A 1300 DIY under the new mandatory route system - practical comment

[jsabine]

Rabbit, that's a great write up, and you visited some nice parts of the country in a properly audacious manner 

I'll second that. A good read, and an interesting disquisition on the practicalities: IMO, exactly the sort of thing the proposal around mandatory routing was supposed to enable (and far more audacious than going off to France with 6,000 others)

But I think the sticky point for the pedants and procedural anoraks is, what does "sticking to the route" mean exactly ?

I think  Jo has already covered that in terms of what happens on the ground and how you deal with it.

If  you want a technical answer you could say that less than 2% of the track points can differ from the planned route by more than 200m. Which means you could go off the planned route for a max of 26km on a 1300km ride, or 12km on a 600km, 8km on a 400km, 6km on a 300km, 4km on a 200km. In reality most necessary diversions would be far shorter rather than one big diversion

But personally I'd go with Dave's assessment "Don't take the piss" . Lay the two tracks on a map and it's obvious to the human eye if someone has taken the piss.

Yep, pretty much. No fixed tolerance (nobody wants to be the organiser who tells the rider they've had a 2.01% deviation so the track's been red-flagged), but a healthy recognition that there will be variation. Off-route in a town because you want to use the cafe that's down a sidestreet, and you'll be fine. A sensible deviation to avoid the bit of road someone's inconveniently holding a (car) hill-climb up, fine, especially if you remember to let the organiser know when you submit your track. A loop round a mountain because you couldn't be bothered going up, not so acceptable.

In essence, organisers look for reasons to validate rather than for excuses not to: make it easy for them to find those reasons, and you should have few problems.

(In related news, rule-fiends, in the interests of simplification the board meeting earlier today agreed a minor textual variation to the mandatory route proposal - if it's approved at the AGM, it should be clearer in use.)

Mandatory routing is very interesting as it does open a digital divide and there is no real analogue equivalent.

[...]
I agree about the potential digital divide and I am aware that a complete move toward use of digital technology for everything would be unfair and isolating of those without devices.  I would really hope that AUK keeps the brevet card based system as well as allow use of GPS validation.  In fact, from a wholly personal and selfish perspective, I wouldn't care whether the mandatory GPS routing option was ever adapted for organised perms and events, as long as it is available for the DIY format.

Mandatory routing for for DIYs requires a GPS as it's the only practical way of validating them, but not for calendar events - a mandatory route would mean that, for instance, the Tour of the Hills didn't require 6 controls and 2 infos in its 115km, with the organiser at liberty to specify a secret control and hide in a hedge if they so chose.

I don't think it is particularly exclusive: sure, it means only GPS-owners can ride a DIY-by-GPS, but equally only riders who were free on the 25th July and able to travel  to Dingwall could ride the National 400. It's a proposal that's meant to improve choice, and make routes simpler and more flexible to plan, while no less audacious to ride - and I think rabbit's ride demonstrates just that.

[Ningishzidda]

Keep it simple.

Design a route between some towns or Garden Centres where hot food, coffee and cake are available.
Ascertain the shortest distance legally allowed on a bicycle. If it doesn’t add up to the required distance, re-design the DIY.

When you ride the route with your GPS, if the result is over the distance, you’ve nothing to worry about.

Knowing GPS units can be up to 3% in error either way, design the route +4% to absorb any negative error in the result.

With recording a GPS track, it may be the case that your ‘actual’ chosen route is +4% greater than the shortest legal distance, whereby you will be successful.

[wilkyboy]

Knowing GPS units can be up to 3% in error either way, design the route +4% to absorb any negative error in the result.

Really?  I didn't know that.

If you use a speed(+cadence) sensor to feed into the GPS — a simple magnet on the back wheel — your distance rolled should be pretty precise (so long as the wheel size is calibrated and the GPS knows what to do with it).  I did a comparison of Garmin Edge 800 and 1000 last week and the distance measured on each was within 100m over 145km (~0.07%) of each other, and just a few hundred metres greater than predicted by RWGPS (since the distance travelled includes off-route excursions, like the suggested garden centre ).  Both were set to auto-calibrate the wheel size, which I've checked previously and it matched the measured diameter spot on.

I guess relying purely on the GPS position to measure distance is less certain, especially if you ride in a lot of canyon-like areas like central London, or the Grand Canyon.  My GPS track from London Sightseer in parts of central London was all over the place and had me in very strange positions — due to the canyon effect of tall buildings close together — the error was in the tens-of-metres range at times!  It made navigating by GPS hard, and the routesheet for 100km ran to seven pages*!  Great fun    No distances on the routesheet, either 


* formatted the way I like it; my PBP routesheet was only eight pages formatted the same.

[Von Broad]

I completed a 1300 DIY this week drawn up on the new 'Mandatory Route' system.

Not following the ins and outs of AUK, I had no idea what the Mandatory Route system was until your ride report. I do now.
Consecutive 300's, staying in Travel Lodges......all this reminds of another pursuit going on at the moment.
What a ride. Chapeau!

[jsabine]

Knowing GPS units can be up to 3% in error either way, design the route +4% to absorb any negative error in the result.

Really?  I didn't know that.

I'm not sure the GPS manufacturers did either.

[Ningishzidda]

Knowing GPS units can be up to 3% in error either way, design the route +4% to absorb any negative error in the result.

Really?  I didn't know that.

If you use a speed(+cadence) sensor to feed into the GPS — a simple magnet on the back wheel — your distance rolled should be pretty precise (so long as the wheel size is calibrated and the GPS knows what to do with it).  I did a comparison of Garmin Edge 800 and 1000 last week and the distance measured on each was within 100m over 145km (~0.07%) of each other, and just a few hundred metres greater than predicted by RWGPS (since the distance travelled includes off-route excursions, like the suggested garden centre ).  Both were set to auto-calibrate the wheel size, which I've checked previously and it matched the measured diameter spot on.

I guess relying purely on the GPS position to measure distance is less certain, especially if you ride in a lot of canyon-like areas like central London, or the Grand Canyon.  My GPS track from London Sightseer in parts of central London was all over the place and had me in very strange positions — due to the canyon effect of tall buildings close together — the error was in the tens-of-metres range at times!  It made navigating by GPS hard, and the routesheet for 100km ran to seven pages*!  Great fun    No distances on the routesheet, either 


* formatted the way I like it; my PBP routesheet was only eight pages formatted the same.

This just proves if you throw enough money at a problem by buying more equipment, an acceptable result can be possible to achieve.

[wilkyboy]

This just proves if you throw enough money at a problem by buying more equipment, an acceptable result can be possible to achieve.

Or just that the promise of technology (and technologists) isn't always (hardly ever?) borne out by the experience of it.

[hillbilly]

Design a route between some towns or Garden Centres where hot food, coffee and cake are available.
Ascertain the shortest distance legally allowed on a bicycle. If it doesn’t add up to the required distance, re-design the DIY.

Erm.   One of the reasons why mandatory routing was introduced was that, pretty much overnight, the online software that helped members do this was changed and so it isn't clear whether such an approach is "robust" and whether any alternative was "better". 

Layer on top of this the reality that we (AUK) were overly dependent on a particular way of doing something, when in reality its continued immutable function is outside our control, and I think the flexibility to register rides as Mandatory is a jolly good thing. 

Even so, the option to do Advisory/non-Mandatory routes is still there for those who wish to follow that approach (I've just entered a non-DIY on an Advisory basis, simply listing controls, but I found it more cumbersome than it was a few years ago).

[tom_e]

If  you want a technical answer you could say that less than 2% of the track points can differ from the planned route by more than 200m. Which means you could go off the planned route for a max of 26km on a 1300km ride, or 12km on a 600km, 8km on a 400km, 6km on a 300km, 4km on a 200km. In reality most necessary diversions would be far shorter rather than one big diversion

I reckon that's the other way round from how you'd want a hard rule.  If they leave the route and then rejoin it at the same place then they don't need penalising again for the extra distance.  So what you want is not that 98% of the track taken is on the planned route, but that 98% of the planned route is covered by the track taken. 

Either way I think a vague human rule is much more appropriate here.  You don't want a hard rule that actually allows a deliberate short cut.

[wilkyboy]

Erm.   One of the reasons why mandatory routing was introduced was that, pretty much overnight, the online software that helped members do this was changed and so it isn't clear whether such an approach is "robust" and whether any alternative was "better". 

Layer on top of this the reality that we (AUK) were overly dependent on a particular way of doing something, when in reality its continued immutable function is outside our control, and I think the flexibility to register rides as Mandatory is a jolly good thing. 

Even so, the option to do Advisory/non-Mandatory routes is still there for those who wish to follow that approach (I've just entered a non-DIY on an Advisory basis, simply listing controls, but I found it more cumbersome than it was a few years ago).

I agree: we aren't ever going to have immutable control of any third-party technology provider, that's always a given.  Merely at any given moment in time the services offered by one or other provider are a relatively well-known quantity.  In part that's why I keep referencing Garmin Edge measured on-the-road distances with RideWithGPS expected distances — the calculated-distance aspect of RWGPS routes I find hard to fault.  For now ...

As for minimum-path routing, that does seem to have gone to pot in the "new" Google Maps, not helped by the 10-destination limit.  RWGPS and sites like it appear to still have access to the "old" Google Maps API and so in effect are using the old routing engine.  Setting RWGPS to walking mode gave identical results to "old" Google Maps — I verified this several times before we lost access to old maps directly.  It's harder to verify that a rider submitting a route via RWGPS has been careful with their routing, though, as there's no "recalculate route" option in RWGPS — you have to wiggle every other point to get the engine to recalculate, so there would inevitably be an element of trust between rider and DIY org (or org and AUK) as to whether a given RWGPS (or other mapping site) route was true minimum distance or not by the previous Google boots standard.

Personally I route my DIYs and perms with a significant amount of over-distance — anywhere from 5 to 30km — as I'd much rather spend an extra few minutes on the bike and know I rode the requisite distance than have my 12-hour ride come in too short to be validated.  And it's flat enough around here that hardly anybody should be troubling the time limit even with the extra distance.


PS. I've still to try a mandatory-route event ... not enough free weekends ...

[Somnolent]

Either way I think a vague human rule is much more appropriate here.  You don't want a hard rule that actually allows a deliberate short cut.

But a vague human rule needs a vague human to apply it ?   Do any of our current DIY organisers qualify ?

The end point is an automated system* that can validate most rides and only spit out the 'edge cases' for the vague humans to look at.

* since AUK is unlikely to be spending its fabled wealth on cutting edge AI you could be pretty sure it will be coded with some unvague rules.

[tom_e]

The end point is an automated system* that can validate most rides and only spit out the 'edge cases' for the vague humans to look at.

Is it?  The main purpose is a bike ride; the GPS track is only used for [relatively low volume] DIY validation, not calendar rides.  Looking at the bigger picture, I don't see you need the automation bar particularly high.

If it was me specifying the software, I'd have every bit of the planned route which wasn't covered by a recorded track point within [say] 200m highlighted for a human.

[Ningishzidda]

The ‘Mandatory distance’ of years ago was the distances between the little blue arrows on the AA Road Atlas.
The DIY designer strung them together until the required distance was exceeded.

No internet mapping packages involved.

The route wasn’t mandatory. If the rider chose to take a country lane route, that was their choice.

It was easy for the DIY Organiser to check the route on his own copy of the AA Road Atlas.

 

[Kim]

The ‘Mandatory distance’ of years ago was the distances between the little blue arrows on the AA Road Atlas.
The DIY designer strung them together until the required distance was exceeded.

No internet mapping packages involved.

Well yes, but the only difference (apart from the amount of effort involved^[1]) there is that AA Road Atlases only get published once a year, and the precision is much lower.


_{[1] Pretty much any shortest-path heuristic more accurate than "it looks like that one" becomes deeply tedious when implemented by hand.}

[Hummers]

Good stuff Rabbbit and another adventure in the bag 

H

[TOBY]

Well done on your ride! 

Now I’m going to let myself down by contributing to discussions on rules:
The issue I can see (if I understand things correctly) is with just the need to submit and follow a mandatory route (that isn’t pinned out by controls) I could probably ride a 200km “event” without ever being further than 10km from my house . . . is that right?

[NeilH]

The issue I can see (if I understand things correctly) is with just the need to submit and follow a mandatory route (that isn’t pinned out by controls) I could probably ride a 200km “event” without ever being further than 10km from my house . . . is that right?

No - it's not. One still needs to nominate controls "at intervals of approximately 50 - 80 km" along the route (but for mandatory routing those controls don't need to fix the minimum possible cycling distance). And other traditional routing rules would still apply, forbidding (for example) repeated loops of the same section of road.

http://forum.audax.uk/index.php?topic=751.0

Regards,

Neil

[TOBY]

Ah, OK thanks, I did check the Auk DIY pages but didn't think to search the forum for the rules.

[hillbilly]

It's not quite as straightforward as that.  But I'm not going to bore the world with my insight into the points of details.

Basically it boils down to the DIY organisers wouldn't accept such a proposed route.  And a hope that they won't accept them in the future.

[Phil W]

The British/ CTC thing of having few or no distances on route sheets really annoyed me when I moved to the UK.

Yeah my local CTC audaxes don't have distance on the route sheets . They are quite vague in places, and since it is laney, there's always a few to mop up towards closing time. If I ever became an irganiser and took over I'd soon sort that out, whether it's GPS or route sheet, at least do it well if you're going to do it at all.

[LittleWheelsandBig]

If you live deep in the heart of suburbia, you could roll around pretty much every nearby road in a grid pattern and not repeat any section of road in the same direction. I guess your route might look like a well-ploughed field. Your controls would be at 50-80km intervals along your proposed track. If you judged it properly, you could do 100km that finished at your house for lunch and then reverse the morning route.

[rabbit]

Interesting discussions folks 

In part that's why I keep referencing Garmin Edge measured on-the-road distances with RideWithGPS expected distances — the calculated-distance aspect of RWGPS routes I find hard to fault.  For now ...

Be aware - RWGPS has a bug in the system they have yet to find and fix.  Apparently it's 'super rare', but I have managed to activate it three times in three different parts of the UK.  I think it is related to drag editing a route but am not sure yet.  It screws up both the elevation and the distance so it says you have a route of say 505 kms, when in fact its only 481 kms.  I would recommend, if you use RWGPS for route production, that you download the route and upload to a different routing program to check this before submitting it as a mandatory DIY proposal.  I suspect the DIY organisers would flag up the issue themselves when they upload the route to check the suitability, but if they don't you could find your wonderfully planned hilly 500 km ride is actually underdistance (yes, this did nearly happen to me and now I use GPX Editor and the good old 'Bikehike' programs instead). 

[Somnolent]

If you live deep in the heart of suburbia, you could roll around pretty much every nearby road in a grid pattern and not repeat any section of road in the same direction. I guess your route might look like a well-plowed field. Your controls would be at 50-80km intervals along your proposed track. If you judged it properly, you could do 100km that finished at your house for lunch and then reverse the morning route.

You forgot Regulation 27, chapter 14, subsection 3, part d
which if memory serves correctly says something along the lines of:
Extracteth not the urine.
 

[LittleWheelsandBig]

A good friend says that he wants to do a DIY perm where all of the controls are railway stations and all proofs of passage are train tickets but he thinks that the grins aren't worth being bankrupted.

All good clean fun!

[jsabine]

A train ticket to the station that happens to be the next control, of course?

I've wondered about submitting a petrol receipt as PoP, but I'm not quite generous enough to fund filling up a stranger's car.