MSR Ride Statistics Interpreted

In the March Magazine I printed the accident statistics for 2009 and 2010 which showed there had been a spate of accidents including four on one ride. But what does that really mean? Is that any better or worse than other years? Is there a trend? And if there is, then what? Further investigation was required. The fundamental question we are trying to answer is, “How many thousand kilometres between crashes?” Which can be further broken down into, “How many kilometres between crashes versus crashes with injuries?” The MSR Ride Statistics spreadsheet is the product of that investigation.

How did I do it?

Having always had a great sense of history, both being part of it, and not wanting to lose what has gone before, I have been collecting stats continuously from my first ride - October 1982, 33 bikes. As the years have passed I have become a bit more sophisticated, the statistics eventually forming the basis of the Club Participant of the Year, a more scientific approach to the previous Club Person of the Year which was susceptible to a popularity vote. Another conversation. This discussion centres on the Club accident statistics.

The Club Participant of the Year provides the raw data of who went on what ride and how long the ride was, all contained in an Excel Spreadsheet. See the latest 2009-2010 3^rd Third elsewhere. I also capture the forecast temperature; what I will do with it I don’t know.

On the Itinerary Page of the Club web site we used to capture the incidents, everything from running over dogs to being carted off to hospital. That information forms the basis of the Club Captain’s Report printed in each magazine. Removed from the public facing Itinerary Page, this information is now captured in the Members area for 2010 onwards in the Incident Logs.

The third source of information is my hard copy Orange Book in which I write down all the attendees at Club functions whether it be a ride, Social Sip, birthday, or General Meeting. And it is useful to capture everything else that could slip my memory like money transactions.

Throw in a surprisingly vivid memory (and photos, Julie notes) of incidents, most of which I attended, and I now have five sources of raw data to compile fairly accurate Incident Logs. Clearly, the crash and injury severity data is a subset of the total incidents data.

The Maths

I have used data for the last five and a half years which is probably long enough to capture a trend, so back to 2005. I have arbitrarily used the calendar year rather than the Club Year (July 1 to June 30). It was easier.

The bottom left of the MSR Ride Statistics spreadsheet tabulates the raw kilometres the Club travels each month – the sum of the distance between the start of the ride and the end of the ride. This is the official Club ride length but for most people add another 80-100 km to reach the start point and get home again from the finish point. Think Berwick or Whittlesea start and Longwarry North finish. Incidents that happen to or from a Club ride are not considered to be part of the Club Ride and are not captured.

The top left of the spreadsheet takes the raw kilometres and multiplies them by the number of people on each ride, summing per month, then per year. We need to know how many people kilometres the Club travels each year. This also picks up pillion riders who are just as susceptible to injury in the event of a crash. Now the first corruption of the raw data begins.

The calculation assumes all people complete the full length of the ride. They don’t. This is not so critical on Sunday rides, but is more problematic on long weekend rides like the Melbourne Cup weekend, Tasmanian trips and the Xmas Camp. For the Melbourne Cup weekends, I have excluded non-riding partners from the calculations. For Tasmania, I have estimated the average number of kilometres, say for 2009, at 2500 km per rider despite people like Cliff and Ben doing 4,000km, as others did much less. That was over 9 days. But in 2007 we were only there 7 days ... At the Christmas Camp, say 2010, there were 39 people, but only 22 bikes. I used 22 bikes x 1,000km. Best guess. Unless I capture exactly how many kilometres people do on these events, the data won’t be that accurate. Nevertheless, I have been consistent over the years making an educated guess and if we are only looking for trends, then that is good enough. Another option is to exclude the inaccurate data associated with weekends away, but that’s where quite a few big crashes occur!

The middle right data on the spreadsheet is concerned with the number of crashes per year, and the severity of injury, using a scale of 1 to 4 where 1 is no injury to the rider – they walk away, through to 4 – death. Luckily there has only been one fatality in the five and a half year period.

Analysis of the Data

Now that the data is captured we can look for interesting “facts” and trends. For the past five and half year period I can state:

The club averaged 29,143 km a year if you sum up the kilometres per ride. The range was fairly tight – between 29,840 and 31,011, excluding 2006 at 24,555 which was a particularly wet year with many rides abandoned. It looks like 2010 may be down as well though the back end of the year is normally heavily weighted with Melbourne Cup and Christmas, so maybe not.

The Club (people x km) averages 324,241 people km per year, but in a fairly wide range, from as low as 238K in 2006 (wet) to 411K in 2009. The trend is definitely up, the long term drought providing brilliant riding weather from 2007 onwards. Given we do roughly the same number of raw kilometres a year, yet the Club kilometres are going up steeply, this can only mean that the number of riders per ride is going up. This is obliquely confirmed with the gradual increase in membership over the period.

Now we get to the pointy end. We average 14.3 crashes per year. 2010 is already up to 11, well ahead of the curve. Dividing the number of kilometres per year by the number of crashes will tell us how many kilometres between crashes. On average, 20,756 km. Not good. If we only include crashes where people are injured or worse, then the figure improves dramatically to 111,563 km between injuries. This is the bottom line, the answer. It is pretty much the same every year – 100,000 km between injuries, except for 2008 where, despite 13 crashes, no-one was seriously hurt. So 2008 boosts the average bringing it up to 111,563 km. And yes, 2010, at 40,953 km between injuries, is 70,000 km down on the average. Not flash. But it does appear to be an aberration.

Lies, Dam Lies, and Statistics

Statistics are open to interpretation. Paul Southwell has already noted that based on 137,000 km between injury crashes, the Club owes him at least 500,000 km of injury free riding!

What now?

Thirty-four of the 86 crashes involved riders either on their first, second or third ride, or who have been riding with the Club for less than three months. This is a significant problem. Forty percent of our crashes involve new riders, but they only make up probably less than 10% of our riders. A strategy needs to be put in place to minimise this risk. One option is to send new riders home after the first leg of each of the first three (?) rides. We are obligated to do something as part of our Risk Management Strategy, part of our Risk Management Policy. Ideas very welcome.

Conversely, 60% of our crashes involve members – about 8.6 per year on average. How does this compare to the state average? Is it acceptable? What can we do to minimise the risk, without removing the fun aspect of riding? Thoughts?

As an aside, the 2009 Incident Log shows 12 extra-curricular crashes (including track days, private rides, etc) as distinct from “only” 16 crashes on regular Club rides. Nine of the twelve non- Club ride crashes were members. Similarly, of the 16 crashes on Club rides, 9 involved members. So there are as many members crashing outside of the Club as there are on Club rides. Of course non-Club activity crash data is almost impossible to track, and hence not tabulated for the other years. But it does raise another series of questions along the lines of: Are the same people crashing inside and outside of the Club? Is this a fair question, given that they may just be the most active motorcyclists – do the most active motorcyclists crash the most? Or are certain individuals over represented in the crash data? Why? Too hard to answer without a lot more data. But worth pondering.

I’ll keep this database up to date and place in the Committee Only section. I am happy to discuss conclusions, assumptions, and data with anyone.

Hot Spot Crash sites

Where	2010	2009	2008	2007	2006	2005	Total
Tasmania		4		1		2	7
Noojee-Icy Creek-Hill End Road	1		3	2		1	7
Thomson Dam, Walhalla area	1	1	1	1	1		5
Jamieson/Torbreck River Road			1		1	3	5
Korumburra Warragul Road		3	1	1			5
Xmas Camp and surrounds			1			3	4

Miscellaneous Incidents

What	2010	2009	2008	2007	2006	2005	Total
Puncture	2	9	6	8		4	29
Ran out of fuel	2	4	4	1	2	1	14
Holed radiator		2	4	1		1	8
Police (speeding and fine)		1	6			1	8
Mechanical (clutch, engine, pads)		4	3	1			8
Animal (kangaroo, dog, snake, bird)		2		3			5

Ben Warden