## The lottery of drunk driving fatality

Several years ago, I read of a news account of a mother driving with her infant son in a minivan. A drunk driver was driving on the same freeway but in the wrong direction and collided with the minivan. The mother died upon impact. Amazingly the infant survived. News accounts like this one are always heart wrenching.

I always think of dying from a crash involving an alcohol-impaired driver is a lottery. It is a negative lottery for sure since no one would want to win it. Los Angeles Angels pitcher Nick Adenhart and two of his friends were the unfortunate “winners” of this negative lottery in 2009.

What are the odds of winning this negative lottery? How often is this negative lottery “won”? Let’s compare the number of drunk driving deaths with the number of winning Lotto tickets that pay out $1 million or more. The comparison is done on an annual basis and as a rate per certain number of minutes. It turns out that winning this negative lottery is far more frequent than winning the Lotto. Driving under the influence of alcohol is a serious problem. The numerical data presented here bear that out. Drivers are considered to be alcohol-impaired when their blood alcohol concentration (BAC) is 0.08 grams per deciliter (g/dL) or higher. Thus any fatality occurring in a crash involving a driver with a BAC of 0.08 or higher is considered to be an alcohol-impaired driving fatality. According to a report from the Center for Disease Control and Prevention (CDC), one alcohol-impaired driving fatality occurs every 45 minutes (see note 1 at the end of the post). This rate of occurrence is derived from the fact that in 2008, 11,773 people were killed in alcohol-impaired driving crashes. What is the total number of Lotto winners in a year? Here, we only focus on the winning prizes of$1 million or more. Would the number of Lotto winners approaches 11,773 in a year? I do not think so. I hunted for data from the website of the California Lotto, I found that as of today’s date, there are only 247 winning tickets that pay out $1 million or more since the inception of the California state lotto in 1985. So in 25 years, there are only about 250 winners in the state of California. Thus on average there are only about 10 “million dollar plus” Lotto winners a year in California. Multiplying across the 50 states in the United States, the total number of “million dollar plus” Lotto winners should be no more than 500 in a year. This should be a pretty conservative estimate since California is the largest state in the country (with the largest population). Not all 50 states in the country have Lotto. Some states do not have their own Lotto and are part of a multi-state Lotto. So 500 is a good upper bound on the total number of “million dollar plus” Lotto winners in a year. With 500 winners a year, there is one “million dollar plus” Lotto winner every 1050 minutes (see note 2). Or one winner in every 17.5 hours (see note 2). Thus alcohol-impaired driving deaths occur at least 23 times more frequently (one death per 45 minutes vs. one per at least 1050 minutes). Of the 11,773 alcohol-impaired driving deaths in 2008, how many of them were the drunk drivers? According to a report from National Highway Traffic Safety Administration (NHTSA), an agency within the Department of Transportation, we have the following breakdown of the drunk driving fatality in 2008. $\displaystyle \begin{pmatrix} \text{Role}&\text{Number}&\text{Percent of Total} \\{\text{ }}&\text{ }&\text{ } \\\text{Driver With BAC=0.08+}&8,027&\text{68 percent} \\\text{Passenger Riding w/Driver With BAC=0.08+}&1,875&\text{16 percent} \\\text{Occupants of Other vehicles}&1,179&\text{10 percent} \\\text{Nonoccupants}&692&\text{6 percent} \\{\text{ }}&\text{ }&\text{ } \\{\text{Total Fatalities}}&11,773&\text{100 percent}\end{pmatrix}$ Most of the deaths were self-inflicted (68%). According to the same report, 216 of the 1,875 deaths in the above table were children. The group of 2,087 (=216+1,179+692) were truly victims and their deaths were senseless. Angels pictcher Nick Adenhart and the mother mentioned at the beginning belong to this group. They did not do anything wrong and were just in the wrong place at the wrong time. If we just focus on this group of children riding with the drunk drivers and the occupants of other vehicles and nonoccupants, the rate of occurence is still pretty high (one fatality every 252 minutes; see note 3). That comes up to be about one fatality every 4.2 hours (see note 3). Focusing on this group alone, the rate of drunk driving fatalities is over 4 times higher than winning the Lotto (one death per 252 minutes vs. one winner per at least 1050 minutes). Eliminating the category of fatality represented by these 2,087 senseless deaths would drive down the rate of fatality and would go a long way to address the problem of drunk driving. Though this would be only modest improvement, it would be a step in the right direction. ************* Note 1 To derive the rate of one death per 45 minutes, we need to calculate the total number of minutes in a year. There are 365 x 24 x 60 = 525,600 minutes in a year. Then divide 525,600 by 11,773 to obtain 44.64 minutes = 45 minuties. We can get a perspective of this calculation by looking at an example of taking an exam. For example, if you have two hours (120 minutes) to take an exam and the exam has 10 problems, then on average you have 12 minutes to work one problem. Thus if you can work one problem per 12 minutes, you can expect to finish the exam in the allotted time. Back to the alcohol impaired driving situation, there are 525,600 minutes in a year and there are 11,773 events. Thus on average there are 45 minutes allotted for each event. $\displaystyle \frac{365 \times 24 \times 60}{11,773}=44.64 = 45 \text{ minutes}$ Note 2 $\displaystyle \frac{365 \times 24 \times 60}{500}=1051.2 = 1050 \text{ minutes}$ $\displaystyle \frac{1050}{60}= 17.5 \text{ hours}$ Note 3 $\displaystyle \frac{365 \times 24 \times 60}{2087}= 252 \text{ minutes}$ $\displaystyle \frac{252}{60}= 4.2 \text{ hours}$ ## An American in Paris We in the United States still use the English system of measurements while most of the rest of the world uses the metric system. For example, we use pounds and ounces in food packaging. Our motoring experience is measured in miles per gallon (MPG) and miles per hour (MPH). The distances shown in road signs are in miles. Suppose you are an American driving on a French highway toward Paris and you see a road indicating a distance of 270 kilometers to Paris. How many miles are in 270 kilometers? From my road race running experience in the past, I learned that a 5K road race is 3.1 miles long. This information provides a quick and dirty way of making the conversion from kilometers to miles and vice versa. To convert from kilometers to miles, the key is to determine the number of multiples of five in the distance with kilometer as unit. For example, 10K would be 3.1 x 2 = 6.2 miles, 15K would be 3.1 x 3 = 9.3 miles and so on. In other words, multiple of 5K would be the same multiple of 3.1 in miles. The distance of 270 kilometers is 54 times of 5K. Thus 270K = 54 x 3.1 = 167.4 miles. The above is a speed limit sign from Australia. Note that 80 = 16 x 5. Thus 80 K = 16 x 3.1 = 49.6 miles. So the speed limit of 80 K/h is about 50 MPH. To convert from miles to kilometers, we do the opposite, that is, determine the number of multiples of 3.1 (or 3) in the distance with mile as unit. For example, the driving distance from Los Angeles to New York City is about 2800 miles (903 multiples of 3.1). Thus 2800 miles is approximately 903 x 5 = 4515 kilometers. For an even quicker estimate, let’s say the distance from LA to New York is 3000 miles. Then the distance in kilometers would be roughly 5000 kilometers. The same conversion may be done more formally using the following conversion factors. 1 Km = 0.621371192 miles 1 mile = 1.609344 km I would like to point out that for everyday purpose, the quick and dirty way of 5 K = 3.1 miles would work just fine. From 5 K = 3.1 miles, 1 K = 3.1/5 = 0.62 miles. Likewise, 1 mile = 5/3.1 = 1.62 K. These are quite close to the formal conversion factors indicated above. Unless you are an engineer building a highway, the approximation of 5 K = 3.1 miles would work just fine. In the United States, sometimes miles and kilometers are shown together. In some cases, the side-by-side information can actually clue you in on the conversion factors. The following is a warning sign in Hawaii. The sign indicates that 0.5 miles = 0.8 kilometers. Thus 1 mile = 1.6 kilometers. The following sign indicate that 100 kilometers = 62 miles. Thus, 1 kilometer = 0.62 mile. On rare occasions, road signs in the United States are kilometer only. See the following sign from Vermont on I-89. ## How big is the median household income? The U.S. Census Bureau announced on September 16, 2010 that real median household income in the United States in 2009 was$49,777. How big is this amount relative to, say, the income from earning the minimum wage? The current Federal minimum hourly wage is $7.25. But states are allowed to set their own minimum wage rates. The hourly minimum wage among the 50 states currently ranges from$5.15 to $8.55. Obviously, the median household income is more than the annual income from full time employment paying the minimum hourly wage. By how much? This post presents a back of the envelop calculation that does not require a calculator or math formula. For the interested reader, the precise calculation is also presented for comparison. The median household income of$49,777 as reported above is essentially $50,000. Suppose someone is making$1,000 a week. This will add up to $52,000 a year (since there are 52 weeks in a year). Assuming working 5 days a week and 8 hours each day, this means that the person makes about$200 each day ($1,000 divided by 5) and about$25 per hour ($200 divided by 8). So in this back of the envelop calculation, the annual median household income is equivalent to an hourly rate of about$25 per hour. Thus the median household income is about 3 to 5 times the income from full time employment paying the minimum wage.

The annual income of $52,000 is more than the median household income by about$2,000. So the estimate of $25 per hour is an overestimate of the hourly rate equivalent to the annual income of$49,777. The precise hourly wage is:

$\displaystyle \frac{49777}{52 \times 5 \times 8}=23.93$

Thus the estimate of $25 overestimates by about$1 per hour. But it still gives the essential information that the median household income is 3 to 5 times more than the minimum wage income (depending on the state). For a rough comparison, this quick and dirty calculation works quite well.