Motostats 2008

[beginner]

Because only fatal crashes are reported in enough detail to do this kind of analysis.

You got me there. I followed your links and found summary data. Is there any place to read the individual accident reports?

 

[netjustin]

Say both riders loose traction in the same turn at the same speed for the same reason.

One thing I've learned in my puny experience is that it's never the same turn twice.

 

[beginner]

One thing I've learned in my puny experience is that it's never the same turn twice.

No two turns are exactly alike so there are an infinite number of turns but there aren't an infinite number of mistakes or skill and experience would be irrelevent.

 

[netjustin]

No two turns are exactly alike so there are an infinite number of turns but there aren't an infinite number of mistakes or skill and experience would be irrelevent.

I dunno. My take is that Dan's info is not to report on actual rider error or lapse in judgment at the point of crash. Instead it looks like trends that act to "set up" a particular type of recurring accident, irrespective of which turn or intersection or onramp.

I'm not sure why you'd be interested in specific accidents, but if it's for learning and accident prevention on your part, then I suggest using a bicycle to travel some of the roads you plan to ride. At speeds of 5-25 miles per hour, you can really see how disgusting the road surface gets and where, especially following inclement weather. :thumbup

 

[DataDan]

With motorcycles there can be two identical crashes where one person walks away and the other person dies of injuries. Say both riders loose traction in the same turn at the same speed for the same reason. One rider slides into the grass, picks up the bike, and rides home. The other rider makes the identical mistake 10' earlier and instead of stopping on the grass hits a tree. There's no mystery why the second rider died, he hit a tree. What I'd like to know is what led up to the loss of traction.

Luck often determines injury severity. We're in complete agreement on that.

As far as modern bikes being safer than old ones, the reverse is true. Which bike is going to kill more people, the 600cc super bike with the best of everying or a bike that's underpowered with marginal tires, poor suspension, and crappy brakes? The modern bike is going to be the killing machine for two reasons, because speed increases risk, and because there's a lot of temptation to operate the modern bike beyond the skills and experience of the rider.

Again, we agree. I just have to persuade you of that.

What you're saying is that it isn't grip or steering response or lean angle limit or brake controllability of the modern sportbike that makes it deadly, it's rider perception of the risk and his adjustment to it. I call that the paradox of risk perception: the more dangerous you think something will be, the less dangerous it will actually be. And conversely, the less dangerous you think something will be (because of the highly capable machine, for example), the more dangerous it will actually be. For more on that see the 1Rider thread Risky Business. One of the themes in this forum is sorting out the contribution to risk of rider decisions vs. fixed elements like the machine and environment.

I followed your links and found summary data. Is there any place to read the individual accident reports?

Not that I know of. However, among the fatal crash data in the FARS database is the sequence of events and one or more rider factors that contributed. For example, here's the sequence in a 2008 crash where a Gixxer 1000 rider at 90mph was killed when he ran wide in a turn:

1. Overturn/Rollover
2. Cross Median/Centerline
3. Ran Off Road - Left
4. Guardrail Face

So he went down on his side of the road in a right-hander (the FARS/Google Earth mashup linked in this thread would show exactly where), slid across the centerline, and hit the guardrail. It occurred in daylight, and no adverse pavement conditions were reported. The significant rider-related factor reported is "driving too fast for conditions or in excess of posted limit".

My take is that Dan's info is not to report on actual rider error or lapse in judgment at the point of crash. Instead it looks like trends that act to "set up" a particular type of recurring accident, irrespective of which turn or intersection or onramp.

And indeed, that kind of detail is seldom available. The exception is where there are witness accounts that appear in the media or a forum.

But I do collect enough detail to know, for example, which vehicle crossed the centerline in a head-on, or who violated right-of-way at an intersection.

I haven't yet posted my analysis of the circumstances of 2008 Bay Area fatal crashes, but I've done the research. For example, 36 of the 98 crashes occurred when a motorcycle ran wide in a curve and hit either a fixed object (like the example above) or an oncoming vehicle, and 18 occurred when a motorcycle collided with a crossing vehicle.

 

[beginner]

Luck often determines injury severity. We're in complete agreement on that.

That's why the thing that interests me is the sequence of events that led up to the bike falling or colliding in the first place. I'm talking about the technical things, the rider and motorcycle dynamics.

What you're saying is that it isn't grip or steering response or lean angle limit or brake controllability of the modern sportbike that makes it deadly, it's rider perception of the risk and his adjustment to it. I call that the paradox of risk perception: the more dangerous you think something will be, the less dangerous it will actually be. And conversely, the less dangerous you think something will be (because of the highly capable machine, for example), the more dangerous it will actually be. For more on that see the 1Rider thread Risky Business. One of the themes in this forum is sorting out the contribution to risk of rider decisions vs. fixed elements like the machine and environment.

Besides perception of risk I'd add the riders ability to perceive and manage lean and traction because that work is done by a different part of the brain that perceives things differently. I'll argue that the "perception" of risk needs to be a broader concept. Some of my falls happened because my motor balance system reacted intuitively and incorrectly or because it didn't perceive the lean or traction change fast enough to react appropriately. Most of those mistakes can't be put in words and even if they could the human balance system doesn't learn from words. I'm still discovering unobvious hazards related to deficient technique even after significant hundreds of hours of riding time. I'm sure there are more to come so I try to encounter those things in a safe place a low speed. (I won't use the P word.)

...among the fatal crash data in the FARS database is the sequence of events and one or more rider factors that contributed. For example, here's the sequence in a 2008 crash where a Gixxer 1000 rider at 90mph was killed when he ran wide in a turn:

1. Overturn/Rollover
2. Cross Median/Centerline
3. Ran Off Road - Left
4. Guardrail Face

So he went down on his side of the road in a right-hander (the FARS/Google Earth mashup linked in this thread would show exactly where), slid across the centerline, and hit the guardrail. It occurred in daylight, and no adverse pavement conditions were reported. The significant rider-related factor reported is "driving too fast for conditions or in excess of posted limit".

Unless the reporting officer was a rider with some experience he probably couldn't evaluate the technique errors that might contribute to an accident. Those would be instructive details when they are contributing factors.

There are two things the safety folks could do for motorcycle riders that I would appreciate. First, maintain a publically accessible database of the accident reports that excludes the fields with personal information, includes a Google Earth placemark, and can be filtered by vehicle type, severity, location, etc. In urban areas like yours you might see trends suggesting places to avoid or be extra careful. A second thing I'd appreciate is if GPS data was collected and sumarized on a bunch of police riders to see what they actually do with their bikes. That would include moving time, moving average speed, max speed, then how much cruising and how much PLP.

 

[CaptCrash]

I haven't yet posted my analysis of the circumstances of 2008 Bay Area fatal crashes, but I've done the research. For example, 36 of the 98 crashes occurred when a motorcycle ran wide in a curve and hit either a fixed object (like the example above) or an oncoming vehicle, and 18 occurred when a motorcycle collided with a crossing vehicle.

Thanks for the effort. This is really important to remember! We are hurting ourselves more than we are being hurt by others.

 

[beginner]

I dunno. My take is that Dan's info is not to report on actual rider error or lapse in judgment at the point of crash. Instead it looks like trends that act to "set up" a particular type of recurring accident, irrespective of which turn or intersection or onramp.

My goal is to improve my chances of avoiding crashes that are likely to cause a significant injury, meaning a doctor visit or worse. One way to do that is get better skills and experience in situations where dropping the bike isn't likely to cause a serious injury, PLP for example. Another way to do it is learn about mistakes other riders make.

I'm not sure why you'd be interested in specific accidents, but if it's for learning and accident prevention on your part

In aviation the investigator tries to report in the most complete detail possible why an accident happened. Those reports are available to all. If that quality of information was available about motorcycle crashes I'd put some serious time into reading it to see what could be learned.

People in the safety business will say, he was going too fast or he was drinking or he 'lost control', and that is supposed to stand as the analysis of the crash. May be that's enough for talking about car crashes but it leaves out a lot when bikes are involved. I suspect there are bike crashes caused by freezing at the controls in a ways that are particular to bikes. I also suspect there are turning crashs where the rider senses loss of rear wheel traction and instinctively leans forward to lower his center of gravity which only shifts body weight forward, increasing slip and making the situation worse. I learned about the second possibility in the parking lot doing figure 8s with one hand steering. Turning right is harder than turning left because I have to lean forward in the seat to push the bar far enough to establish the right hand turn. The weight shift forwward increases rear wheel slip compared to a left hand turn. It's those kinds of mistakes that interest me.

 

[netjustin]

My goal is to improve my chances of avoiding crashes that are likely to cause a significant injury, meaning a doctor visit or worse. One way to do that is get better skills and experience in situations where dropping the bike isn't likely to cause a serious injury, PLP for example. Another way to do it is learn about mistakes other riders make.In aviation the investigator tries to report in the most complete detail possible why an accident happened. Those reports are available to all. If that quality of information was available about motorcycle crashes I'd put some serious time into reading it to see what could be learned.

People in the safety business will say, he was going too fast or he was drinking or he 'lost control', and that is supposed to stand as the analysis of the crash. May be that's enough for talking about car crashes but it leaves out a lot when bikes are involved. I suspect there are bike crashes caused by freezing at the controls in a ways that are particular to bikes. I also suspect there are turning crashs where the rider senses loss of rear wheel traction and instinctively leans forward to lower his center of gravity which only shifts body weight forward, increasing slip and making the situation worse. I learned about the second possibility in the parking lot doing figure 8s with one hand steering. Turning right is harder than turning left because I have to lean forward in the seat to push the bar far enough to establish the right hand turn. The weight shift forwward increases rear wheel slip compared to a left hand turn. It's those kinds of mistakes that interest me.

I am not in the business for anyone's safety except my own. That said, I have a pretty good idea of what will get me in trouble and a modest idea of what won't. In your example, I don't think leaning forward on the bike will create a problem other than upsetting the suspension. The more your weight is biased towards the front of a moto, the less important rear traction is to maintaining control.

Also, turning right might be more difficult with one hand, but that doesn't necessarily mean it's always difficult. If you are riding an enduro, particularly with stock geometry and a 21" front tire, then yes it takes some effort to effect countersteering especially with the inside hand only, e.g. pushing only at the bar. But it's different for other bikes, particularly ones setup at the factory for sport riding.

Anyways sorry if I sound like I'm trying to harsh your points, they are thoughtful I'm glad there are people thinking about this stuff as it relates to preserving our lives. It really isn't as critical to people moving inside cars! :laughing

And good on you for not only putting together ideas and theories but actually getting out and practicing them first hand. I'm only able to do that every month or so. Wish I could do it more. :ride

 

[Marcoose]

I am not in the business for anyone's safety except my own.

There's a smile at the end of the statement, so I'll take it with a grain of salt, but, Justin, that's a short-sighted idea.

Many accidents happen with and because of other road users, ie, cars, motorbikes, bicycles, pedestrians, horses, dogs, etc. The safer other road users are, the safer you are.

All the best,

Marc.

p.s. Dan, sorry about hijacking your thread.

 

[netjustin]

Many accidents happen with and because of other road users, ie, cars, motorbikes, bicycles, pedestrians, horses, dogs, etc. The safer other road users are, the safer you are.

More than anything I meant that as a disclaimer, that nobody should use my thoughts on traction as anything other than my own experience. I.e., your mileage may vary. What you've said is still spot on though. Thanks!

 

[Ogg]

I really need to get back to some analysis work that I was doing to supplement Dan's contributions. One thing that really stood out from some of the early peeks at the data is the kind of roads that are more likely to have fatal motorcycle incidents.

A lot of the discussion I see is about taking curves because a lot of the fun, and perceived danger, is in the twisties. We tend to see this as the challenging riding where we are more likely to exceed limits of traction, ability, judgment, etc. However, (again, preliminary result) from what I can see, the most likely place to die on a bike in the Bay area is on a freeway.

 

[beginner]

However, (again, preliminary result) from what I can see, the most likely place to die on a bike in the Bay area is on a freeway.

Can you share the link that has details like the location of accidents?

 

[Ogg]

Can you share the link that has details like the location of accidents?

Not at the moment, but as soon as I have it ready I will. The FARS data includes accident location. When I mapped it out a significant number of the reported fatalities in the Bay area lined up with freeways and/or urban settings (there's a bit of slop in the location data and thus some ambiguity). The point is that I was at least a little bit surprised to see relatively few fatalities off in the twisty roads.

All of that said, I do not have data for non-fatal accidents so we're left to wonder whether the observed result is due to more miles being ridden on slab than twist, or perhaps that riders in the twisties are more likely to have protective gear, or that something about canyon carving is inherently less fatal -- but could easily be more likely to cripple. It's all speculation and thus an area where I'd be very careful about making claims (empiricism beats intuition every single time).

So, as soon as I can get back to work on the stuff (this week, really, I swear), I'll start to post up results. I have data going back quite a few years.

 

[Nemo Brinker]

Hey DD and Ogg--I'm taking some GIS classes--if y'all have wreck data in Excel format, I may be able to map it soon.

 

[Ogg]

Hey DD and Ogg--I'm taking some GIS classes--if y'all have wreck data in Excel format, I may be able to map it soon.

Might be nice if you have access to ArcView and TIGER line files to do some accident overlays. In the interim google maps or google earth seem to work okay because an interactive web interface is included. Haven't tried either of the google tools with large data sets though. Either way, happy to share what I come up with in a format you can play with.

 

[DataDan]

Nemo and Ogg--If you want an Excel file with lat&long for Bay Area fatal crashes, PM me with an email address. A map would be an informative addition to this thread.

 

[beginner]

Might be nice if you have access to ArcView and TIGER line files to do some accident overlays. In the interim google maps or google earth seem to work okay because an interactive web interface is included. Haven't tried either of the google tools with large data sets though. Either way, happy to share what I come up with in a format you can play with.

I work with google earth features a lot, overlays, placemarks etc. and share those with others. I also have comercial mapping software called Global Mapper which can convert all kinds of mapable data to Google Earth KMZ files. There may also be free utilities that convert various map data to kmz. I'm not interested in doing hours of laborious hand work to get map data into GE but I'd use Global Mapper for an automated conversion if that's possible.

I find I'm most interested in single vehicle motorcycle crashes. Either the bike loses traction or it collides with something. There is a finite list of motorcycle dynamics issues that cause one or the other. I'd like to understand those a lot better and know more about which ones are most often involved in crashes.

 

[Ogg]

In order to keep the data sets small I've been focusing on the SF Bay Area. FARS is a national data set though, so you can pull it down and poke through it for Michigan (or the whole country) if you want.

 

[beginner]

In order to keep the data sets small I've been focusing on the SF Bay Area. FARS is a national data set though, so you can pull it down and poke through it for Michigan (or the whole country) if you want.

What is the link where it can be pulled down from? Or how about some key words that will bring it up in a search.