John S. Allen's Bicycle Blog

A fundamental principle of traffic engineering is that the road user who is required to yield right of way (in this case the motorist) must be able to see the one to whom yielding, unless a traffic signal establishes who must yield.

Bike boxes stand in stark conflict with this principle by having bicyclists swerve out in front of waiting motor vehicles at a traffic signal. Traffic engineers seem to have decided that the fundamental rule does not apply to bicyclists, as it does between motorists — or for that matter, between motorists and pedestrians.

The much-derided John Forester had an explanation, cited here:

Highway people had training deficiencies because of the overwhelming success of motorization. They never had to teach any traffic engineers how to drive. They never had to teach the theory of traffic safety — the theory was implicit in everyone’s driving knowledge. Therefore, these people never questioned the principles of the ‘bike safety training’ they had received. They didn’t recognize that it conflicted with the theory behind vehicle safety.

But in the case of a bike box, the disconnect is so stark and obvious, that I just have to wonder.

Research on bike boxes in Portland, Oregon showed that about 95% of bicyclists do not use the bike box. The bicyclists understand the risk of swerving into the bike box. They are less cautious about overtaking on the right to pull ahead of the first waiting motor vehicle, as also shown in the photo at the top of this article. Fatalities have resulted.

This bike box in the photo above has  much worse problems than a square-across bike box, because bicyclists enter the bike box diagonally from behind. The Portland, Maine senior engineer’s explanation of this bike box is based entirely on the static geometry of the intersection. It pays no regard to the inability of motorists to see bicyclists to yield to them. The bicyclists approach outside the field of view of the right-side rear-view mirror. If the vehicle has windows behind the front row of seats, motorists could swivel their necks more than 90 degrees and see approaching bicyclists as long as there is no vehicle waiting in the next lane to the right. Without windows behind the right-side front door, or with passengers or cargo blocking the view, bicyclists approach unseen.

The bike box was painted in 2020 (also see this) and mercifully had largely faded away by November 2022, suggesting that it lacked full support of the city.

Here are the requirements for a bike box as stated in the MUTCD interim approval. The Portland one does not or may not comply with some of them, as described below.

A bicycle box shall be formed by an advance stop line placed at least 10 feet in advance of the intersection stop line.

At least one bicycle symbol shall be placed within a bicycle box (see Attachments IA-18-1 and IA-18-2 for placement details).

This is in compliance.

Where a bicycle box is provided across multiple lanes of an approach, countdown pedestrian signals (see Section 4E.07 of the 2009 MUTCD) shall be provided for the crosswalk across the approach on which the bicycle box is located to inform bicyclists whether there is adequate time remaining to cross to an adjacent lane before the onset of the green signal phase for that approach.

I can’t tell whether countdown signals are at this intersection from the photo but in any case, there is no crosswalk across the approach on which the bike box is located. The lack of one reflects an understanding of the same problem that the design of the bike box ignores: pedestrians approaching from beside and behind waiting vehicles. There is a crosswalk across the far side of the intersection.

Turns on red shall be prohibited from the approach where a bicycle box is placed using a NO TURN ON RED (R10-11 series) sign.

I don’t see a NO TURN ON RED sign at the far side of the intersection. I also can’t tell from the photo whether right turns are permitted from the lane to the left of the bike box.

At least 50 feet of bicycle lane should be provided on the approach to a bicycle box so bicyclists will not need to ride between lanes to enter the bicycle box.

I can’t tell from the photo. Note that the MUTCD statement is defective in that a bike lane may be to the right of all the other lanes. It is a “should” rather than a “shall”, so it is not a firm requirement.

A STOP HERE ON RED (R10-6 or R10-6a) sign should be provided at the advance stop line defining the bicycle box with an EXCEPT BICYCLES (R3-7bP) word legend plaque below (see Attachments IA-18-1 and IA-18-2).

As of November 2022, there was none. What “stop here” would mean is unclear as the stop line is

Green-colored pavement (see Interim Approval No. 14) may be used within a bicycle box and the approach bicycle lane, where one is provided. A separate request for Interim Approval for green-colored pavement is required if the agency has not already received such an approval.

Has it? I can also add: the MUTCD diagram in the Interim approval does not show a diagonal bike box. This design was never contemplated. However, design is supposed to reflect engineering judgment. A facility which does not allow safe yielding reflects poor engineering judgment. What specific approval has the City of Portland received for this unusual and troubling design, if any?

Is the bike box swerve accommodated in Maine traffic law or Portland ordinances? I don’t think so.

Also, the MUTCD is authorized by the FHWA. The NACTO guide is not an official guide and does not reflect engineering standards.

I expect that the bike box will see very little use. Bicyclists will ether merge before turning left, or go to the two-stage turn queuing box at the far right side of the intersection. Either of those maneuvers is accepted and normal.

I also expect that motorists will regularly encroach into the bike box, and bicyclists will not use it, as at this one in Boston:

An article contends that Dutch people are competent bicyclists because they can ride while carrying an adult passenger on the rear rack or the handlebar, or holding an umbrella, or walking a dog. The author disparages two authors who teach bike-handling and traffic skillsl: s”‘[c]ompetence’ for Franklin – just as for Forester – means the ability and willingness to ride in motor traffic.” and that Dutch bicyclists simply would prefer not to. “Of course, those Dutch riders were in an all likelihood perfectly capable of cycling on British roads – they just didn’t want to.”

People get good at improvised and risky bicycle acrobatics by having to use a bicycle that is not designed for the particular use. Knowing how to mount and dismount gracefully, propel the bicycle efficiently, use the gears, scan for overtaking traffic (including on bikeways and quiet streets, where it is no less necessary) are another type of competence and from what I have seen of Dutch cycling, these skills are neither taught nor widely practiced in the Netherlands. My observations on this: https://john-s-allen.com/blog/2019/05/dutch-bike-handling-skills/ and https://john-s-allen.com/blog/2019/08/another-embarrassing-dutch-bicycle-education-video/

Cyclist and planner Andrew Besold comments:

I must say that I’m a cyclist and I was all for the Idaho Stop Law. Then I lived in Sun Valley, Idaho for a year and saw this law so totally abused by local cyclists that every intersection was total chaos.

It all sounds good on paper but human nature is to push the rules until they are totally broken. The end result was bicyclists blowing through stop signs and red lights even when high-speed motor traffic had the right-of-way. Since local cyclists interpreted this law to mean that they didn’t have to stop at all (yes, one yelled this at me when I confronted him about it after he nearly ended up in my windshield!), local drivers have no choice but to stop for cyclists even when the drivers have the ROW according to the law. That might sound like a free bonus but as a cyclist who tried to follow the rules and knows a thing our two about traffic dynamics from 30 years of riding, yes the one driver sees you and is letting you go but the one in the other direction or in the far right lane didn’t and if I would have proceeded would have been hit by the unsuspecting driver (it’s a hazard called double jeopardy). End result? Total chaos that would have me waiting for minutes to clear an intersection that should have taken a dozen seconds.

Reply from blog host John Allen: Let me take the long view: bicyclists’ taking license (in the sense of licentiousness, not drivers’ licenses) can go either of two ways. One is the direction in which populist advocates would like to go: that bicyclists who are unable to operate according to the rules of the road, or would rather not have to, come to control the expectations for road use. This is tied in with the “8 to 80” concept: and with what John Forester called the “toy-bike syndrome”. Bicyclists are then a special class of road users, with special legal protections, as are pedestrians, recognizing that they may be, as the saying goes, of all ages and abilities. Other road users, and in particular, motorists, have to watch out for the unpredictable bicyclists and are held responsible for crashes which bicyclists cause by violating the rules of the road. There is a mythology, wishful thinking by US bicycle advocates, that strict liability law in the Netherlands always holds motorists responsible in collisions with bicyclists. That is what the American advocates would like to believe. Actually, this law applies only to damage claims, same idea as no-fault insurance in the USA, as described on the BicycleDutch site, https://bicycledutch.wordpress.com/…/strict-liability…/. Despite the great political power of bicyclists in the Netherlands, that is as far as strict liability goes. It also is fair to assume that Dutch bicyclists are not as licentious as US ones, by virtue of more experience on average, mentoring etc.

In the USA, where motoring dominates, the licentious behavior of bicyclists is more likely just to lead to bikelash, in my opinion. And advocacy based on championing bicyclists’ licentiousness also results in preventable injuries and deaths.

Here’s a video from BicycleDutch.

For the cyclist, this is a variant on the two-stage turn queuing box (2STQB), or what I once called the “cross-street bike box” — but with a couple of significant differences.

A 2STQB as in the USA on the far right corner is used only for left turns. In the Dutch design, the cyclist who is traveling straight ahead heads into a 2STQB before the cross street, and if turning left, then another one to make the left turn.

Dutch drivers are required to yield right of way to cyclists both when entering and when leaving the intersection, just as they would to pedestrians. Dutch drivers understand this and do this — however, if bicycle traffic is heavy, with serious effects on efficiency and capacity for motor travel.

The video says nothing about signalization, if any, and it completely avoids the issue of delay.

If you step back and look, this reveals itself as a conventional four-way intersection inside a roundabout. Cyclists travel in the roundabout, and motorists must cross it when entering and leaving.

“This is a standard US road from the latest urban design guide with the official lane widths.”

Wrong. The reference is to the NACTO guide, which does not have official status. The NACTO guide was created by advocates for greater bicycle mode share, and includes numerous treatments which have no official status and which create conflicts. The treatment shown would be regarded as an experimental treatment by the US Federal Highway Administration. The double weave, with one lane crossing another, is not an approved treatment in the official reference, the Manual on Uniform Traffic control Devices. Furthermore, this treatment has been used in the USA only since approximately 2008. As of 2025, NACTO also has disowned this treatment.

The narration in the video says that motorist left turns can often be made as quickly as they could be made in the street, but an uninterrupted left turn can only be made on a stale green, so the green for the cross street starts just as the cyclist arrives on the far right corner.

I have another post commenting on an American consultant’s promotion of this type of intersection, linking to more information from BicycleDutch, and raising additional issues.

In the crude “it just sort of happened” narrative, the driver bears no responsibility, and in the cosmic New Age narrative, the driver bears no responsibility. Got it.