This is the wonky details of my post found here for those who like to quibble.
Note that throughout this exercise, I have focused on "externalized" fatalities: deaths to parties other than the drinker, the driver, and the gun owner. Matters such as suicide, single-car auto accidents, etc are excluded or explicitly accounted for. While obviously these are still a matter of public policy, they are on a different level than harm inflicted upon innocent third parties.
For automobiles, there are approximately 233 billion trips made per year in the US, and bit over 32000 traffic fatalities. However, roughly a third are self-inflicted and only involve the driver. This can be inferred from the facts that about half of all fatalities occur to a driver and that about a third of fatal accidents are single car, thus implying that roughly two thirds of the half of fatalities that occurs among drivers happens to the at-fault driver, while the other third of a half occurs to an innocent driver in a two car accident. Thus, there are approximately 21500 externalited traffic fatalities each year. This leads to the conclusion that with respect to driving there are approximately 10.7 million trips per externalized fatality.
For drinking, the average American drinks 9.4 liters of alcohol per year. If you assume that the average drinking session involves three ounces of alcohol (about three drinks), that implies the average American drinks 88 times per year, or 28 billion drinking events per year among all Americans. There were 10228 drunk-driving related fatalities in 2010, but approximately 45% of them are likely self-inflicted, resulting in 5625 externalized drunk driving fatalities. Additionally, approximately 28% of violent crime is alcohol related. There were 12664 homocides in 2011, and assigning 50% of the responsibility of "alcohol related" homocides to the alcohol results in 1773 externalized deaths. Combined with the traffic fatalities, this gives alcohol a total of 7398 externalized fatalities, or approximately one externalized fatality per 3.7 million drinking events.
The data for guns is a little more sparse and I have to make some assumptions. One is to look at ammunition use. The domestic ammunition market is $1.9 billion dollars per year according to a recent industry report. At $0.20 per bullet wholesale, that is 9.5 billion bullets. Of those, the military uses 1.8 billion, and other federal agencies 400 million. I could not find data for state and local police use, but considering there are more than five times as many state and local law enforcement officers as federal ones, my assumption that their bullet consumption is the same at 400 million is probably conservative. Thus, there are approximately 6.9 billion bullets available for civilian use annually. For a rough estimate, the assumption that people shoot six bullets on average when using a gun would lead to 1.15 billion uses per year. Of course, "using" a gun could range from no shots fired (an unsuccessful hunt, brandishing it in order to scare off an intruder) to a few boxes full at the gun range. The figure of six per use is literally made up but probably reasonable. Note that since there are about 64 million gun owners in the US, this would imply the average gun owner uses a gun 18 times per year and expends and annual total of 110 bullets. That seems very reasonable to me based on my experience.
We can attempt to look at this from the bottom up as well. There are 13.7 million active hunters in the US. Assuming 30 hunts per year, that is approximately 400 million hunts. As for shooting ranges, their annual industry revenue is about $560 million, so at $30 per trip, there are about 19 million visits to gun ranges. However, these would disproportionally add up to a lot of bullets fired, as most people shoot a fair number each time they are there. Even the wildest estimate of defensive gun uses is 2.5 million per year. Assuming the average gun owner uses their guns about ten times a year in addition to the aforementioned uses puts us right back at about 1.1 billion uses. To be generous, I will overweight the (likely exaggerated) defensive use estimate 100-fold, just for good measure, resulting in 1.35 billion gun uses per year.
Hand guns make up a bit under one third of all guns in the US. I will assume they make up no portion of the hunting uses, 80% of the 100-fold overweighted defensive uses, and a proporitionate 1/3 of the remaining miscellaneous uses, resulting in 510 million uses per year, with the remaining 840 million uses being long guns (rifles, shotguns, etc). According to the FBI, there were 8583 firearm deaths in 2011, with 6220 from handguns, 776 from long guns, and 1587 from "unstated". Assigning the "unstated" deaths to handguns and long guns proportionately, this results in 7631 hand-gun deaths and and 952 long-gun deaths in 2011. Thus, 1.15 million uses per externalized fatality for long guns, and 67200 hand gun uses for externalized fatality.