Author Topic: [LEL17] Managing the spikes at controls  (Read 33362 times)

Re: Managing the spikes at controls
« Reply #175 on: 31 August, 2017, 07:32:31 pm »
"in PDF format" !?  - apparently there is something called HTML these days. Online and linkable from the screen it is explaining (even embeddable in the screen).

Kim

  • Timelord
    • Fediverse
Re: Managing the spikes at controls
« Reply #176 on: 31 August, 2017, 08:15:28 pm »
"in PDF format" !?  - apparently there is something called HTML these days. Online and linkable from the screen it is explaining (even embeddable in the screen).

PDF on the other hand can be output sensibly to dead-tree format, for ease of access without tying up the screen in question.

There are advantages to both.

alfapete

  • Oh dear
Re: Managing the spikes at controls
« Reply #177 on: 31 August, 2017, 08:16:24 pm »
"How many had gone through" wasn't at all obvious to the control team I was part of (several of whom possess proper big brains, unlike mine!).

Not normally one to venture in to heated debates, but I'm surprised that the relevant info wasn't easily discovered by anyone scanning cards.

The failing may have been that the training given by the controller (who MAY have had time to read Phil W's guide, amongst a hundred other tasks in the last few weeks) was not disseminated to the volunteers sufficiently. After all, however simple it may be to ride a bike, if you've never seen one before and have no advice on how it works you might just try using it as a mangle
alfapete - that's the Pete that drives the Alfa

Kim

  • Timelord
    • Fediverse
Re: Managing the spikes at controls
« Reply #178 on: 31 August, 2017, 08:35:46 pm »
There's a difference between ease of discovery and a thing actually getting discovered, thobut.

I'm highly computer literate and confident with digging my way around new software - something which you certainly can't assume of everyone who might volunteering - but that doesn't mean I'll actually explore a system sufficiently to discover a given feature if I'm either  a) too busy doing something urgent  or  b) afraid of destroying some not-as-yet-understood state[1] that would prevent me from continuing with (a) without documentation, assistance or sufficient time to work things out from first principles.  Particularly if I'm not aware that an immediately useful feature is there to be discovered.


[1] Eg. logging something out that I don't know the credentials for.

CrinklyLion

  • The one with devious, cake-pushing ways....
Re: Managing the spikes at controls
« Reply #179 on: 31 August, 2017, 08:45:22 pm »
Yeah - I didn't go rummaging until a More Knowledgeable Person showed me stuff, because the last thing I wanted to do was to muck up the scanning in process.

It was remarkably straightforwardly once I knew where to look. I completely grok that other people might not have looked or been shown.

Planet X Paul

  • The Green Machine
Re: Managing the spikes at controls
« Reply #180 on: 31 August, 2017, 11:07:55 pm »

We had a 71% scan out rate, so very positive considering it was encouraged but not mandated. I think its still right to treat forecasts as a guide. You'd have a better idea what the weather was doing to push rider arrivals either backwards or forwards. It is why we didnt try and predict rider arrivals more accurately than to the nearest hour. Any more granular is a promise of false accuracy.


Would separate scan in and scan out desks be better?  I was going to scan out at Louth (northbound) but the queue of riders scanning in was out the door, so I didn't bother.

Re: Managing the spikes at controls
« Reply #181 on: 01 September, 2017, 12:04:11 am »
"in PDF format" !?  - apparently there is something called HTML these days. Online and linkable from the screen it is explaining (even embeddable in the screen).

PDF on the other hand can be output sensibly to dead-tree format, for ease of access without tying up the screen in question.

There are advantages to both.

PDF can also be online and linkable, both within itself and in both directions with HTML based stuff.

I've created substantial technical manuals (for non-cycling related matters) in both, and the downsides of HTML are such that I'm unlikly to choose that option should I be faced with the choice in future.
 

Re: Managing the spikes at controls
« Reply #182 on: 01 September, 2017, 10:17:29 am »
Would separate scan in and scan out desks be better?  I was going to scan out at Louth (northbound) but the queue of riders scanning in was out the door, so I didn't bother.

Or even a "during the ride" section of the rider area in the website, allowing you to scan yourself out from your smartphone.

Re: Managing the spikes at controls
« Reply #183 on: 01 September, 2017, 12:15:14 pm »
'Managing Spikes', has a long history. In 2001 analysis was done on the results, and a graph appeared at the bottom of the Arrivee roundup of the results. http://www.aukweb.net/_resources/files/results/lel/lel2001.pdf

That informed the provision of controls in 2005, especially the twinning of Langdon Beck and Alston. Langdon Beck wasn't far enough from the previous control, and it ended up being very quiet there on the first day, especially as there was a tailwind. It had been proposed that LB would open later, for the Lea Valley starters to take advantage of it, but Heather resisted that, as she wanted to prepare the control. It picked up later on and became one the best controls, especially as it had a bar.



The combination of the Thorne start and the Lea Valley start meant that there was a chance of a mega-spike, with two sets of riders along the course, and back. The controls were small back then, and the entry fee wouldn't stretch to much in the way of luxury.

A lot of lessons were learned from 2005, and Simon Doughty was very much the man to make sense of those lessons. Unfortunately Simon had a life changing interaction with a drunk driver on the way to work at Manchester velodrome. A nurse called Mel, who had been involved with organising CTC events visited him a lot, and was the source of updates on his condition here. Mel had been the controller at Thorne in 2005.

With Simon out of action, there was an election at an AGM to appoint an LEL organiser, the candidates were Mel, and Pam, an organiser in the Hampshire area, who was a travel agent, and booked the AGM venues. Mel proposed carrying on with the plans for an AUK flagship event, Pam proposed a more bare-bones approach.

Mel won the day, and commented to me that the AGM had little conception of what handing over the event to a strong-willed Lancashire redhead might mean. This clearly would be something to watch.

I had a front-row seat, as Heather was down to control at Alston. The plan was to use the school at Alston. That booking possibility foundered at short notice, apparently on insurance grounds. We'd taken to skiing at the Yad Moss slope, and Heather hired her equipment from the Alston Outdoor Centre, so when it was suggested as a control, as the Youth Hostels were booked, we already had an in.

The Outdoor Centre was clearly inadequate, so I went up with lots of prepared food, largely tins, and erected our tent in the grounds as a provisions store, to make space in the kitchen. The start was now Lea Valley only, so there was a measure of predictability. If anything slowed riders in the Alston area, the control would resemble an improvised dressing station at the Battle of the Somme, and that's what happened when the rain and the wind came.

There had been lots of input from Nigel Hall on predicting flow rates for 2009, and discussion on here. It was early days for YACF, and there were loads of theories. At some point someone would have to set the actual policy in stone, which Mel did at a meeting in York. That's the last we saw of Nigel, who'd been down to run a control. That was the last meeting prior to the event, so we wouldn't meet the new controller for Thorne.

2009 was the first year of a single start from Lea Valley, and about 600 were signed up. That was twice as many as had previously started from any single point. If there was a tailwind, the bulk of the field would want to sleep the night at Thorne, at 300km, which was clearly not possible. I was sanguine about this prospect, as Heather is from Thorne, and her mum lives there. I would have a bed and a bath, and something to eat at Jean's house. So I got to see Thorne twice on each passing, before and after sleep.

On the way back, Thorne had a voucher system to ensure that riders only went round for food once. The other controls were ad-lib, especially Alston, where we'd catered for 600, rather than the 500 who'd turned up. We still have tinned peaches in our larder from 2009.

I got the impression that the controller at Thorne in 2009 was keen not to see anyone else parachuted into a situation with a large capacity for apparent failure. The other controllers had seen the situation developing over the years, and had experience of controls such as Loudeac on PBP.

There are two camps on the subject of spikes. Those who feel that the problem is amenable to a technical solution, and those who quite like finding themselves on the set of a remake of M*A*S*H.

I'm reading a book called 'The House of God', about interns in 1970s US hospitals. I'm starting to suspect that medics and the military like the chaos and sleep deprivation of LEL, because it reminds them what it was like to be young. I can see that for someone who thinks they've just booked a holiday, it might look different.

Phil W

Re: Managing the spikes at controls
« Reply #184 on: 01 September, 2017, 01:05:09 pm »
"in PDF format" !?  - apparently there is something called HTML these days. Online and linkable from the screen it is explaining (even embeddable in the screen).

You do know that both PDF and HTML were first published  in non draft format in June 1993 over 24 years ago? By "these days" I guess you mean for the last quarter of a century. Your comment reminds me of this.

http://youtu.be/OXDK3x5lAYI

We could of course subject some volunteers to the "gramophone test" to find out where on the spectrum they are.

Phil W

Re: Managing the spikes at controls
« Reply #185 on: 01 September, 2017, 02:05:50 pm »
ESL your latter camp referring to those who do no planning or contingency of course.

A bit like saying to your 1500 riders, yeah just set off when you like, our emergency control teams are ready.

Then at the controls they wait there in complete isolation , no idea what is going on out in the war zone, a Radar character (why does Radar remind me of Matt?) at each gate , watching for riders and shouting "incoming" . As the riders come in, a list of needs is read out "Food, drink, neck gone, knees hurt, saddle sores, broken spoke.." And the emergency control teams are scrambled. The riders patched up and sent back out into the war zone.

Suicide is painless...

Re: Managing the spikes at controls
« Reply #186 on: 01 September, 2017, 02:28:59 pm »
M*A*S*H approach summed-up my ride.  On the basis of "no plan survives first contact with the enemy" and "plan too early, plan again", I just decided to 'wing it' as usual and ride the bike and see what happened. No spreadsheet, no checking average speed etc, just making sure I had time in hand and only occasionally did I actually check where I was with regard to control times. I wonder what the split was in terms of folk like me and those who had a cunning plan?

Re: Managing the spikes at controls
« Reply #187 on: 01 September, 2017, 02:36:02 pm »
I'm guessing that might depend on the age of the riders.  Younger people (assuming a reasonable amount of intelligence) shouldn't find LEL too gruelling as long as they are not too ambitious for their ability.  It must get harder the older you get, although experienced riders can bring that to bear.  The older you are, the more a plan becomes helpful, I would have thought.

Peter

Phil W

Re: Managing the spikes at controls
« Reply #188 on: 01 September, 2017, 02:42:47 pm »
Matt spotted at St Ives


Re: Managing the spikes at controls
« Reply #189 on: 01 September, 2017, 03:00:54 pm »
In 2011 we got notice of changes in PBP arrangements from an article by Jan Heine. We arranged to get an interview with Jean-Gualbert Faburel about the new arrangements.
https://youtu.be/t68hZ54Bpuw

The gist of the changes was that the start time was brought forward, to enable more riders to get beyond the Loudeac bottleneck. There would be an intermediate stop at St Nicolas de Pelem. Wise riders had long gambled on the 'secret' control at Corlay, or St Martin, and this formalised that arrangement. Jean-Gualbert also emphasised the benefits of 'living off the land'.

What actually happened is that a sudden rainstorm hit the area between Loudeac and Carhaix, and riders holed up at St Nicolas, turning the marquee there into a convincing replica of M*A*S*H. I got it in the neck from some of the victims, for guiding them to St Nicolas.

I had been asleep in Heather's car at Loudeac during the rain. Heather likes the PBP atmosphere, so I have to ride or film. It gives me an opportunity to report on the event, without getting dragged down by 'events'.

The PBP process is fascinating. IT is one component of it, but it's not until you get the medal, brevet card and brochure in your hand, that it is over.

Re: Managing the spikes at controls
« Reply #190 on: 01 September, 2017, 06:41:25 pm »
I can remember that night ;)
I was riding behind the storm, could see the lightning further up the road. When arriving at St. Nic the bike parkings were packed so I asked a volunteer if it doubled as a secret control. The answer was no so I immediately continued. It took until the following day for the St. Nic sleepers to overtake me.

Re: Managing the spikes at controls
« Reply #191 on: 01 September, 2017, 08:29:24 pm »
Here's the footage I have from St Nicolas in 2011.


There's a clear line from the 'Loudeac Problem' through preparations for PBP 2011 to St Nicolas.
Data acquisition on PBP2007 and PBP 2011 was similar to LEL 2013. There was no check-out, so information on time spent at controls would have been anecdotal. Heather shot  comparative video at Loudeac in 2007 and 2011, during the 90 minutes I slept, which illustrates conditions.

PBP 2011 had electronic trackers, I still have one attached to my left shoe. That provided data on control occupancy, so would have informed control provision in 2015. PBP adopted the start system used at LEL 2013.

Pre-qualification for PBP 2015 enabled the start groups to be tailored to length of ride in 2014. That meant that the most experienced riders went off first. That seemed to ease the Loudeac problem. Data acquisition was compromised by a problem at Fougeres.

There was a problem with the pre-ride meal, which meant that the early starters scoffed the rather nice finger food. Security issues also meant the finish was disappointing compared to earlier years. Subsequent events around Europe show the wisdom of that decision.

The conclusion I reach is that these events are not perfectible. Completion depends on strength, cunning, resilience and experience. Those characteristics are rewarded, and there will never be a magic formula that ensures fairness in the result, where the 'deserving' get their just desserts.

The problems in these events are more social and political. There's an ageing volunteer base, and organisers are afraid that analysis of performance will discourage retention of volunteers and recruitment of new ones.

Statistical analysis seems to offer a way looking at the way the event runs without making value judgements.

Phil W

Re: Managing the spikes at controls
« Reply #192 on: 01 September, 2017, 10:24:57 pm »
I've had a lifelong ease with and love of mathematics, it is what my degree is in, and it equips you with a number of great tools and skills for life.

You're right statistical analysis is one of those great tools for looking at an event without starting off with the barrel of distortion that is your perspective and memory of it.  There has been some great studies into the nature of our memories, and basically how piss poor they really are, how you can end up with false ones etc.. Memory when sleep deprived is even worse.

The trick with any analysis is not to jump to conclusions before you've done it. That would be a bit like putting the cart before the horse.

Re: Managing the spikes at controls
« Reply #193 on: 01 September, 2017, 10:36:42 pm »
My instinct from an 11.45 start was that I needed to reach Pocklington on the first night to sleep, which would give me the best chance of finishing within time.

I saw riders trying to book beds at Louth northbound. I thought that was a terrible idea. I would expect that the majority of riders on a similar start time would have failed to finish if they'd slept at Louth. But it would be interesting to see how the numbers play out. I had a conversation with NeilV on this subject - he'd slept at Louth! And finished just fine, thank you.

Re: Managing the spikes at controls
« Reply #194 on: 01 September, 2017, 11:16:58 pm »
My feeling from an 1145 start was pretty much the same, that Pock would make subsequent days that bit easier - but I started feeling sleepy as I neared Louth, and really didn't fancy another 100km through the night with few or no services.

In the event the queue for a bed was about 20 mins, but the queue for food was more like 45, so from a 2250 (ish) arrival it was getting on for 1am before I was in bed. (I know - at least I got a bed and food.)

Finished fine, without ever feeling under massive time constraints, but equally never felt like I'd got enough of a cushion to be truly comfortable.

Re: Managing the spikes at controls
« Reply #195 on: 02 September, 2017, 12:10:56 am »
My instinct from an 11.45 start was that I needed to reach Pocklington on the first night to sleep, which would give me the best chance of finishing within time.

If you're fast enough to reach Pocklington while it's still night time from an 11.45 start you don't need to worry about not finishing within time.

I was in the same start group as you. If there'd been beds available, a couple of hours sleep at Louth would have fit my schedule perfectly.

Re: Managing the spikes at controls
« Reply #196 on: 02 September, 2017, 03:30:52 am »
When the anonymised data is released...etc...etc.
"Yes please" said Squirrel "biscuits are our favourite things."

Re: Managing the spikes at controls
« Reply #197 on: 02 September, 2017, 10:29:53 am »
I'm going to defend Matt, and his comments about the tracking system.

As has been noted Matt was with us at St Ives.  The first control after the start.  As Controller I took the view that the tracking system northbound was of little help to us, as we knew what the start pattern was so worked with that.  I suppose it might have been possible to determine how many actually started compared with the anticipated schedule, but I never discovered how to do that.  In any case, with nearly 200 riders hitting us per hour for quite a bit of the day, the controllers didn't have much time for digging around in the system for more information, which at that stage wouldn't add much to our knowledge anyway.  As we were also recording riders through on good old paper, we did have a pretty good idea how many had passed through at any time.  Again, it wasn't particularly relevant for most of the day.

Matt wasn't with us for the southbound riders, so has no experience of our use of the system after he left.  We did use the analytics, and found them very useful.  They would have been more useful if I could have spared another PC to run them alongside the card-logging function.

Re: Managing the spikes at controls
« Reply #198 on: 03 September, 2017, 02:09:04 pm »
There was no food voucher system at Thorne in 2009.

Re: Managing the spikes at controls
« Reply #199 on: 03 September, 2017, 02:13:49 pm »
Mattc possibly has difficulty with IT, as he was unable to register himself as a volunteer.