Commentary: Why Are We Building Bikes Lanes That Are Hurting People?

Photos by Joshua Hart
Photo: Joshua Hart

As one of the certified bicycle safety instructors working with San Francisco’s Bike Ed program, the most important safety concept we try to get across to our students is that cyclists should never ride closer than 4 feet from any parked car. The reason is that getting ‘doored’ is the single most common cause of injury by motor vehicle users to people riding bikes in San Francisco.

People have been injured and even killed by riding in what is ostensibly a ‘safe’ space. Suddenly a door opens in front of them and they either have to swerve into motor traffic or hit the door itself.

Inevitably a student asks us, “But what about the bike lanes I see all over the city? A lot of them are totally within the ‘door zone.’ Where am I supposed to ride if the bike lane isn’t even safe?”

It’s a good question. Many of the city’s bike lanes have been built so that there is no clearance whatsoever between an open door and a passing cyclist. In the worst case scenario, a person with an older American car (Cadillacs have the longest doors) parks lazily, 2-3 feet from the curb. When they open their door, it can block the entire bike lane.

You might ask, isn’t it the responsibility of the driver to look over the shoulder before opening their door? And you would be right. CVC code 22517 requires that a vehicle occupant check for traffic before opening a door. Yet in the real world, people open their doors without checking all the time. And a person on a bicycle cannot reasonably be expected to look into every car to check whether there is someone inside. Under state law, cyclists have the right to ride where it is safe, which courts have affirmed is outside the dangerous door zone (CVC 21202).

Photo: Joshua Hart
Photo: Joshua Hart

So why are we designing bicycle facilities that are dangerous to use?  According to Caltrans chapter 1000 design manual [pdf], a bike lane adjacent to on-street parking must be at least 1.5m (just under 5 feet) wide. The minimum parking lane width is 7 feet (San Francisco Municipal Transportation Agency standards require 8 feet). Essentially that means if you are going to follow widely recommended safety practice and ride out of the door zone, but still in the bike lane, you have about one foot left to maneuver.

Photo: Bryan Goebel
Photo: Bryan Goebel

On San Francisco’s Fell St, the city’s designated east to west bicycle route, despite the new green paint, if you want to ride a bike safely, you are put in the unsafe position of staying within a one foot strip between the door zone and the motor traffic zooming by to the right.

Even if you are Kate Moss on a bicycle, you need more than one foot of width to safely maneuver. The American Association of State Highway and Transportation Officials (AASHTO) recommends that cyclists are given at least 4 feet of width (preferably 5 feet) to ride safely. Even if you are an expert at riding your bicycle in exactly a straight line, the likelihood is that you will veer into the door zone, or into the traffic lane at some point along your 3 block running of the gauntlet. If you follow official bicycle safety guidance, you should technically take the next lane over. But drivers see an empty bike lane and a “car lane” blocked by a cyclist in front of them and you have a perfect recipe for road rage.

So what is the solution? First off, bicycle design standards need to be changed to require cross hatch markings that extend 4 feet from a parked car so that even novice cyclists realize this a “no riding” area. If there is insufficient width on the street for a buffer and a bike lane, then a parking or travel lane needs to be removed. If providing for the safety and dignity of all road users is politically infeasible, then the bike lane itself needs to be removed and replaced with sharrows indicating that cyclists should take the travel lane.

It is no longer acceptable to lure people onto bicycles with a network of bicycle lanes that look inviting, but in reality can cause injury or death when used as directed. Our bicycle infrastructure design and our bicycle safety curricula should complement, not contradict, one other.


  • Neale

    I wonder whether in some cases the dooring problem can be relieved by diagonal parking, either in the centre of the street or at the curb. These are not without dangers, esp. cars backing into the bike lane. But in some circumstances they might be a better solution. Here are some pictures of such parking common many years ago in Vancouver, BC. (Photos from Vancouver Public Library).

  • smushmoth


    comment #97 is doing just that.

  • From #97:

    “(2) So what of combining bikes with pedestrians? It works reasonably well on trails where the pedestrians are all running or exercise walking. It works horribly on a sidewalk, where pedestrians do a half-dozen different activities, only one of which is going somewhere. Pedestrians mill about, dart from side to side unpredictably, and look one way while walking another way. This kind of pedestrian behavior generates many conflicts with cyclists. Yes, this has been formally measured in a study. I don’t have access to the study any more. I read it back in 1978.”

    emphasis on “It works horribly on a sidewalk”

  • patrick

    @smushmouth, the fact that something is illegal, particularly in traffic law, often has little to do with safety.

    Exactly who is advocating opening up sidewalks to bicycles? Certainly not me, as I already posted earlier, I’m perfectly fine with keeping it illegal.

    I didn’t even say anything about whether cycling on the sidewalk is safe or not, I simply asked Dave to define exactly what he’s complaining about cyclists doing, and then provide some evidence that there is an actual problem.

  • The SFMTA data for 2008 showed 24 reported injuries to peds involving bikes (almost 4,000 involving cars during the same period !). However only 3 were on the sidewalk.

    Obviously these incidents are under reported but nevertheless you have to conclude that the sidewalk riding bicycle marauder leaving a trail strewn with injured elderly pedestrians might be something of an exageration.

  • Dbarchitect –

    Can you specify where that information is for future reference?

    If we do some quick statistical analysis with it:

    24 / 3,500 (without the exact number, being generous): 0.7%
    Estimated vehicular mode share for bicycles in 2008: 6%

    Even if only 1/3 bike-ped collisions are reported (which says something about how severe the problem is[n’t]), they would only make up 2.1% of ped collisions.

    Even at a very generous 2.1%, a bicycle is 3x less likely to hit you than a motor vehicle.

  • The data I posted is in an SFMTA report on crash injuries for San Francisco, the one I dug through was 2008 I believe.

    In these discussions I think it’s difficult, though important, to separate analog effects, like that door popping open in front of you scaring your pants off, from what is the number of injuries/deaths from various situations, what the safety priorities should be. Bike lanes that are wide, 14′ is a sufficient parking/lane total width in my book (13′ being barely adequate, and 12′ being the threshold to switch to “Sharrows”) work very well and are very safe for a cyclist who sticks to the outer part of the lane. Putting down some more paint to try and keep riders out of the door zone seems like a good idea to me, but it’s a lot more paint. That idea has not been officially adopted by the SFMTA. Education about the door zone, also very important. Not striping lanes because the risk of dooring is too great? Probably not a supportable strategy, particularly when combined with the “safety in numbers” effect that is a direct result of striping more lanes.

    It’s all very complex, but going into drama mode doesn’t seem to serve end results very well, either of individual safety or of increasing overal urban cycling “comfort” and numbers.

  • David Baker,

    I’m forced to disagree with you on many points.

    I think you and I are on different operating systems. Yours, and forgive me if I misstate, seems to be just about entirely statistically based. As in, “If it hasn’t been shown to be a problem why sweat it?” Mine is mostly operationally based, using statistics as a guide. As in, “If X juxtaposition of people, places, speeds, distances, blind spots, sight distances, can be shown to cause a danger, we ought not to use traffic control devices to encourage or coerce people to put themselves in juxtaposition X.”

    Additionally, I urge you to plumb the Internet for the many available criticisms that absolutely skewer Jacobsen’s “safety in numbers.” The short criticism: the equation he uses to produce that nice looking “declining accident rate” is a mathematical tautology, that produces that curve shape regardless of the data you put into the equation. Three researchers I know have used three different random number generators to plug fake data into Jacobsen’s equation, and all three have gotten that lovely curve, from data that is all noise and no signal. Ergo, Jacobsen’s conclusion is a crock of nonsense, and relying on it instead of sound, proven design is professional malpractice. Think about it: you have a behavior which many thousands of people have proven is unsafe, such as door zone riding or blowing through stop lights. Without specifying a causal mechanism, Jacobsen blithely predicts that if more people engage in that dangerous behavior, it will become safe. Doesn’t that sound just a little too easy to be true?

    You give bike lane measurements that put people in the door zone, not to mention the startle zone. You want to squeeze a bike lane inside a 14-foot line, but only the outer few inches of that bike lane are outside the door zone. Center yourself in that bike lane and you ARE in the doorzone. It’s a bad place to be! And if the rider swerves to avoid it…. guaranteed, any collision resulting from the swerve is the rider’s fault. If it were a safe place to ride, cars would drive there. But the cars don’t. That should tell you something.

    Safety happens to individual cyclists for individual reasons. Looking at mass data, especially after guys like Jacobsen have polluted it, obscures that fact. A cyclist who claims the lane is far, far safer than a gutter bunny. A cyclist who obeys traffic signals wins brownie points (and jaw-dropping astonishment) from other road users. Plus he’s safer, and once he gets used to abandoning the messenger mode of riding, he’ll realize it lowers his stress.

    John Schubert

  • Jim

    @ John Schubert,

    I agree and disagree on your stance on taking the lane. It is certainly safer in lower-speed environments than gutter bunnying, but nearly suicidal on some higher-speed roads. The enmity and not-so-subtle messages generated by doing so, even on many sharrowed roads, will tell any rider where much of the car-driving public thinks the bike should belong. Nearly all instances of these types of aggression do not cause “accidents”, but certainly leave an impression.

    While the SFBC has done incredible things locally to broaden cycling awareness and spin (sic) cycling in a positive way, relying on the MTA to generate a bicycle infrastructure with no prior experience in designing one will lead to many structural errors, such as stated in the above excellent article. The MTA, to its credit, has been responsive to criticism of its work and has changed much of its early signage and striping.

    It’s really too bad cyclists have to play the role of crash test dummy beta testers for an infrastructure to mature.

  • @Jim,

    It isn’t like the SFMTA has to reinvent the wheel. Plenty of cities, even within the US, that have done this before and are several steps ahead of us. If anything, we are playing catchup which gives us the advantage of learning from THEIR mistakes.

  • EL

    Newsflash – This afternoon, an eastbound bicyclist on Market was hit by a southbound car on Van Ness. According to people standing at the corner that I spoke with, the bicyclist ran the red.

  • @EL – Your merit badge is in the mail.

  • b

    @EL – Any word on what happened? I was on the bus when I saw all that. It was pretty horrifying. Any idea if the biker ok? Terrible dreams last night about that…

  • Jim,

    There are fast roads I don’t take the lane on, but fewer than you might think. And I don’t get honked at (much). Avoiding the door zone puts me smack in the middle of the travel lane.

    If I think it’s okay to get over to the right for the convenience of others, I do. Point is, it’s my choice, and I don’t put anyone else’s convenience ahead of my safety. I choose when to do so. I do so often, but never when my safety is at issue.

    I don’t see bicyclists claiming the lane as a big deal in cities where the traffic is slow anyway. Motorists are already unable to maintain a speed above bicycle cruising speed. The presence of the bicyclist doesn’t change things for the motorist, other than reducing the top sprint speed between red lights.

    In the case of one-way streets with syncronized lights, I find that for my own driving (and I’m a brisk driver) I’m happy with a sync speed in the low 20s. So compromise — make it 19 or 20 — and all will be well.

    There are many excellent videos of lane-using behavior at One really instructional video compares the view of the road (and it’s a big, fast, six-lane traffic sewer) from the viewpoint of a bystander, a gutter bunny, and someone using the right lane. The first two views are really scary, and from them, one would probably think the third view would be the worst. But… the third view is entirely different and far, far more peaceful.

    John Schubert,

  • Al

    That animation at is really excellent. Everyone should watch it. Thanks for the link.

  • SF94114

    This is a silly article…there is no substitute for sharing traffic lanes on streets with parked cars. Bikes cannot effectively ride in car lanes. Folks, bike lanes next to parking lanes ARE A COMPROMISE…and as such, things are NOT PERFECT. Drivers will HAVE TO get used to checking for bikes BEFORE OPENING THEIR DOORS and bikers will have to keep an eye out for people getting out of their cars. REMEMBER: Having no bike lanes is far worse than having to look out for the occasional bad driver who doesn’t check for oncoming bike traffic.

  • SF94114, many people who have already posted would disagree with you. You can read back through the comments since I assume you didn’t.

    Bike lanes are a way to keep bikes out of the way of cars by putting them in a dangerous spot. If there was no bike lane, then the cyclist would be able to take the full lane which would “slow” drivers.

    Except for the ignorance of drivers, I would prefer to ride on sharrows. However, drivers think I have to be all the way to the right which again puts me in the door zone. If only there was some sort of test…

  • What mikesonn said!

    SF94114, you can see for yourself how well taking the lane works by seeing some of the many videos at I suggest starting with

    Most of these videos were shot in Orlando, hardly a bike culture city. The models are mostly women, mostly over 40. What they show is that people who AREN’T athletes, who AREN’T “young and fearless,” can easily use the safest possible practices — driving their bikes like vehicle operators, rather than being fearful road sneaks on the edge.

    Here’s why I don’t like SF94114’s concept of “compromise.” I make compromises in life every day, often willingly so. But I don’t compromise my personal safety for the illusory convenience of someone else.

    You can’t ride in the door zone safely. You can’t see inside cars to detect an impending door opening, and you should NOT be looking. There’s too much else to look at to navigate traffic safely. And riding in the door zone invites too-close overtaking passes from behind and makes you far less visible to traffic at intersections ahead.

    mikesonn, contact me at, so I can buy you a beer next time I’m in your area!

    John Schubert,

  • Another reason for bike helmet laws – NOW

  • Tony Belway,
    Spell out your arguments.
    I’ve worn helmets every day for 38 years, so it’s not like I dislike them.
    But I _do_ dislike laws based on poor assumptions.
    Helmets prevent only a fraction of head injuries (the ’85 percent’ statistic you may have heard is based on a long-discredited study that many experts believe is just plain junk science).
    Head injuries are only a fraction of all injuries.
    A helmet law is seldom enforced.
    Better to put a guard rail at the top of the cliff instead of an ambulance at the bottom of the cliff. In other words, give cyclists the knowledge to improve their own safety. Once they feel empowered that way, the overwhelming majority will choose to use helmets anyway. You’ll get more helmet use than you would have gotten with a law.
    It’s easy to write a helmet law with legal consequences that utterly screw cyclists. The language in the national Uniform Vehicle Code is a prime example of that. Hint: could you explain “liabilty exclusion” to a law school professor and get an A for your recitation? If not, you shouldn’t be telling people what should be in a helmet law. (No, I’m not a lawyer, but understanding some law is essential in this life.)
    Many laws that get written are pretty awful. Maryland’s new 3-foot passing law is a classic example. Cyclists are far worse off with this law than they were before it was passed. If you don’t know the risks you’re stirring up by asking for helmet laws, you’re in for a rude awakening.

  • It has come to my attention that drivers aren’t aware of what a sharrow is or what it means – both the term and the actual painting on the ground. I thought it was pretty straight forward, it marks the suggested location of the cyclist in relation to the lane. If drivers are unable to comprehend such simple and obvious lane markings, then we have a serious problem with driver education.

  • Mike – that isn’t a problem with driver education – that’s a problem with our general education. Which is getting indicted on a lot of fronts lately.

  • grosser schwanz

    The problem is a lack of common sense. You don’t need to be rocket or nuclear scientist to realize, collision of a bicyclist with a car, is 100s times more catastrophic for bicyclist, than for a walker, hit by bicycle.

  • Morton


    Actually, as a driver, the meaning of this so-called sharrow isn’t obvious at all. I took it to mean “be watchful – there may be cyclists on this route” i.e. it was a popular bike route although not a designated bike lane.

    So I took it as advisory not mandatory. I’d have figured the the “slower traffic keep right” rule still applies, as it does on every road.

  • Typically, the sharrows are located about as far right as a cyclist can safely go (when parked cars are present). So a motorist trying to pass should wait for a safe chance to do so, as they would on any road.

  • Peternatural, we should be so lucky. All over the country, I’ve seen absurdly too-far-to-the-right sharrow placement, instructing cyclists to be the victims of doorings, right hooks, and “I didn’t see him” intersection collisions.”
    Americans, including most bicyclists, have an odd belief that a bicyclist should cower out of the way of overtaking motorists under all circumstances. This creates many collisions.
    When a bicyclist takes a safe lane position, he is far, far safer. The _real_ delay to overtaking traffic is negligible. Since one principle of traffic law is “first come, first served,” and another principle is “pass when safe,” this notion that one vehicle operator has to cower for the illusory convenience of another vehicle operator, at detriment to his own safety, is loathesome.

    Mikesonn, thank you for your comment. It’s useful to hear how these things are perceived. I think that if you and I had a chance to go riding together, I’d be able to convince you that what is safest for cyclists is no real hardship for anyone else.

    — John Schubert,

  • grosser schwanz: everyone knows that.
    Here’s my ongoing gripe: I repeatedly hear people make that argument, apparently thinking that it proves that their opponent’s points about how to ride safely are invalid.
    I think we all agree to set aside the Darwin-award behaviors (ninja night riding, blasting through stop lights, etc.).
    What behaviors are you advocating? What behaviors are you discouraging?

  • Clutch J

    @ John You wrote–

    >In other words, give cyclists the knowledge to improve their own safety.

    Sure, cyclists can take steps to make themselves safer. Reading Effective Cycling was the single most important action I ever took as a bicyclist. I’m now at 18 years of daily bicycling on streets– approximately 10,000 trips– with no collisions or injurious falls.

    But from a collective advocacy perspective, educating individual cyclists is a losing strategy, as it’s a very rare bicyclist who’s willing to acknowledge that his or her skills are lacking.

    The preferred method to improving safety is creating safer environments by reducing the volume and speed of motor vehicles. Cycling-specific facilities are secondary but can help. Looking ahead, let’s hope the quality of bike facilities improves, as the deficiencies of bike lanes are well-known.

  • Mike F

    Great article!

    I think it raises a bigger question of bike facilities in general, not just bike lane width. I’ve recently started working (and subsequently riding) downtown and although I applaud the idea of the separated green bike lanes on Market, the way that they require you to weave in and out of traffic feels dangerous.

    The switchbacks seem haphazard and have inadequate signage for motorists, which is going to lead at least one of them to drive right through the turn-lane into the bike lane at some point. Again, I applaud the concept, but I can’t help but wonder fit the current implementation doesn’t put cyclists more at risk than before.

  • grosser schwanz

    I use bicycle as my primary transportation for 20 years+. Considering number of bicyclists, cars and pedestrians, most sidewalks can easily accommodate few bikers. I also drive van and motorcycle and observe idiots, who don’t know the rules and are deadly for bikers. While biker and walker are about the same.


A bike commuter high-fives an advocate for protecting him on the southbound Valencia bike lane during a May protest. Streetsblog/Rudick

Safety Vigilantes Strike Again on Valencia

Some 30 cycling advocates, wearing bright yellow t-shirts, stood along the southbound bike lane on Valencia Street between 16th and 17th streets and prevented Ubers, Lyfts and other cars from blocking this popular bike route during Friday evening’s rush. The protest, which emulated an earlier action on Golden Gate, was intended to ratchet up political pressure for […]

Should I Wear a Helmet Today?

The Naparstek boys riding last year’s Summer Streets event… wearing helmets. Sarah’s "Too Much Emphasis on Safety" post yesterday brings up the question in the headline above. A Canadian Broadcasting TV crew doing a documentary on biking is filming me as I take my two sons to school on our Dutch cargo bike today. While […]