The trials and tribulations of planning my first DIY

[freeflow]

Last year I got introduced to Audax and have taken to it in a big way.  I love the relaxed atmosphere and the fact that it is non competitive.  I’ve now had my first experience at planning a DIY Audax and I thought it might be helpful to share some of my pain/learnings.

I’m based near Cambridge and the DIY is from my house to Bath.  The reason for such a journey is that me and the missus have been going to watch Badminton Horse Trials for the past 15 years or so and we stay in Bath for the three nights we are there.  Whilst I enjoy the event I do find 2 days of watching dressage a bit hard going and as I have now given up riding a horse there is less to do shopping-wise.  So it has been agreed that I can ride down to Bath on the first day of the event which is Thursday 7th May.

So having got the agreement I then thought – Why not do it as a DIY Audax?  A quick scan of google maps, cycle travel, mapmyride etc convinced me that a 300k route would give me the best bang for the buck.  So I plotted a route using mainly back roads ( I use GPS for navigation so it doesn’t matter how twisty things get) which gave me a nice 305 ish km ride to Bath.

WRONG

I then looked up how to organise a DIY Audax and following kind help from Paul (the SE DIY representative) I finally understood that I needed to set some control points and that this was best done using google maps in wailng mode (actually that should be walking mode but I think the spell checker version is very apt) .

I soon realised that I couldn’t manage my original planned route as I needed more than 25 control points to force google maps to use my route.
So I then set up what looked like the same route using waypoints only to find that it was woefully short at only 270km. This then meant I had to modify the route to ensure that the waypoints and walking mode gave me a route over 300k.

NOT GOOD

Having placed what I thought were acceptable waypoints I then joined the ‘dots’ using back roads and ran into two problems.  The first was that the ridden route came out at over 340k.  I’m not a wuss but 300k is pretty much my limit at the moment.  The second problem was that the last third of the ride was now through terrain where small roads were limited and not particularly oriented in the direction I wanted to go. Consequently the route had long stretches of A road (which I don’t find very appealing) or an unnecessarily wiggly route on back roads (I’ve come across this before when trying to plan a ride from A to B.  It would be a good geography lesson learning how roads are constrained by the territory they cover).

WAYPOINTS

Are a pain in google maps.  You think you have your route sorted.  Then you zoom in to compare your ridden route proposal with the google map waling route to make sure that you go through each of the waypoints.  What you then find is that the waypoint corresponding to the location you typed in is 200m into a field or a half mile trek up a road and then back again.  Adjusting these waypoints means that google changes the waypoints from meaningful names to meaningless road/junction combinations (which are still valid control points, just not so easy to find when just looking at a map).  It also means that your 301km google route suddenly become a 299km google route forcing you to tweak the waypoints once more.  So round and around you go.

You also have to check your ridden route quite carefully.  I use mapmyride and it will quite happily take you across fields and throughfarms even if there is no right of way.  My bike is set up so riding such territory isn’t an issue, but I’d prefer not to have to make unplanned deviations when I know I’m riding at my limit.
I finally have an acceptable google route which Paul has confirmed is OK, and a riding route that uses the same waypoints but is only 17km longer.

The google route is

http://goo.gl/maps/czpmM

The ridden route is

http://www.mapmyride.com/routes/fullscreen/608002340/

If all goes well I’ll be departing at 10pm on Wednesday 6th May.  Full moon is on the 4th May so if the skies are clear it should be a good night for riding.
If anyone wanders why the route meanders so much when I get to Bath its because I was a post grad student at the University and spent some time living in a one of the houses on Lansdowne Crescent (up in the servants quarters of course), and the Royal Crescent is because I want to.

I hope the above wasn’t too boring.  I’m very happy for anyone to pass on hints and tips about planning DIY Audax routes.

[Chris S]

I hope the above wasn’t too boring.  I’m very happy for anyone to pass on hints and tips about planning DIY Audax routes.

I think you summed up the process very well  .

Practice will give you an eye for a route, and the iterations you need to optimise a route around a set of controls will decrease. Probably .

There are some real experts out there - who seem to have no problem in devising a 200.1km ride which comes in barely over-distance at all when ridden (no names, obv!). Of course, we don't get to see the hours and hours of agonising over GM that might have gone into these endeavours.

Welcome to the Wonderful World of DIYs.

[Euan Uzami]

did post a reply but wasn't very diplomatic so going by the rule of if you haven't got anything positive to say don't say anything at all, so yhpm instead

The diplomatic version is that there comes a point when the amount of controls you have to insert to get it up to distance makes it too much of an effort to make it into a DIY of that distance.
You could either do it as a 200km diy or just not bother doing it as a DIY at all.
Have a think about effort to reward ratio.

[Huff n Puff]

I did my first DIY in September, have now done nine and only just perfected the route!

In addition to the issues you encountered, I was trying to devise a winter route that worked for AAA and therefore avoided roads that flooded, iced, got too much gravel, flints and mud after rainfall, did not have excessive potholes etc. yet started near to my home. Ride with GPS was quite happy to route me through fields and all sorts of non-cycleable terrain, which meant tweaks to the route after each test ride. Ride with GPS also shows less metres ascent than the Audax UK software, consequently I was struggling to get the required ascent when planning on Ride with GPS, until realising it under counted. Luckily, Steve Snook (top man) was really helpful and gave me the exact Audax UK readings after each ride so that I now have a feel for the differences.

I now have a good AAA winter ride in the Chilterns of about 120k, but I did plenty of 'Wailing' along the way!

Paul Stewart was/is excellent too.

[Jaded]

Sherston as a Control is taking you away from some nice back roads.

You can have a grid reference as a DIY Control, so could come out of Malmesbury on Foxley Road, follow the green dotted line on the mapmyride until Grittleton, turn right there, then cross the M4. There's a very nice tea room at http://www.fossefarmhouse.com/ might sort you out for the remaining 15 miles!
From there it is easy to get back onto your planned route.

[redfalo]

I've done four DIY's since last summer, all via Paul. I find it easier to plot the route manually in Basecamp, using free OSM maps, doublechecking distances with Google Maps on wailing mode. Only when my definite route is set, I replicate it in Google Maps, chose as little controls as possible and file it to Paul.

These are my different steps:

1) Play around with Google Maps wailing mode got get a rough idea about distances.

2) Make a decision about the start, the finish and the 'structural turningpoints'. If I were planning a 300k ride from Haverhill to Bath, I'd probably chose Stratford-upon-Avon as the fundamental point to turn south, which is 282 km according to Google. http://goo.gl/maps/7BUb6

3) I'd then plot a route manually in Basecamp, using free OSM maps. I'd draw a straight line from Haverhill to Stratford, and then from Stratford to Bath. This gives you a general sense of direction. I'd then draw a route that follows the straight line as closely as possible, using the smallest roads available and avoiding busy A roads altogether. I don't trust the automatic routing functions and want be in full control of the roads I'll be riding on. Drawing the route manually may be tedious at the beginning, but once you got some practice, it does not take much longer than using automatic routing.

4) When the ultimate route is fleshed out in Basecamp, I then replicate it in Google Maps, and add a few more controls, but only as little as necessary, and file it to Paul.

[freeflow]

come out of Malmesbury on Foxley Road, follow the green dotted line on the mapmyride until Grittleton, turn right there, then cross the M4.

Excellent tweak.  Changing the control point from Sherston to Grittleton changes google walking route to 303km and the ridden route to 316 km.  Thanks for the tip(s). Route now updated.
 
The only caveat is that the area around the M4 at Grittleton is where you need to start being canny about the roads used.  For example, I would'nt take a route through Acton Turville down Tormarten Road  as this is a prime route for traffic going between Badminton and the M4 at .  My current schedule has me at Grittleton anywhere 2 and 4 pm.

[freeflow]

I find it easier to plot the route manually in Basecamp, using free OSM maps, doublechecking distances with Google Maps on wailing mode. Only when my definite route is set, I replicate it in Google Maps, chose as little controls as possible and file it to Paul.

Sums up nicely that you need to define a working method that minimises the effort.  Mapmyride uses OSM maps.  The tip about grid references above is helpful.

[Ivan]

I've recently found myself reverting to using 1:250000 OS maps for planning DIYs - with a ruler it is very quick to establish potential direct routes between controls using minor roads only and bypassing any major towns. I then enter these into Google Maps for the total distance check.

[IanDG]

I don't have thes problems with routes on Lewis and Harris - need plenty of controls for the out and backs on my longer routes tho'.

[wilkyboy]

I live just outside Cambridge also, and I ride at least several DIYs per year to keep my RRTY going.  What I can confidently tell you is that almost all the routes with convenient commercial controls end up being 198 or 199km on Google wailing.  I often end up having to control in an inconsequential village near home just to force GM to not wail me on footpaths through fields near the A14/M11 junction (and I mean right underneath that junction).

Another annoyance is that RideWithGPS doesn't recognise the difference between a footpath (no right of way for ridden cycles) and a bridleway or byway, so I end up using it in driving mode to avoid unnecessary off-road sections, and forcing the route to use lanes by adding extra dots.

FWIW, I end up with roughly one dot (not a true GPS waypoint, but a mapping dot) per 1-2km on open roads, at least one per segment of road between junctions, much closer together in towns.  So a 200 would have around 150-200 dots on the map.  Some of those dots are necessary for fixing wrong-way around roundabouts and green triangles, as well as forcing the route onto smaller roads.  It's probably overkill, but the routes are then reliable on my Garmin Edge 800.

If you hate doing the route prep then it's probably a drag, but I do enjoy all that, otherwise I'd end up watching Master Chef instead (which I do as well as  ).

[Banjo]

I find DIYs I do almost allways end up with me doing a substantial amount further trhan the nominal distance in order not to have loads of controls.

Organizers will not be thrilled at a DIY with more than usuall controls.

I would suggest making yourself a nice 200 km DIY in the general direction you wish to go .Have a rest and a meal then ride the rest of the way at your own pace.

[Manotea]

The first problem is the original A-B route is really only about 275km, so it needed to be wigglified (technical term) to make it a 300km.

The second problem is that attempting to nail down every turn of the 'on the road' route will generate too many controls on wiggly routes.

Here is my effort, rendered with 10 'gazette' controls as a street legal 301km and remarkably faithful to the OPs original 'on the road' route.

I've always found prepping a route which meets the standard using the smallest number of controls part of the DIY 'challenge'. Sad but true. I guess it appeals to my OCD side.

For DIY Perm purposes, we really don't need the full on the road route, just the controls that satisfy minimum distance. When it comes to DIY controls,  'less is more'. 

[mmmmartin]

For my RRTY I used a few DIYs but then found out very quickly that it was a lot easier to enter a perm sorted out years ago (by "A Certain Well-known South-East Organiser"). Bung a cheque in the post, get the brevets, do the rides. Simples.
But I agree that if this option is not open to you a DIY is the answer. Sort out a decent 200k startng from home and you're laughing.

[jsabine]

'less is more'. 

Fewer is more, shirley?

Sheesh

</pedant>
</smartarse>

[Ian H]

When I was SW DIY man I stopped recommending Google walking mode as it had even then become too random with footpaths and all kinds of small but incremental shortcuts not available to a bike. That was at least two years ago.

[Jaded]

For my RRTY I found DIY by GPS to be invaluable.

I had at least three routes that were A to B for a reason, and there was no way any perm would have done that.

[Somnolent]

'less is more'. 

Fewer is more, shirley?

Sheesh

</pedant>
</smartarse>

 

GPXeditor is rather good for this IME.
You can use the Google routing algorithm on it .... whilst switching views between gmaps, OSM, OCM render of OSM, and OS mapping at a variety of scales.
And you can have two tracks (of different colours) on the screen at the same time - one the shortest walking distance per Google, and the other the way you actually want to ride.   This helps greatly in visualising the effect of moving controls around, even by quite short distances.
 

[vorsprung]

I wanted to ride my own 400 (the Avalon Sunrise) as a DIY GPS the year it wasn't running as an event.  I wanted the points as i needed a 400 to get SR

I had to use 21 controls to make the event route more or less fit where I was going and add in a trip around a motorway roundabout which went above a point only accessible by circuitous paths to "hack" up the stated distance to over 400km.  The real distance was over 400km of course

[mattc]

Eh? Surely the Controls on a "400km" Cal event HAVE to enforce the minimum distance to be at least 400km 

[must be some crossed wires here ... ]

[Bobby]

GPXeditor is rather good for this IME.
You can use the Google routing algorithm on it .... whilst switching views between gmaps, OSM, OCM render of OSM, and OS mapping at a variety of scales.
And you can have two tracks (of different colours) on the screen at the same time - one the shortest walking distance per Google, and the other the way you actually want to ride.   This helps greatly in visualising the effect of moving controls around, even by quite short distances.

Ooooh, that sounds cool <goes to play>

Thanks PpPete

[freeflow]

GPXeditor is rather good for this IME.

I've tried using gpxeditor on a number of occasions but I find it to be a rather opaque and unhelpful application.  In my limited experience of these things I'd probably only give 1/10 for its user interface.

The second problem is that attempting to nail down every turn of the 'on the road' route will generate too many controls on wiggly routes.

I'd agree with this.  With my first attempts to recreate my route using google maps I was too wedded to the route I'd already devised.  Once I'd accepted that I just needed 'A' route it was much easier. 

[Somnolent]

GPXeditor is rather good for this IME.

I've tried using gpxeditor on a number of occasions but I find it to be a rather opaque and unhelpful application.  In my limited experience of these things I'd probably only give 1/10 for its user interface.

It's a trade-off.  More user-friendly controls would take screen real-estate away from the map.
IMO its other good features make it worth the trouble to learn your way around an admittedly less-than-obvious-at-first glance UI
Coming (as I do ) from a pre-Windows 3.1 generation it didn't seem that hard.

"YMMV" as they say.

[Euan Uzami]

GPXeditor is rather good for this IME.

I've tried using gpxeditor on a number of occasions but I find it to be a rather opaque and unhelpful application.  In my limited experience of these things I'd probably only give 1/10 for its user interface.

The thing about the  UI is it's only there when you want it to be - like Pete says it is right click menu driven in order to maximize map space, so everything you need to do can be accessed by right clicking, and some menu functions are sensitive to WHAT you right click on or WHERE you right click. It is also open to suggestions for improvement unlike probably most other online route planners. It's also intentionally kept as simple as possible - if you need things fading in and out, animating up and down and generally whizzing and banging around and looking like facebook then it's probably not for you.

[PloddinPedro]

The biggest hassle with the composition of DIY-by-GPS  courses is the need to use "walking" mode because it's the closest GMaps gets to "shortest". It would be fine except that it naturally incorporates footpaths. That's why I always liked Autoroute when it was accepted - it's dumb enough to take " shortest" literally and yet sticks to roads (because it assumes you're motorised). It was a retrograde step when we abandoned it.