I also ride my bike in Berkeley without a helmet.

Some other considerations which influence me:

  • helmet use seems to make crashes more likely (by making bikers and/or drivers less cautious), so it's misleading to use data about harm that's conditioned on there being a reported accident.

  • I'm fairly careful to avoid roads with heavy traffic, or with cars driving more than about 30 mph. I expect that fatality rates vary a lot by these road factors.

  • I use a cheap bike that doesn't go as fast as the average bike.

Alas, I don't have good evidence about how to quantify these considerations.

7ChristianKl1yIt seems like it would be most effective to wear something that doesn't look like a helmet but protect like [https://hovding.com/]one [https://www.kickstarter.com/projects/267784424/anti-ordinary-a1-a-beanie-thats-as-safe-as-a-helme] .

Interesting. The ability to fold it enough to fit in a backpack should reduce the hassle of storing it at my destination, which has been part of why I've been reluctant to use one.

How dangerous is it to ride a bicycle without a helmet?

by habryka 3 min read9th Mar 201930 comments


Epistemic Status: An hour of googling combined with some highly dubious statistics and very rough notes. Take with at least 500 grams of salt.

I don't like wearing bike helmets, and have a bunch of friends who were horrified at me not wearing helmets when I occasionally ride a bicycle. On their request, here is a rough cost-benefit analysis of riding a bike without a helmet to my office and back (a total of ~20 minutes):

I decided to break down the question into the following three subquestions:

  • How great are the benefits from exercise when riding a bike?
  • What is the baseline risk of riding a bicycle?
  • How elevated is the risk by not wearing a bicycle helmet?

How great are the benefits from exercise when riding a bike?


According to the first episode of More of Less (haven't gotten to the second one yet), the benefits of cycling outweigh the risks by about 5-10 times.


Micromorts are a unit of risk measuring a one-in-a-million probability of death. You can say that a mode of transport results in so many miles per micromort and the result is a number you can really use. For cycling, this is 10-20 miles, for driving it is about 230 miles.
Microlifes are a millionth of an (average remaining) lifespan = 30mins of life, so activities which extend your life or decrease it in a chronic rather than acute way can be measured.
In looking at the risk associated with cycling to work, people often quote the mort factor You also have to include the life factor. First 20mins of excercise = +2 microlifes . Sedentary behaviour = -1 microlife


First 20 minutes of moderate exercise: 2 microlifes
Subsequent 40 minutes of moderate exercise: 1 microlife

However, this is more complicated by the other costs of wearing a helmet, as well as a preference of mine to not die at a particularly young age. All-cause mortality is really low in my age-range, so bicycling might be a significantly higher proportional increase for that age range, and with shorter GCR timelines I might not care super much about my long-term health, which also reduces the value of exercise (though then there are also immediate cognitive benefits of exercise beyond all-cause mortality that start taking effect immediately, which seems maybe like a sufficient counterbalancing consideration).

Concretely, if I cycle for 20 minutes every day and that is my only exercise, I gain about 2 microlives. If I have some other source of exercise, I gain more around 0.5 microlives. On average I think it's more something around 1 microlive for the 20 minutes of cycling. So let's go with that number for now.

What is the baseline risk of riding a bicycle?

Wikipedia for one micromort: Travelling 10 miles (16 km) (or 20 miles (32 km)) by bicycle (accident)

Average bicycle speed (Wikipedia): 15.5 km/h

Micromorts per hour of cycling (baseline): ~1

That results in about a gain of a third of a micromort over 20 minutes of cycling.

How elevated is the risk by not wearing a bicycle helmet?


Results: Estimates of helmet effectiveness were similar from odds ratios (ORs) using hospital controls or from relative risks (RRs) using helmet use estimates (Seattle: OR = 0.339, RR = 0.444; Victoria: OR = 0.500, RR = 0.353). Additionally, the odds ratios using hospital controls were similar when controls were taken from a larger cohort for head injury of any severity (Seattle: OR = 0.250, alt OR = 0.257; NSW: OR = 0.446, alt OR = 0.411) and for serious head injury (Seattle: OR = 0.135, alt OR = 0.139; NSW: OR = 0.335, alt OR = 0.308). Although relevant exposure data were unavailable for The Netherlands, the odds ratio for helmet effectiveness of those using racing, mountain, or hybrid bikes was similar to other estimates (OR = 0.371).

Ok, so if we assume that all micromorts come from head injuries, then this suggest a reduction of around 2/3 in risk.


Results: A total of 43 studies met inclusion criteria and 40 studies were included in the meta-analysis with data from over 64 000 injured cyclists. For cyclists involved in a crash or fall, helmet use was associated with odds reductions for head (OR = 0.49, 95% confidence interval (CI): 0.42–0.57), serious head (OR = 0.31, 95% CI: 0.25–0.37), face (OR = 0.67, 95% CI: 0.56–0.81) and fatal head injury (OR = 0.35, 95% CI: 0.14–0.88). No clear evidence of an association between helmet use and neck injury was found (OR = 0.96, 95% CI: 0.74–1.25). There was no evidence of time trends or publication bias.

Ok, so this also suggest something around the 2/3 number, so I think that’s a pretty decent bet.

For arguments sake, let's just assume that all of the bad consequences of accidents are covered by measuring head-injuries, resulting in a total risk increase by a factor of 3.

Conclusion (very sketchy)

Ok, so if you add this up super naively in terms of micromorts, then we get a (1/3) * 3 = 1 micromort loss at a 1 microlife gain for my average office commute. If we commit a statistical atrocity and just add those up naively, we end up with a net gain of 0, suggesting that cycling without a helmet is roughly equally risky as doing some random sedentary activity.

There are obviously a lot of complicating factors to this, but I am not sure in which direction they point. I care more about short-term injury risk than I care about long-term gain from exercise, but I also care about the cognitive benefits of exercise and want to maximize my peak-potential more than my average potential.

I also know that you can’t just add micromorts to microlifes together, though it seems hard to figure out what the correct thing to do is, and I only set aside around an hour of time for this.

Overall, my conclusion is that if I have the choice between riding my bike without a helmet, and staying at home, I should probably be mostly indifferent between the two. Since the alternative is usually paying for an Uber to my office, or walking, or riding with a helmet which I find quite actively annoying, I think I will continue riding without a helmet for now, in the absence of me noticing some new considerations or evidence (or discovering some way to be less annoyed by helmets).

Or maybe someone feels motivated and does a better analysis than I did here and corrects me. My overall estimate on the importance of this isn't super high, so I probably won't do much more analysis of it.