The Prisoners’ Dilemma & Bicycle Helmets

OK, so here’s the Prisoners’ Dilemma:

Two guys are arrested on suspicion of being the pair behind an armed bank robbery. The investigating officer is cunning though, and separates them, making each an identical offer:

  1. Confess to the robbery, and thus provide the evidence to put your accomplice away for a long stretch, and we can cut you a deal - you’ll get to walk free.
  2. This works the other way - if your accomplice rats on you, he walks, while you take the rap.
  3. If you both confess, well I get TWO convictions, but I’ll go easy, and won’t object to early parole.
  4. If neither of you confess, then all I’ve got is this minor* possession of a firearm charge that I can stick on you both. But it’s nothing really, and you’ll both walk.

One of the things I teach in businesses is negotiation, and we use a variant of this as a game to help people understand the value of trust. The basic lesson is that there are good reasons for doing what’s right for yourself, but there are better reasons for doing what’s right for everyone.

Anyway. Bicycle helmets.

Over the last couple of weeks there’s been a low level flame war on the helmet issue running in the comments of the Yehuda Moon strip.

I generally don’t wear a helmet, as it’s just another thing to fuss with, it leaves my forehead sweaty & red when I get to meetings, and I question their magical life-protecting properties under most circumstances. I also worry about the message that the kind of ‘thou shalt wear a lid’ you hear from some quarters sends about the risks & rewards of riding a bike.

Because the risks are small (see below), while the rewards are huge - better overall health (like that of someone ten years younger), and life expectancy that’s overall two years greater than people who don’t cycle regularly.

But I also do believe that the potential upside of helmets is greater for different kinds of cycling. People who are just learning to ride a bike fall off quite a lot in their first few hours; people doing proper off-road stuff over lumps & bumps; BMX kids defying gravity doing jumps and tricks; people racing where they push the physics of cornering & braking, and particularly when they do this in groups or where fatigue is likely to become a factor.

But for the kind of cycling that uses a bike as a simple, pleasant form of transport? No, I don’t believe that the benefit’s particularly significant.

So I’ve been thinking about it more and more, and I think I can frame the helmet thing in similar terms as the Prisoner’s Dilemma:

  1. If I wear a helmet, then I may gain some safety benefit in the event of a fall. Potentially you get a downside if I don’t tell you about my new helmet-wearing ways, as the hot-shot legal team representing the driver who nearly killed you when he ran you over will now claim that your injuries were caused by your contributory negligence.
  2. If you wear a helmet, then you too might gain some benefit. And if I don’t know about your new ways, then I’m open to the ravages of the hot-shot lawyers.
  3. If every cyclist wears a helmet, then we all gain the same potential benefit. But we are all bear the other costs of this (both financial in buying the helmets every two years (explanations here and here) and social as I’ve explained above), to try and avoid something that is incredibly rare.
  4. What if neither of us wear helmets? There’s an obvious potential downside, as neither of us have anything between our heads and the tarmac in the event of a fall, or the windscreen of a car that hits us. BUT we do get to ride our bikes, and be seen to be doing so by the car-obsessed people around us. We can turn up to meetings looking smart in normal clothes, if not a little flushed (but never sweaty). We can use a bike to pop to the shops without worrying about another damned thing to carry around. We can show them all that it is a perfectly normal, safe thing to do. And maybe get more people to ride their bikes. Because of all the stuff you read on the Internet, all the statistics that the pro- and anti-helmet people YELL AT EACH-OTHER, the one that they all seem to agree on is the simple fact that there is safety in numbers. In societies where more people ride bikes, the accident rates are far lower. And at times of crisis, when people are driven out of their cars and onto two wheels (like, the ’70s oil crisis, the immediate aftermath of the London Tube bombings, and the gas price jump of 2007/8), accident rates drop.

So, this is the crunch. Do we take the prosecutor’s offer of looking after ourselves? Or do we take the counter-intuitive, potentially not best for the individual route, which may ultimately yield the best overall result?

Please do feel free to comment. But no shouting!

(As I write this, I’ve a feeling that it can’t be original - surely someone else has made this connection? If you have, and I’ve read it but then forgotten it on all but a subconscious level, please post a link below!)

Filed under: 'A'-List Blogs, Assassination Attempts, Bike Culture, Bike to Work, Everyday People, Ranting, Road Safety, helmet

17 Responses to “ The Prisoners’ Dilemma & Bicycle Helmets ”

  1. Treadly and Me on April 7, 2009 at 2:05 am

    Just to nit-pick a bit, I don’t think your bike helmet case is a true example of the Prisoner’s Dilemma. The underlying idea of the classic Prisoner’s Dilemma is that the prisoner must make a one-shot, irreversible decision that implications for both the prisoner and a known other (the other prisoner).

    The decision to wear or not to wear a helmet is neither one-shot, nor irreversible. You can choose whether or not to wear a helmet every time you ride.

    In options 1 and 2, the down side is not borne by a specific and known other, rather it potentially happens to many unknown others. And the reverse is true: the individual helmet-wearing decisions of others have a tiny potential impact on me. So in the context of the Prisoner’s Dilemma you’re probably not motivated to care so much about how your decision affects the other.

    However, your options 3 and 4 are where things get interesting. If all (or almost all) make the same decision, the outcomes can be quite different. Not least, the perception of everyday cycling as a ‘normal’ or ‘risky’ activity.

    In passing I’d note that in a jurisdiction where bike helmets are both compulsory by law and the norm behaviourally, Ian Walker’s finding that motorists use the presence of a helmet as an indicator of cyclist competence and predictability is not valid. I would suggest that this is probably a good thing, although more a side-effect in importance.

  2. Pippi on April 7, 2009 at 3:07 am

    I just found your site and I have to say this is the best helmet post I’ve read. This is exactly my opinion but you’ve expressed it much better than I usually do ;)

  3. electric on April 7, 2009 at 3:36 am

    Sorry but your Prisoner’s dilemma is weak. I will be wearing a helmet. Is that snitching on you? No.

    “Essentially, however, it is alarming that the Judge concluded that, in general, a cyclist’s decision not to wear a helmet may be regarded as “contributory negligence” (i.e. compensation for someone else’s negligence should be reduced on the basis that the cyclist was also partly to blame for their own injuries), in cases where the helmet would have made a difference.

    So, if the helmet makes no difference don’t worry. You go right ahead and not wear it. BUT maybe you feel you’re being persecuted by big bad lawyers and the ‘man’. What else do you want, to be absolved of taking the proper precautions? Maybe they can’t prove it makes a difference. The thing is, I bet it does… because most accidents won’t be of the ten-tonne truck going 80km/h over you variety(good God). Those only account for a small percentage of events, most of the events you’ll encounter will me more moderate and happen in a manner where the helmet does make a difference.

    Don’t get me wrong, i’d love to ride without it(and not have to face the added consequences) but, honestly, when i do it feels like i’m playing russian roulette and i just put another round in the chamber. *click* Ah. another ride to work without bad hair and red spots *click*…. *click*…

  4. Karl On Sea on April 7, 2009 at 5:39 am

    Yeah, it’s not a true Prisoners’ Dilemma, due to the marginal upsides from ‘not ratting’ on the other prisoners, and the fact that there are many repeats of the ‘experiment’ - every time you get on a bike in fact!

    The point about the ‘real’ one only being played out once is interesting. In the version I use in my negotiation workshop, I get groups of people to play ten rounds of the game. Rather than time in prison, they play for money (real money, but not very much of it), as this adds a whole bunch of realism to the behaviour. The game is played with as many groups / people as possible, so that there are more people to ‘rat’ you out that you have to trust to ‘keep quiet’. We play seven rounds where no inter-group discussion is allowed, and three rounds with raised stakes (2x, 5x and the final round, 10x) where all players can discuss how they’re going to vote.

    Almost without exception, human greed & self interest are not overcome. Someone will almost always choose what’s right for them rather than for the society as a whole. From this experience, I don’t think there’s a huge difference between single- and multiple repeats of the test. Of course, ten rounds isn’t actually that big a number, and my guess is that if you played a LOT of rounds (like the computer in that film, War Games did), you would get different results.

    But it’s not like Russian Roulette either. The odds there are 1 in 6, and if those were the odds of dying whenever you got on a bike without a helmet, then I’d just never get on a bike at all. The death rate is reported here as around 1 per four million hours’ cycling, or just over half that seen in drivers & passengers of cars. Yet we don’t wear helmets to drive / be a passenger in a car, unless we’re actually doing something ‘risky’ like taking part in a race. Is cycling any different?

  5. spacemonkee77 on April 7, 2009 at 6:27 am

    Brilliant post, game theory is a clever way of structuring the debate, not perfect, but very useful.
    that link is a fantastic example of the two sides of the argument squaring up to each other. as an “anti-helmet” person, perhaps i am inadvertently cherry-picking but the pro’s seem to rely on personal anecdote and the position that its a benefit, not a cost, so hey, why not? whereas the anti’s seem to offer data to back up argument.

    Round here, on my commute into work it seems to be about 75% helmet wearing, which I find surprisingly high and worrying, as “the government ” has been quoted as saying they won’t bring in helmet legislation whilst helmet use remains low, which is another reason why I am an actual anti rather than an “it’s up to you”. Say it loud, I’m anti and proud! ahem. Hope that’s not shouting, more declamatory.

  6. Ken on April 7, 2009 at 7:47 am

    Hi. For Me It’s About Choice.

    I Cycled Over Forty Years (25 In London) Without One (That Includes The Broken Clavical & Head Injuries (Never Been The Same Since LoL) Only In The last Year Have I Wore A Helmet.

    For Me I Cycle Right Through The Year So Black Ice Can Be A Problem (Came Off Twice On The Stuff This Year) Also There Are Some Nature Trails With Gravel Etc Near By. Where Dogs Are Unsupervised Etc.

    So For Me It’s About Weighing Up The Journey. Sometimes I’ll Just Cycle Without One Or Have It On The Bike.

    So As I Said It Should Be About Choice. Altho I Was Gratefull When When I Came Off In The Black Ice (Being Older You Feel These Now)

    All The Best

  7. Andy in Germany on April 7, 2009 at 11:51 am

    I see the point, dunno if I’d put it that strongly though, as in the game the person would be understanding this as a possible consequence, whereas it’s a bit more indirect in this case.

    On the other hand, if cyclists did stand up and refuse to wear helmets it would stop lawyers using them against us.

    I just wish people would stop trying to bully me into wearing one…

  8. hardly on April 7, 2009 at 3:18 pm

    The problem with this logic is that it ignores basic human nature; we just don’t get long-term effects very well.

    Better frames on this would come out of the decision-making sciences, not the art of negotiation.

    Like this: If you spend $100 every other year, you can save $X (where X = odds of a head injury while cycling * costs of being down w/a head injury).

    I suppose the one problem with this is the Natasha Richardson cost, e.g. death = ?

    Additionally, this whole argument reminds me of the argument we had in the US before seat belts were compulsory wear. Since they went into effect, the highway death rate has plumetted to new lows.

    If you don’t want to be safer, fine, you absorb the cost of that. But please, start breeding after you stop bicycling.

  9. Karl On Sea on April 7, 2009 at 4:00 pm

    We had the same thing here too - on both seat belts and, if you can believe it, drink driving. I remember seeing archive footage of some fine upstanding member of society getting very incredulous when asked her views on proposed drink-driving laws: “But that’s an outrageous assault on my civil liberties. If I want to have a few drinks and then drive home, I know that I’m in control of my car!”

    In the UK, only 5% of people don’t wear seat belts, yet they account for 29% of in-car fatalities, meaning that they’re six times more likely to be killed than if they’d belted up (anyone better than me at this sort of stats, please feel free to correct this).

    What’s missing from the discussion on cycling though is a sense of proportion on the risks. The risk of head injury in cycling is lower per mile traveled than for pedestrians, and less per hour spent doing it than for motorists. Does the logic that helmets should be worn by cyclists doing regular everyday things (i.e. not racing, not mountain trail riding, or not learning to ride a bike for the first time) also apply to motorists & pedestrians? Ought we to be campaigning for pedestrian helmets, and perhaps mandatory driver & passenger helmets for people in cars?

    I don’t think so.

    Would mandatory helmets reduce cyclist head injuries & deaths?


    But experience has shown that this has stemmed largely from reduced numbers of cyclists being on the road, and that this is accompanied by an increase in the per-cyclist risk as they have less of the safety in numbers effect on their side.

  10. hardly on April 7, 2009 at 7:35 pm

    I dunno, I get the feeling that there’s a mixing of statistics and that tends to make me think that a dodge is being played.

    A number of years ago, I went in for my physical and the Doc asked me a bunch of risk factor questions, e.g. how much do you smoke, how much do you drink, do you work out, etc. I proudly told him of my bicycling regimen and he said, “Well, that’s certainly going to improve your cardio health, but you’ve significantly increased your risk of injury due to an accident…”

    The issue to me is not simply risk of a given event. It’s also the cost of that event. Risk * cost = the ability to make an informed choice.

    For example, my guess on low head injuries in cars is due to cars getting side-impact air bags so the hit the head makes to the window when the car is struck from the side is cushioned. And, this is a feature that people want to buy because they understand that when their head impacts a solid object at 35+mph they are going to take significant punishment.

    Too, I probably fall into the category of the ‘converted’ because of a bike accident I had years ago where a helmet definitely saved my bacon (helmet was split open, so it did its job, and my head was fine). Based on that experience, I’ll continue to wear a helmet, and I’ll strongly recommend the same to everyone I know.


    Because you never know.

  11. Brad Hefta-Gaub on April 7, 2009 at 10:39 pm

    I’ve said it before, and I’ll say it again… I love this topic because it really makes me think. I am very much torn on this issue.

    I want to see more people riding their bikes. I am convinced that the number of people who get sick and/or die from heart disease and diabetes is far greater than the number of people who will have serious injuries from riding a bike. And so if the perception that bicycle riding is dangerous is really dissuading people from doing it, I’d like to see us take steps to convince people it’s safer.

    But, here’s where I get stuck.

    When I was that fat kid, and that fat executive, I saw people riding bikes all the time… and NEVER ONCE did I think… gosh that looks dangerous, after all they’re wearing helmets, it muse be really dangerous.

    Nope, my reaction was even more shallow than that… and I bet its the reaction that 99% of people have when they see cyclists… My reaction was: I don’t want to wear those tight fitting clothes, my butt will look sooooooo fat in those pants!

    Once I lost 50lbs, and was actually ready to start riding a bike, I was still afraid of looking silly. I vividly remember walking into Gregg’s Cycles and asking the clerk for help in finding some “cycling clothes” that wouldn’t look so silly. I didn’t want to wear fancy brightly colored jerseys and tight fitting pants.

    I think you had a great article about this recently Karl, the concept that people are mostly turned off because they feel like they don’t belong. Its one of the reasons I have reservations about Critical Mass… I think it’s too easy for people to feel out of place, or unwelcome… and I don’t think it’s the “helmet” that really makes people feel unwelcome.

  12. electric on April 8, 2009 at 1:11 am

    The example of russian roulette doesn’t take into account the odds, exactly. Let me rephrase.

    If you played Russian roulette with two handguns, haha!

    Handgun 1 has many small caliber bullets which can’t penetrate steel.
    Handgun 2 has 1 large caliber bullet(it says .50 on the side) this bullet can penetrate steel.

    For each hour you cycle you’re forced to pull the trigger on gun 2 and then gun 1.

    Now lets say our theoretical revolver can hold 4 million bullets, but we’ll only put 1 in the large handgun(handgun 2). A fired bullet from handgun 2 represents you losing the game. This act is equivalent to the above average of 1 death per 4 million hours of cycling quoted above.

    Now lets say for each death on the road there are another 200 accidents of moderate seriousness(enough to involve police). Lets assume that these accidents fall under the dear judge’s “contributory negligence” clause which means a helmet can be PROVEN to have prevented injury. To take into account those 200 accidents. we’ll put 200 smaller caliber bullets into handgun 1.

    Ok! So lets put on a our metaphorical russian roulette cyclist helmet, this one is made of 10mm of steel! ;)

    In 10 years we’ll bike 20,000 hrs. So we’ll pull the triggers on the handguns 20,000 times.

    Now we shouldn’t be surprised when sometime during those 20,000 pulls a bullet is fired from handgun 1. After all 4,000,000/200=20,000. Maybe handgun 1 will go off a few times or none at all(the latter is more dangerous for you, because with each pull you gain a greater feeling of safety, “nothing happens! i bet those jokers didn’t put any bullets in this gun haha!”) So one day you’ll take off your steel helmet, feeling reasonably sfve in your experience so far, and play the game. After all that steel helmet is heavy and it’s hot out today.

    What you’ve done, is taken those 200 or so bullets from handgun 1 and upgraded/loaded them into handgun 2. Now your odds of having a quite serious accident today are 4,000,000/201=19,900 in 1. This is in comparison to 4,000,000 in 1 when you’ve got the steel helmet on.

    Now. I know the gun metaphor is somewhat misleading.. since it makes no allowance for the open nature of real life. i.e. each hour we’re given a new gun. So you won’t get a situation where you’ll be the guy who has to pull the trigger on the last un-tested chamber. Anyways i digress your odds are still 19,900 in 1. Odds. Not fate. but still… people expect to win the lottery everyday. look at those odds.

  13. town mouse on April 8, 2009 at 8:49 am

    I think one of the things the statistics don’t capture is that people don’t just randomly fall off bikes unless they’re five years old or learning how to ride. They come off their bikes because they’re drunk, because they’re reckless, or because they’re knocked off. Not cycling drunk and not cycling recklessly are under your control, and to a certain extent so is not being knocked off.

  14. Karl On Sea on April 8, 2009 at 9:26 am

    Nice modification of the Russian Roulette analogy. I think you’re wrong with the adjustment of the odds, as it’s 4,000,000 to one for all cyclists, not just those with helmets, but it’s still an interesting way to think about it. Crucially, I still don’t think I’d play, even with a 4,000,000-chamber revolver!

    In the UK something like just 35% of cyclists wear helmets, and yes they do account for disproportionately fewer head injuries than those without helmets. It’s estimated that the benefit of helmets in relation to all falls and accidents is around a threefold reduction in head injury risk (Maimaris, Summers, Browning, Palmer. British Medical Journal Vol 308 pp 1537-40, 1994. Head injuries reported in 11% of non-helmeted hospital cases, and in 4% of helmet wearers; by the time other factors were taken into account, it was estimated that a helmet provided a 3.25 protection factor.)

    In other words, of that 4,000,000 man-cycling-hours between deaths, around 3,600,000 are taken by non-helmet wearers, and 400,000 by those with helmets. With those kinds of numbers, it’s easy to see why legislation mandating helmets gets passed.

    BUT, whenever helmets are made made mandatory, this is accompanied by a decrease in overall cycling. Logically it shouldn’t, but whoever said that people are logical? Also, non-third world countries with the lowest risks for cyclists seem to be those with the lowest helmet use, but critically the highest number of cyclists, and the highest standards of cycling infrastructure.

    Head injuries are just under three times more likely when a motor vehicle is involved in an accident with a bike - provide infrastructure that separates bikes and cars, and crucially makes cycling convenient (no stop-start routes going places that no-one goes to / from) and you get the same safety improvement as you would with mandatory helmets, while at the same time making cycling more attractive to the 98% who don’t use a bike as basic transport.

  15. Tim Beadle on April 8, 2009 at 3:41 pm

    Interesting stuff. I’d not encountered The Prisoners’ Dilemma before, but I have emcountered the Tragedy of the commons:

    It “describes a dilemma in which multiple individuals acting independently in their own self-interest can ultimately destroy a shared limited resource even when it is clear that it is not in anyone’s long term interest for this to happen.”

    Sounds like it should apply to cycle helmets (i.e. an individual’s seemingly rational and sensible self-interest making things worse for the collective), but I can’t find anyone make the explicit link between the two elsewhere.

  16. Karl On Sea on April 8, 2009 at 3:54 pm

    Tim - that’s a superb, and probably superior way of thinking about this problem - thanks for pointing it out!

  17. Learning to Ride a Bicycle on November 29, 2009 at 2:16 pm

    [...] Do The Right Thing [...]