Discussion: Butterflies and course setting software

What are my software options if I want to set a course with many butterflies (orientshow style)? I want to define the butterflies and then have the software generate the variants, much as condes manages with the hagaby format.

I've never done anything like that except for a relay using OS2003 relay software. I know OE2003 has a one man relay mode, but I've never used it.

In Condes, I defined forks, not butterflies. But a butterfly would just be a set of forks with repeated ABC options. For a three-loop the diagram view would look like:

common control

A B C variants (as long as needed)

common control

A B C variants (repeat of above)

common control

A B C variants (repeat of above)

common control

You can export this in XML format and try importing those courses into OE. I was able to do it into OS with minor errors---the distances didn't display correctly.

In OS, you have to tell it what variant each person has. OS didn't like this method, because it wants to assign repeat variants such as AAA and AAB, but you never want repeated letters, at least for these three loops.

For large events, we usually use excel for convenience to easily enter competitor data, so I had to figure out a way to define in excel each competitors variant to the satisfaction of OS/OE and to the actual correct variant selection that you've set up. For the ABC loop above, there are only six options, so the excel lookup table is pretty easy. (I think what I did with the relay was backup the event, then have OS assign (incorrect) variations. I then exported those variations, studied the format, then picked my own variations and parroted those new variations back to OS in the right format.) Basically I make a table that codes each variation, and then I number them 1 to whatever. Then I just assign courses by typing 1, 2, 3, etc, and the table figures out that 1 is AABCA, 2 is ABCAA or whatever.

I think Condes can print maps with names and/or bib numbers on them. I've never done that, but I'm going to try to figure it out before I do another relay. I always used a printout of the excel spreadsheet to print the right number of each variation manually, and then used that same spreadsheet to find the right map and write that competitor's bib number on it. It's an easy chance for an error, and scary because of that.

I'm really interested to see what other people recommend. When I did the above, it felt like I was making it up as I went along, and the help files didn't seem to help much.

I recommend you mess around with the one-man relay stuff in OE. I have a feeling OE or OS will let you design much more complicated variations than Condes allows (seems to allow?). The big issue is can you keep everything straight.

That sounds a somewhat inelegant solution. As you say, the risk is keeping everything straight. The attraction of condes with hagaby is that the maps for each course can be generated with the appropriate control structure and control descriptions. Finn indicated the next condes version should handle this, but I was hoping that maybe the software proliferation for event administration might have been replicated to a degree for course design software. Because my event is in a month.

Autodownload copes with butterfly loops automatically, you upload the default course and it works out butterfly loops when people download

It has limitations but I believe these are in development at the moment (I asked for them anyway!):
can't do both types of butterfly on the same course (normal and figure of 8)
doesn't let you view the variations

a lot of it is automatic behind the scenes stuff, but it works

To be clear, Condes is printing the courses and the map numbering and control descriptions are all correct. I've just manually selected the correct variation and printed it, rather than importing a competitor list into Condes and having Condes print the map with the person's name or bib on it---I think it can do that, but I'm not sure.

{edit: I'm looking at the online help, but I can't find any documentation of this feature. So---I'm guessing it can't.)

Condes 8 has added some features for relays where you can force relay legs onto specific forks, and where you can bind a leg to multiple course pieces (If the course chose A at some point, you can bind later pieces to also choose A later.)

New Relay Features

Condes Relay Support Online Help

I think Finn is currently developing the ability to add names to the forked courses. But I'm not 100% sure, and I don't know any timeline

Condes can be forced to generate butterfly courses as course variations (quite complex ones, e.g. the prologue course at the 2009 Australian 3 day championships), but the problem with event software is that some don't recognise the variations, or if they do, treat them as individual courses, rather than variants of a single course. In Autodownload, you can define a single variant and the software will recognise all variants, thus achieving the objective of a single course, but it is limited (e.g. won't handle loops within loops or combinations of butterfly/phi loops). Manipulation of result files might be an option post-event to publish aggregate results if your loops aren't recognised by the software, but for checking on the day, event software that recognises variations (e.g. AD) will at least provide you with an option.

I am trying the cedarcreek option for now. Thnx. I will be using Geco as the event software. Will report on how it all works.

Regarding Geco, it's one course per variant. So the typical setup for a finale is:
- one course per runner = one variant
- one course does not necessarily mean one butterfly. Actually, we often set two butterflies, each with 4 loops, for the finales
- ranking based on category (so both men and women can run on the same courses and still be ranked separately)

If you plan to perform the full workflow of an orientshow, with quals, semi-finals etc. don't hesitate to ask further questions, there are some subtleties (actually the zip contains a dummy simulation of multiple stages for a full orient-show in the demo folder)

I have never use Condes, just Ocad, so can't really help on course setting. There is one loophole reported by a friend with Condes though: numbers should be formatted with the dot separator, not the comma (like 42.315 not 42,315). I guess this is linked to the regional settings of your computer, so probably a non-issue in english-speaking countries.

And don't forget to test! Geco is still a relatively new software.

Here are some things I looked at when I was trying to figure out how people did forking:

This one used to blow my mind regarding how they printed maps and got epunching to download: DxDeluxe 2005 (JS Open). Look at the gaffling diagram and the excel spreadsheet, as well as the maps. What's most surprising to me about this event is how simple the diagram appears but how the courses are all over the place.

The Jukola map archive (which needs to be updated!) usually shows the map with the forking diagram. Here's a good example from 2003. You can see how the Condes 8's new features are just picking up what is standard practice at the big relays---certain forks are reserved for specific legs, and certain later forks are "bound" to earlier choices.

I looked at how OS handled the courses, and it actually is a little easier to understand than Condes. Condes has to have the entire course modeled beginning to end (so it can print a full map), while OS (and I presume OE) doesn't---at least it doesn't appear as a long continuous string---it's broken into pieces. It looks like OS breaks the course into codes like 1A, 1B, 2A, 2B. For one of my courses:

1A = 208 191 211 213
1B = 208 227 228 213
2A = 231 218 221 232 222 234 203 205 225 214 215 219 202 237 238 239
2B = 209 233 221 232 222 234 203 205 225 214 215 219 202 237 238 239

I underlined the forks. I guess here each relay leg has to run one variant of 1 (A or B) and one variant of 2 (A or B). I'm guessing OE will be the same way, except with your many loop butterfly you'll have higher numbers and more than A and B. I'm guessing it can handle many numbers (probably up to 4 or 8 or more), and at least A-D for choices of variations.

Anyway, choose the software you feel more comfortable with. As long as it can export in IOF XML format, you are good to go.

I suggest you also try OEvent:
http://www.oevent.org

It should recognize variations and import them as variations of same course, and then you have a special dialog to distribute those variations to runners.

Just coming back to this discussion. I've always just used purple pen for setting but I'm setting a relay and as far as I can see purple pen doesn't do gaffling. I had been just doing multiple iterations of courses as I don't have many but that feels a bit clunky, with a lot of room for error.

So, what is the best/easiest bit of software to do this in. Preferably something with an interface as simple as purplepen. I have OCAD but find it a bit user unfriendly - so CONDES maybe?

Thanks in advance

I only have experience doing forks with Condes. I think you can try out the software as a limited controls/limited courses free download:

http://condes.net/index.php/download/

If I recall correctly, the sample course file that is included is quite good at showing what is possible.

I will say, though, that forks take some time to figure out. The basic trick is to use the course editor, where you see the map and add controls to the course by clicking on the control circle with the correct connection line highlighted, but then you double click on the course name to open a dialog box for the course. It will have a list of the controls in order. Then you have figure out how to select where each fork goes, and insert the fork (of the correct type), then manually add the correct control numbers for each leg of the fork using sort of a click-button interface. I *think* that it's best to not enter any of the forks in the map-based course editor, but just the common controls before and after each fork, and then insert to add the forked controls in this dialog box. (What I usually do---incorrectly---is enter the AAA variation, and then end up deleting controls in the fork that aren't "in the fork" as defined in this dialog box. I hope that makes sense.) I usually draw a little diagram for each fork on paper showing the common controls and the controls in each leg of the fork, so I can just add them without changing back to the map view (course editor).

My point is that the Condes interface is a little intimidating if you're starting out to do forks, but it is my favorite course setting software, and well worth the time to learn. The Condes license is for the club, so it's extremely convenient that way. I have said before that I think Condes is the single best value of any software our club has purchased. I've emailed bug reports to Finn (who wrote Condes) at midnight my time, and found new software ready to download when I got up. (But I haven't submitted bugs in a long time. My point is that he's really responsive.)

If you have any questions ask here or feel free to email me.

The one interface button in Condes that I tell people about looks like an opening door with a little left-right arrow: "Show or hide print area frame on screen". It toggles on-and-off the print window lines. I was pulling my hair out one week (years ago) trying to hide those stupid things, so I like to point it out.

I will second those comments about Finn's bug fixing timelines. How many of him are there?

I'll third support for Finn and CONDES.

Making courses in condes can be very easy, you just click where you want each control. For relays its more fiddly. The trick is to create all the controls first, then get the course dialog box open, which creates and shows all the forks. Then you make the course by clicking for forks and typing in control numbers in the dialog box.

Once its done, condes automatically generates all possible course combinations, for individual maps and generates allowed combinations for teams, reducing the scope for human error.

Before you finish the courses and print the maps, make sure that the epunch software that will be used can handle the variations. MeOS deals with butterflies very easily, but when it generates the splits, normalizes the times so that it appears that everyone ran the course in the same order. This is fine for comparisons, but doesn't work if you want import to RouteGadget, for example. Erik Melin of MeOS kindly gave us a patch so that we could export our results to RouteGadget, but the product off the shelf does not have this feature yet.

I understand that OE2010 handles things like butterflies, but I've never been able to use it. Always consult with the Epunch crew on course design. We had a rogaine last month that had 69 controls. OEScore only handles 64. Most people had punches that could hold 60 or fewer.

I understand that OE2010 handles things like butterflies, but I've never been able to use it.

The way to do butterflies in OE2010 would be to import each course variation into Courses, and them assign the course individually to the competitors - Courses/Competitors, after checking the option for Individual Courses under Courses/Classes.

Which means that you must assign a variation to each runner, instead of having them pick up a map at random and having the software assign the variation when they download.

Yes, that is correct. I guess the other option would be to make the code checking as Mixed, and then put the butterfly controls as Mixed controls, i.e., gotten in any order, then some manual review would be required.

Or use the feature to read the courses when they download, but that has a mess of stipulations.

So, when it comes down to it, your suggestion that course setters check with the SI Manager before devising interesting new course formats is a good idea.

Am I the only one using OCAD for course setting as well as map making? :-)

I have used OCAD for several years for one-man relay events, in which case it will print out all the individual maps for each competitor, marked with their bib number and map number/order.

I have not tried butterfly/phi loops, but I assumed that would also work OK?

Thanks everyone. I've been sent some ocad instructions so as I already have it I'm going to have another go at that. If I'm still having trouble I'll look into condes

Interesting about course assignation. This is for sprint relay so I need to make sure whatever variant the first man/woman gets, the second gets the alternate gaffles. I suspect the easiest way to do it is to assign to competitors myself.

I've only ever used 0CAD for course-setting, though I have worked with other programs when I was reviewing courses designed by others. 0CAD seems fine to me, and I've never felt a need to use anything else.

I recently used PP for the Traverse which used 2 butterfly loops. Basically I just made 4 different courses, and assigned one to each competitor by name.

yeah, something like that was my original plan - depending on how many gaffles I do it's still viable - just trying to reduce the chance of human (my) error

I also have only used ocad. Have done butterflies but never tried gaffling-yet.

Never mind jargon on A Meets, Re-Entrants etc what the hell is "gaffling"

I have been orienteering for over 40 years in the UK, Australia and Canada and this is new jargon to me and I feel excluded!

"Gaffel" is the Swedish word for "fork".

Well now I'm forked

haha, I only actually realised it was swedish a couple of months ago - it's the only term I've ever heard used in the UK!

For a sprint relay (two women, two men), 2 x 2 Motala method is the simplest option. Forking is illusory, although you can pretend it's happening so that your course planning software can facilitate map printing.

The Blue Course on Day 2 of Mountain Lakes had a Butterfly loop. The course setting was done in Ocad 11.

Is Motala 2 x 2 Vannas ?

Why is forking illusory ?

Motala is one course per leg (e.g. courses a, b, c for a three person relay), Vannas is basically two Motalas (courses a, b, c and courses x, y, and z) with a common control typically half way around so team members run one of a, b, c and one of x, y, z but in total the team runs a, b, c, and x, y, z.

For a sprint relay (two women, two men), I am assuming that the men's and women's courses are complementary (i.e. the courses in total run by the women members of the team are the same for all teams, and likewise for the men), so even if you have forks in the course, they are defined, so that you effectively have a Motala structure as there are only two possible courses for two people.

I just made a quick test in OCAD 11, setting up a mixed sprint relay with leg 1 & 4 reserved for the women and legs 2 & 3 for the men:

I defined four first controls, split into two groups using leg grouping, i.e. 1/4 vs 2/3, then within each group I defined a two-way fork. The only problem here is that OCAD does not allow you to do that directly, for the second group it wants an intermediate control before the internal forking, but it was easy to work around with a temporary control which I deleted after defining the team variation:

In Ascii graphics:

S1
+----------------+
1 2
4 3
+-------+ +------+
31 32 33 34
+------+ +------+
| |
+----------------+
35

etc.

This should allow OCAD to directly print all the maps for each team, with team/leg numbering, while making sure that the forking is kept correct for all of them.

Terje

@Hawkeye

I'm missing something, I think - surely you can 2-fork the course n times, giving 2^n courses. Of course each of those courses has its complement which the other same-gender member of the team has to run.

Seems like more fun (aka confusion) than Motala.

3 forks on each course would be 16 courses all up - should be manageable in PP, maybe jayne has 4 forks ? Cutlery overkill.

Thanks Undy, you are correct. It is me who is missing something (e.g. statistical analytical skills).

Good job I'm a statistical analyst then :)

32 courses is an awful lot to keep track of. 4 forks sake use condes.

The WOC93 Men's relay had 128 different forks. 32 teams of four runners each, and everybody got a different map. And the courses were all offset-printed. Each map wen through the press four times for purple ink. Magnificent overkill. (Likewise for the Women's, though there were not enough teams entered to use all of the maps.)

Undy, rather ambitious for Purple Pen.

In Condes (possibly somewhere in the help, and look at the relay sample in the demo event file).

Course/new/4 relay legs
Insert leg fork
In the leg distribution dialog, drag 4 (column D) under 1 in column A, and 3 under 2 in column B, click OK
In the leg forks, insert either controls or forks.
(CedarCreek's method in his post of Dec 4 describes a useful way to create your courses).

Once the course is created, go to the Relay/Open Teams Allocation dialog box and add teams. If you have a printer that can print double sided, print maps with team, class, and leg allocation.

OS can probably accept Condes XML relay teams export, but Autodownload only has a CSV import, so if using AD, you might have to copy to the clipboard and massage the data, but would appreciate hearing of a more elegant solution.

Thanks JJ, I had forgotten the joy of using an over printer for relays. I think ours is still kicking about...., may have to buy new sticky tape.

JJ is talking about a real, actual offset printing press to print courses.

Too late, I'm going to the gear garage this weekend.