There were 11,208 firearm homicides and 21,175 firearm suicides in the US in 2013 (http://www.cdc.gov/nchs/fastats/homicide.htm and http://www.cdc.gov/nchs/fastats/suicide.htm).
About 16% of firearm homicides occur during the course of a felony of any kind. The largest percentage of murders, more than 40%, occurs during arguments (Bogus, p. 7 http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1140442). Franklin Zimring and Gordon Hawkins have written "Most of the circumstances that generate homicide are not property crimes involving strangers, but arguments among acquaintances that nobody would regard as distinctively criminal until the attack began." If people substituted other weapons for guns, that would reduce overall homicides as other weapons are less deadly. There was a "Tale of Two Cities" study conducted by a team of epidemiologists led by John Henry Sloan and published in the New England Journal of Medicine which compared crime rates over a 7 year period in Seattle, WA and Vancouver British Columbia. Quoting the Bogus article:
Suicide rates with firearms have a fatality rate of 85% (https://www.hsph.harvard.edu/means-matter/means-matter/case-fatality/), far in excess of fatality rates by other means. The total suicide fatality rate of all methods is 9%.
Let's say that in total, banning firearm would conservatively save 33% of the firearm deaths in the US, so 10,000 people each year. For comparison purposes, the number of Americans that died in the Vietnam war was 58,000. One would have to assign a huge likelihood of mass murder by the US government in order to come close to offsetting the lives saved. For example, it would take a 5% chance each year that 200,000 are killed in some statist dystopia to get an expected lives saved that would offset the deaths prevented by eliminating firearms.