[david-gpu]

> The bicycle infrastructure solutions the bicycle crowd comes up with increase the risk of collisions at intersections

Either cyclists don't know what sort of infrastructure makes them safe, or you have an imperfect understanding of the sort of infrastructure that they would like to see.

> The one way to reduce the risk of intersection collisions is

That is one way, but not "the" one way, nor the best way. Dutch-style intersections are probably the state of the art solution when sharing the road with cars is unavoidable. Car-free cyclepaths are even safer. There are other means as well, including the elimination of right turns on red, which are particularly dangerous to pedestrians as well.

[u801e]

> Either cyclists don't know what sort of infrastructure makes them safe

Many cyclists have had no education, training or classroom instruction on how to cycle safely in traffic and have a distorted view of what infrastructure can do for them in terms of safety. For example, this cyclist[1] ended up in a crash because he failed to foresee the situation that could have easily been avoided. He evidently thought that the protected bike lane he was using made him safer. Yet, he could have easily been run over after being pushed out into the roadway. Someone with education and training would have realized that the motorist was not looking in their direction and they should anticipate that they won't yield to them.

> or you have an imperfect understanding of the sort of infrastructure that they would like to see.

I've seen plenty of examples of infrastructure that increases the risk of the collisions I mentioned earlier because the cyclist is hidden from the motorists' view until shortly before both arrive at the intersection. This doesn't give the motorist or the cyclist enough time to determine which of them should yield.

Infrastructure that relies on traffic lights to provide a protected movement through an intersection is the best solution in those cases, but results in longer wait times for everyone. This leads to non-compliance with traffic control signals and people who will try to beat the light to avoid a several minute wait. Unfortunately, most infrastructure I've seen relies on mutual yielding to work. Mutual yielding will work with both are moving at walking speed, but not at vehicular speed.

> Dutch-style intersections are probably the state of the art solution when sharing the road with cars is unavoidable

This doesn't address the numerous mid-block intersections where there isn't sufficient room to install one. Second, these intersections are geometrically similar to modern roundabouts (from the point of view of a motorist making a right turn at one), yet one study[2] has shown that around 71% of motorists exiting a roundabout yield to pedestrians waiting to cross or within a crosswalk. Presumably, the rate of yielding for cyclists are are moving at 15 to 25 feet per second instead of just 3 to 5 feet a second would even be lower because the motorist would be less likely to see them because the cyclist would be further away from crossing the intersection.

> Car-free cyclepaths are even safer.

They are not available in all cases.

> There are other means as well, including the elimination of right turns on red, which are particularly dangerous to pedestrians as well.

But this doesn't address the right on green problem, which is when most right hook collisions happen.

[1] https://www.youtube.com/watch?v=4k6-AI_X1qE

[2] https://nacto.org/docs/usdg/nchrprpt572.pdf (figure 63 on page 83 in the PDF)

[LarryMullins]

> [...] protected bike lane [...] https://www.youtube.com/watch?v=4k6-AI_X1qE

Surely that's not a protected bike lane at all? What's it protected by, a few flimsy sticks of plastic? I assume that when people talk about wanting protected bike lanes, they're talking about concrete jersey barriers at least.

To call that bike lane "protected" is like printing "please don't shoot me" on a tshirt and calling it a bullet proof vest.

[u801e]

> Surely that's not a protected bike lane at all?

Unfortunately, the definition of protection is not standardized, so one may consider flexi-bollards, parking stops, bollards, parked cars, planters, concrete curbs, and jersey barriers forms of protection. The standard term for a facility like this is cycletrack, but that term is not as commonly used.

For what it's worth, the term protected, in a traffic engineering context refers to exclusive movement through an intersection based on traffic signals. The most common example is a protected left turn as indicated by a left green arrow on a traffic signal. The other option for the left turn on either a solid green light or flashing left turn arrow is a permissive left where left turning traffic has to yield to oncoming traffic. Ironically, protected bike lanes rely on cyclists making a permissive rather than a protected movement for every intersection they traverse.

[Beltalowda]

> For example, this cyclist[1] ended up in a crash because he failed to foresee the situation that could have easily been avoided

In general I agree that with some defensive cycling (and driving!) you can prevent a lot of accidents, but I'm not sure if I agree with that example; the car seemed to stop for the cyclist, and then it started moving with about a second for the cyclist to react.

Usually "eye contact" is the best bet, but it's near-impossible judge from that video if that was there or seemed like it was there. It's certainly possible the driver looked to the right in the direction of the cyclist and simply missed him due to situational blindness. Cyclist assumed driver saw him, driver didn't really register the cyclist, with the video as a consequence.

What reasonable expectations are also depends on how common cycling is in the location, how common that sort of cycle path is, stuff like that. I don't even know where that video was filmed. It's always easy to judge these things after the fact from a video sipping coffee from behind your desk, but in real life it's very easy to interpret something wrong, make a mistake, or just not pay attention for 3 seconds.

I take some amount of issue with the phrasing "this cyclist ended up in a crash because he failed to foresee the situation". He ended up in a crash because the driver ran in to him. He could perhaps have prevented the crash by correcting for the driver's mistake – which would clearly have been a better outcome, and is also why these videos are useful so we can all learn from them – but it's still primarily due to the driver's mistake that the crash happened.

[u801e]

> I'm not sure if I agree with that example; the car seemed to stop for the cyclist, and then it started moving with about a second for the cyclist to react.

This comes down to the difference in speed of a walking pedestrian versus a cyclist. A pedestrian walking at 3 mph is moving at about 4.5 feet per second. A cyclist moving at 10 mph is moving around 14.5 feet per second. This means that a pedestrian a second away from crossing the street would only be 4.5 feet away and within the view of the motorist approaching the intersection. The cyclist, on the other hand, would be 14.5 feet away and outside of the field of view of the motorist unless the motorist made the conscious decision to look down the sidewalk to check for approaching cyclist traffic.

The video shows that the motorist is turning onto a one way street where traffic is approaching only from their left. Naturally, the motorist is going to check for approaching traffic from the left by looking at the roadway. They will see a pedestrian about to start crossing in front of them because they would still be in the field of view, but a cyclist coming in the opposite direction moving at least a third of the speed of traffic coming from the motorist's left will definitely be outside of their view.

To increase intersection safety, the number of conflicts a driver has to account for needs to be minimized and separated timewise. When you compare a conventional intersection to a roundabout, you can see that the driver preparing to enter only has to account for traffic coming from one direction and pedestrians crossing in front of them. They also have a degree of time separation between the conflicts they have to deal with when entering the roundabout and those they have to deal with when exiting it. In a conventional intersection, there really is no time separation, meaning that the motorist has to deal with more conflicts in a given period of time.

In the case of the intersection shown in the video, the number of conflicts increased because the motorist has to now check for cyclists approaching in either direction in addition to traffic and pedestrians. The more conflicts one has to check for, the greater chance that one will be missed which increases the chance of a failure to yield and possibly a collision. Just placing the blame on the motorist without looking at how the design contributed to this situation won't solve the problem.

I will say that had there been shared lane markings on the street and had the cyclist been riding in the center of the lane in the same direction of traffic, this collision would not have occurred because the cyclist would have been where the motorist was checking for traffic.

> Usually "eye contact" is the best bet, but it's near-impossible judge from that video if that was there or seemed like it was there.

This again comes down to the speed difference of a walking pedestrian versus a cyclist. Eye contact is something that a pedestrian moving at walking speed can do and they can stop in a single stride length if they realize the motorist hasn't seen them.

A cyclist moving at 10 mph is going to have a difficult time checking whether the motorist has made eye contact and by the time they realize the motorist isn't going to yield to them, they're going to travel 14.5 feet in the second they take to react and another 6 feet to come to a stop (assuming a constant 0.5 G deceleration). 20.5 feet is far longer than a stride length and will take the cyclist past the intersection, meaning a collision is practically guaranteed to occur.

> I don't even know where that video was filmed.

I believe it's on a section of Summit St in Columbus Ohio near Ohio State University. One person I know has been keeping track of crash statistics[1] since these facilities were installed. Since they were installed, the number of crashes has increased substantially.

> I take some amount of issue with the phrasing "this cyclist ended up in a crash because he failed to foresee the situation". He ended up in a crash because the driver ran in to him.

There are always going to be people who will fail to follow right of way rules in certain situations. Defensive driving involves anticipating when this may happen and take action to avoid a collision. This particular facility basically set this situation up so that it's likely to occur from time to time and cyclists need to be aware of this (especially given the crash statistics associated with this facility).

[1] https://drive.google.com/file/d/1UTSvrBxWqFpE1k6L212ppmSUo1N...