing with ebola it is essential for infected individuals to be identified early so that they can get intensive care (most importantly intravenous fluids because of dehydration), but also to prevent further spread. This is probably what has failed during this outbreak, but another problem is also that the size of the outbreak has put tremendous strain on an already weak healthcare system. These outbreaks have also occurred in well populated areas in countries with very weak governments, that also lacked experience in dealing with ebola (both in the administration and among the population).
In many former outbreaks, the initial death rate is very high since early victims are found at late stages of the disease, but when the outbreak is more controlled the death rate drops due to supporting care of infected individuals. I don't know to what extent this has been the case during the current west african outbreak. However, the overall fatality rate of the current outbreak is 52% (tabulated Aug. 28 2014), with 44% in Sierra Leone and 66% in Guinnea (WHO: Ebola virus disease update - west Africa), and this lies well within the normal fatality rate of the Zaire ebolavirus.