28 November 2012

The 15th Percentile - Survival of the Fittest?

Doisneau Traffic
Robert Doisneau - running pedestrians in Paris

We recently covered the disturbing and archaeic 85th percentile method and how it is applied for (and by) vehicles. If you thought THAT was fun, you might also enjoy The 15th Percentile. It is frequently used to determine the time between the WALK and DON'T WALK crossing signals -  in other words, how much time the engineers computer models allow for human beings to cross streets. It's not as rooted as a standard as The 85th Percentile, but it is still widespread.
In a nutshell, we should be paying more attention to pedestrian crossings, when you consider statistics that say that "40% of accidents involving pedestrians occur at these intersections". In Europe, one in four pedestrians die on a crossing. What seems to be the problem?

In the U.S., the Manual on Uniform Traffic Control Devices (MUTCD) establishes a “normal” pedestrian speed at 1.2 metres per second (m/s), so traffic signal times are set according to that speed. It is also an international standard. One of the first things that popped into my head was: “what about the most vulnerable groups such as children or the elderly?”.

After analysing several published studies, children didn't appear to be the most problematic group – at least when it comes to their crossing times. Instead it is the elderly that are a cause for alarm.

The elderly have different 15th Percentile speeds than younger pedestrians, but it is still widely recognized that the the crossing time for older pedestrians is insufficient. In other words, 1.2 m/s simply isn't enough for older people to cross a street. Which, of course, means that it is not a very good standard, especially considering that our population is growing older and older. It's a standard that can lead to increaed death and serious injury.

Once again, we've made the mistake of allowing ourselves to be controlled by machines, like numbers. We know all too well that this has been proven to be insufficient in protecting pedestrians and cyclists, among others.

The fact that 1.2 m/s is not a speed for everyone has proved to be common knowledge. For instance, the Institute for Transportation Engineers is aware of the fact that "(...) the majority of traffic signals [in the U.S.] are timed so that up to half of pedestrians beginning to cross at the start of the clearance interval potentially will be in the street when conflicting traffic is free to proceed."

Even the MUTCD raised awareness about this by saying that a speed lower than 1.2 m/s should be considered when pedestrians with slower speeds routinely use crossings. So, if establishing a traffic signal time limit by the 15th percentile will give slower speeds than the average ones, the theory should be right, no? Then why are so many people getting injured or killed in the place where they should be protected the most - when crossing a road?

The simple answer: The 15th Percentile is still not enough to solve the problem for slower pedestrians.

After watching a number of videos of intersections – pedestrian crossings and its failures – beyond the point of mental sanity, there are a few problems that can be addressed (among many others, I'm sure) and that seem to have failed:
  1. Confusing signals for pedestrians. In the U.S., 50% of pedestrians, don't understand that the red “Flashing Don't Walk” sign means that you may continue crossing. Of course they (including me) don't understand: it's because it is confusing. Farther down you'll find particular cases around the world.
  2. The stop line for cars is too close to the pedestrian crossings. This one goes without saying.
  3. Drivers running red lights. It's obvious but I don't understand how this still happens these days.
  4. Not every sign respects the average speed of the pedestrian. Let's admit for one second that the average speed is 1.2 m/s, regardless of the group of pedestrians. Not every pedestrian crossing signal respects that speed. How many of us had to almost run in some crossings just to arrive safely on the other side? There's also probably a lack of inspection going on in some cases.
  5. 0.9 m/s should be the speed set for the average pedestrian – according to previously cited studies. This would allow the accommodation of all types of pedestrian groups. Several studies (linked here, here and here) have also pointed out that this is the speed for older pedestrians.
  6. The time gap between the red signal for pedestrians and green one for cars should be much longer, in some cases.

There are plenty of examples where the pedestrian has to adapt to the road rules, and not the other way around. (Heard that before, haven't we?). Here are some examples:
  • In the U.S., a crossing with some instructions for pedestrians. Without sounding paranoid, are we moving towards a mandatory walking licence?

  • In Berkeley, California, an experimental program between 2001 and 2004 was set up to test the efficiency of pedestrian flags. We blogged about it before. The methodology was: before crossing, the pedestrian would pickup a flag and waved it while crossing; then he would leave it in situ after crossing. Here’s an image:
Whilst using pedestrians for any experimental program is quite disturbing, the program was rather unsuccessful. Here’s what went wrong:
Nevertheless, other cities in the US have pedestrian flags in place. Seriously.

Sao Paolo Protecting Pedestrians

This is a very noble initiative but it still states clearly that the car rules the streets. Reducing car speeds would make this unnecessary.
  • On the other hand, let’s not forget the “push button” examples throughout the world:
Ottawa Application FormSao Paulo Streets 052Ljubljana Bicycle Life_5Halifax Ignoring the BullDC Overcomplication

The last image is very peculiar. Not only the streets don’t belong to pedestrians but now we have to submit to the power of the car (“THOU SHALL NOT PASS”?). Generally, these pedestrian buttons (or even for cyclists like they have in the Netherlands), require human beings to push a button and wait for the mercy of an engineered computer programme to grant them permission to enter the roadway.

When do you we ever require motorists to stop, roll down the window and press a button before crossing an intersection?

This recent news item from Stoke, UK shows how ridiculous some city councils can be. A spokesman actually says, "it was not that the crossing was not working properly, it was just that people were not using it correctly." So the council will hire people to stand in the crossing and "educate" pedestrians about how to use the badly designed crossing.
  • Traffic signals for pedestrians:
(source: here)
(source: here)

These two signs are the scariest ones: the first one shows a running pedestrian – I'm tired just looking at it.On the second one you have a countdown clock stressing you. The bad thing is that they assume your speed will be (at the very least) 1.2 m/s.

Nevertheless, these examples show some kind of effort, but they are measures to avoid addressing the bigger problem: cars. The cars remain untouchable.

Let’s face it, if safety is really taken seriously, then a lot of good things may happen during a crossing:

(source: © Apple Record)

But we still have an elephant in the room - a bull in the china shop - and the above measures are directed to avoid him while he's walking there. In this case, the elephant is the car – and the (almost) subsequent excess of speed by the driver. Reducing speeds in streets would help tackle pedestrian safety in a greater way.

Also, citizens’ initiatives are the most powerful and effective work-force, in my opinion. They should have a bigger role, considering they are the biggest stakeholders when in comes to traffic safety.

At the end of the day, it's funny that there are engineers entirely responsible for the free flow of vehicular traffic movement, but we don't have the equivalent for pedestrians. Sure, engineers who should be responsible for the free flow of pedestrians. Including their safety, of course.

All in all, The 15th Percentile method doesn't seem that bad at first sight. However, it still fails to protect everyone due to the fact that the elderly are slower walkers. We're still making the same mistake as ever: thinking of people as numbers. They (we) aren't.
Pedestrian Zone
But today's pedestrian crossings show that you have to be fit, be 100% aware, be literate. Basically, you have to be a "perfect robot"  and adapt. Adapt to traffic, adapt to cars, adapt to signals and adapt to crossings. It's Darwin's natural selection: the survival of the fittest.

PS: If you know any absurd activity on crossings for pedestrians, please share with us.


Tallycyclist said...

Sadly, pedestrians are often at the bottom of the totem pole, even though almost everyone is one at some point of the day. Well, maybe the US is an exception since so many people spend hardly any time walking outside.

Oftentimes pedestrians will unnecessarily yield to me, probably either because they don't trust that I won't hit them with my bicycle, or feel that I somehow have priority even when that's not the case. This further illustrates the power of speed, weight and/or size. Though cyclists are much closer to pedestrians than motor vehicles in many aspects, there are still some differences. Obviously the inconvenience, danger or discomfort posed by bicycles is on a totally different level than from cars, but being on one does often gives you extra power over peds.

Melissa Bruntlett said...

Here in Vancouver, local police and council ran a safety campaign earlier this year called "People Are Fragile". The focus was predominantly on things pedestrians and cyclists could do to avoid being hit. Among the absurd recommendations for pedestrians at http://www.practiceroadsafety.ca/pedestrians: Make yourself visible by wearing bright or reflective clothing. Consider carrying a flashlight or flashing red light at night.

John in NH said...

So what are some solutions? Count down timers are great, and I would like to see them count down until the signal comes on too, concurrent timing or automatic timing is also critical, no pushing button crap (though this wont go over well in areas where there maybe 1 or so folks trying to cross in an hour)

I personally like the running signal (they have these in Spain) as it is much clearer than the flashing hand, coupled with a good timer I think is a good combination. I agree moving the stop line back is critical can can benefit double if there are trucks and buses that have to turn on the road. At grade crossings on side streets (ped crossing on a speed table) are also good since it encourages pedestrian priority (cross light or no cross light). What other solutions would you suggest? NYC is implementing elder zones and slowing down the signals, I think this is a good place to start (near nursing/elder facilities or community health/resource centers) what else?

Ed said...

Let us not forget pedestrian bridges. When a street is too dangerous to walk across, hire an architect and spend $27 million to force people on foot to climb twenty feet into the air to cross the street so they don't get in the way of the precious cars: http://ny.curbed.com/archives/2012/11/27/west_thames_pedestrian_bridge_inching_toward_construction.php

Ed said...

Here's an inviting way to cross a street: http://www.flickr.com/photos/jag9889/2655870366/

Alex Rybakoff said...

I am writing about Helsinki. Situation is not that bad like in many other places, but not a good one as well.

One irritating example is when you cross a huge street or square and you have 3-4 different lights. E.g. second and third are green, but first is not. Then first is green, but second not. Then you pass second and third, but fourth is red. Finally fourth is also green. Awful.

Another example is a medium-size street with 2 lights for pedestrians, and second (far from you) is already green, but first (close to you) is still red. I am from another city, so I am always thinking that lights on normal streets (not huge squares) should be green at the same time. Once I almost stepped under the car :(

Third problem are traffic light's buttons and their small lights, indicating that button was pressed. They are not working in quite many places, so people think that they are not working and start to go/cycle on red.

Good point is that we do not have this bizzare bridges/underpasses like Ed mentioned, all of them (except one-two from early 70s) are quite usable for pedestrians and cyclists.

Another good point is that you have enough time to cross the street, without running. In nearby St. Peterburg you do have to run almost everywhere. And not far, in Riga, there are no pedestrian signs/lights at all in the places where you expect, so you have to go 0,5 km around. Disgusting!

Miles Bader said...

Sometimes pedestrian bridges and/or underpasses work very well, sometimes they don't. The picture you showed seems to be pretty horribly designed, though.

A few things that come to mind to make them better:

(1) They should be an addition to surface-level crossings in most cases, not a replacement; different people have different needs and preferences. Providing a pedestrian bridge should be an optimization for pedestrians, not for cars.

(2) In many cases it's possible to extend the elevated portion to reach a common destination, rather than simply going up one side of the street and down on the opposite side. E.g. if there's a shopping mall on the other side, extend the bridge into it; if there's an adjacent road to be crossed, make the bridge a square with access stairways.

(3) Make the bridges and especially the stairways wide enough to handle expected traffic, and then some...

It's pretty easy to tell the good ones: they're crowded, even when the surface-level crossing is well-designed.

Miles Bader said...

Oh, and (4) if possible, make them bicycle-friendly (bridge wide enough to allow easy passing, a ramp in the middle of the wide staircases). [Bicycle-friendly in this case, of course also means luggage friendly...]

Miles Bader said...

btw, I've always wondered:

What happens if you duck-tape down one of those "press to walk" buttons...?

[should be very easy with the big mushroom-style ones!]

Chris Walker said...

In the UK we have a flashing amber phase on pelican crossings, that allows drivers to proceed if and only if the crossing has cleared. Longer flashing amber phases would allow slower pedestrians to cross without forcing drivers to wait unduly.

On puffin crossings, which are replacing pelican crossings, sensors monitor pedestrians' progress on the crossing and only turn the lights green once no more pedestrians are detected on the crossing.

Do other countries not have these features on their crossings?

Chris Walker said...

Chris's British Road Directory (not me, a different Chris) has a good write-up on the various systems that the UK experimented with before standardising on the current signals.

It's interesting to read how averse the authorities were to any form of 'do not cross' signal at the start due to the fact that, aside from one experiment, there is no concept of jaywalking in the UK, and a 'do not cross' signal would have been given "without due authority".

Eventually they came to their senses, and treated the 'red man' as a recommendation only.

Rui said...

Pedestrians = Parasites!

If you don't think so, just learn it in this documentary:


Let's irradicate these peskies!

Dmitri F said...

When I was in Istanbul for a week, I found the traffic situation there horrible. Pedestrian crossings were few and far between, on large, 4 lane city roads, you could find a crossing every 4-5 blocks at best.
And when it turned green, after about 90 seconds (they had timers), you had exactly 24 seconds to cross.
As we were walking around town, I often found it difficult to simply cross the street. I witnessed a crowd 5-10 people, a few older ladies among them, waiting to cross a tiny back street without a crossing, and it was insane that not a single driver stopped to let them pass. We waited for quite a while before I decided to simply go.

In some cases, the sidewalks were so narrow that a stroller barely fit, while the road was huge and lightly trafficked, with trams passing so close (and fast) to the sidewalk that a misstep could land you under one.

As we later drove south through turkey, we drove through many many towns that were cut in half by the motorway, in many places, there was a fence blocking pedestrians from running across, forcing them to use one of maybe, two overpasses in the entire town.

In smaller towns, there was no fence blocking the road, but with cars driving by at 140-200km/h crossing seemed like an insane proposition. In the rare cases that there was a crossing, it lacked a light, and turkish drivers simply do not stop.

I wondered what would happen if one were to live on one side of such a town. Would you even be able to hang out with people on the other side of the road?
I imagine going to see your friend living not 30 meters away could pose quite a challenge.

koula said...

In Florida (or Flori-DUH, if you will) most of the pedestrian bridges and tunnels end up becoming dangerous because they are built out of sight and with very little lighting, and crime settles in quickly. While these tunnels or bridges are in theory meant to help people cross large spans of highway safely by foot, I notice that most people end up avoiding them, for fear of getting mugged or raped. Very poor design.

I've found that the city or state transportation folks are somewhat receptive to changing walk signal times, but are far more likely to react to requests from the public. A better solution would be to reassess the MUTCD's guidelines and change the book to reflect a longer crossing time. Last I looked, books are not carved in stone anymore :)

Pedro Madruga said...

@Tallycyclist, definitely agree on that. The pedestrians "grew up" thinking they had no power. In fact they still have a lack of that.

@Melissa, that is a bizarre initiative. Thanks for sharing.

@John, There are plenty of solutions, a few of them mentioned in the text. I think that the countdown timers and the running signals are rather stressful. At least for me.

@Ed, great example of waste of money. Will definitely considering that next time. Thank you for sharing.

@Alex, desynchronized lights are definitely one of the most dangerous things for pedestrians. Sometimes we have to stand on those islands in the middle of busy and speedy roads.

@Miles, I just don't agree on nr.2. With that option you will be somewhat "forcing" people to actually go to the shopping centre. I think that choice should be up to the people, therefore -- if a bridge exists -- it should be just to cross a road.

@Chris, I read about those example while researching for this post. Those are great examples and I've never been on a crossing with that. Can you tell me if that happens nationwide in the UK? Thanks.

@Rui, I've been watching films from that website. Lot of them are mobility related.

@Dimitri, those are some scary examples that are easily related to where I used to live before. Answering your question: I think it could be possible but it would mean a huge life risk...

@Koula, I agree. If the MUTCD considered 0.9 m/s instead of 1.2 m/s that would probably save more lives.

Miles Bader said...

There should of course also be a street exit—but if 90% of the users immediately enter the shopping center/station/whatever, then a direct entrance from the ped bridge (or tunnel) is a big benefit, it really lowers the perceived effort.

[and in the case of a business, the business can also help pay for it.]

Dmitri F said...

Another fun example from Turkey, this time from a beach resort town of Alanya (and surrounding towns).

A motorway runs all along the coast, going for several hundred kilometers between several towns, all tourist resorts mainly.

The kicker? The motorway separates the town from the beach (just the beach, nothing more!).
I'm not kidding.
You have to use an underpass (often starting in a "beachside" hotel, as well as the street re:@Miles) to get to the beach. Alternatively, you can use rare, unlit, albeit marked crossings.

Note, that this is a cross country motorway with 2 lanes each way, and with drivers mostly going over 100km/h, and those are the sane ones. At least there is an island in the middle, so you can rest between your sprints.

It's a miracle how these towns ever sustain tourism. That damn highway and the constant stress from the horrible pedestrian infrastructure and insane drivers, certainly mean I'm never going back there.

Of course, most tourists there were Russians, and I think the situation in Russia is no better, so maybe they are just used to it. Coming from Stockholm, I might have been spoilen.

Chris Walker said...


You'll find them all over the UK. There's a bit of variation in the design of the Puffin crossing equipment, but nothing major.

The detail that I forgot from my earlier comment is that pelican crossings are only used as standalone pedestrian crossings. They don't work as part of a signalled junction, where the pedestrian phase often just runs on a timer.

A puffin crossing can be incorporated into a signalled junction, and a lot of them have been upgraded.

Pedro Madruga said...

@Chris, do you have any statistics on people injured or killed while crossing, in the UK? They seem to be seemingly safe, no?

Chris Walker said...


I don't have anything with absolute numbers. I've managed to find a document that compares the safety of Puffin crossings with older-style crossings, and finds a reduction in accidents.

Unfortunately, it doesn't distinguish between Pelican and signalled junction crossings (which use different light sequences), preferring to refer to them collectively as 'farside' (referring to the location of the pedestrian signals).