Discussion: Chasing start: Yea or Nay?

For an upcoming Eastern Canadians, we'll have a very technical but limited in square kilometers terrain, that would be perfect for 2 middles rather than a long and a middle. If 2 middles, would you prefer 2 individual races and awards, or a chasing start on day 2 and one award? We will likely have many juniors and masters, total participation about 200 people, not too much visibility.

That is brave of you to ask this on Attackpoint as any minnute now you are likely to get some smart-assed Aussies chiming in saying they are a Master with limited visibility and they prefer ....
However as you asked my suggestion would be for a modified chasing start - two day total time; one award. Those within half an hour of the leader (perhaps less if it is the middle distance) start in the chase. The rest start at a regular interval before the chase starts.
Before? yes, they are presumably the weaker orienteers. Don't have yourself waiting forever for them to get out of the woods and lets have some of them back to cheer on the chase winners.

I am a master with limited visibility and I prefer...

I used to be able to visibly master reverse chasing starts and found them exhilarating orienteering - when I was fit! Always nice to catch glimpses of people ahead as a motivational incentive. Chasing starts though, I found them a bit less interesting - and meant for a m-u-c-h longer day for the officials. gordhun's suggestion is not a bad one, as usually it's the faster orienteers who actually care about chasing / reverse chasing starts and the racing aspect of it.

Thank you people for paying attention! Pardon my limited English and limited experience: could someone please educate me on what's reverse chasing start? Also, we'll likely have a mix of novices and pros, so we would like to make it fun for all, especially teens. I am pretty sure the pros will enjoy the intricate terrain, quality map, and old rivals no matter what.

please educate me on what's reverse chasing start?

You have the slowest person start first and the fastest person start last. With this format, the event staff will have a shorter day, and the orienteers will be running with more people in the woods. This format doesn't doesn't have the "first across the line" aspect, though.

And if done properly, it can result in a mass finish, which can provide all manner of excitement.

Snail: good on the Ramblers for not forcing the wrong format onto a map.
My suggestion would be a mass-start one-person relay format for one of the days. Very good for helping to attract new rivals.

Which would mean a mass start for the previous day's non-finishers & non-participants?

We do a couple of those Chase events per year. Participants love them as they all get a shot at winning and organizers don't have to wait for control pickup (in theory)
Couple of headcams with this one. The results on this was a mass finish even though the spread at the start was 30 minutes.

Chase the Rabbit

Fantastic video Platterpus! May I share?

Nev-Monster: looked it up, sounds super-fun, my biggest fear: we'll screw up something with the electronics, plus explaining it to newbies in 2 official languages that we all half-master.. one of these days!

I've found that mass finishes don't live up to the promise (thankfully for the organizer and the competitors competing for limited finish punches). Most orienteers are too variable. But maybe someone has had different experiences.

Just did a reverse chase with BAOC's sprint weekend and really enjoyed it. Our 11-year-old, the youngest participant by a good bit, also really had fun.

They also had the top 4 men and women run together in an A-final first, so we all got to spectate for that as well. Fun all around.

Snail: Feel free to share of course. I think all competitors finished within 3 minutes of each other in this race and most within 90 seconds of each other.
bksk1lzused a formula that he called the "Slo-Co" (Slowness coefficient). Worked brilliantly or maybe we were just lucky. And the only reason that fingersticks were used was to check punches since lapsed time in this race was irrelevant.

Participants love them as they all get a shot at winning

Really? I thought it was still based on cumulative time regardless of who crossed first in a reverse chasing start. The only difference being that in theory the last starter should chase down many - although probably not all - the previous starters.

Can we get videos like that at major North American events? That'd be way cool. This video should be linked on the OUSA website as a demo of the sport.

The handicap is in the start time so that after the start, it does not really matter what your elapsed time is. The finish order is the only thing that matters. So maybe this is a different format then what I understood the chase was about, but this format worked great because everyone knows these two things while out there: Pass as many as you can, don't be passed and you will do well. In theory, if the math is correct, the person that makes zero nav errors will end up with the win.

One issue I see with that is if someone made an unusually (or intentionally) big nav error in the qualifier and so started further in front than he or she otherwise would have, thus giving him or her an unfair advantage.

Yes.. without enough data the math might fail. In our case we used at least 3 race totals and averaged them out. In most cases we used more than 8. But there might be some way around that.

I have run a number of what we call "Ironman" events. Three sprint races with handicap starts aiming for a mass finish. The result was based on the sum of placings for the three races. I discovered that not only is orienteering performance highly variable, but the lure of slower runners ahead changes the performance of most orienteers. They run faster than in normal sprints with allocated start times. Setting handicaps became heavily influenced by performance in the same event the previous year. But this introduces variability because of changing fitness through time (and the rapid improvement of juniors). The solution was to change the handicap of the winner of the race (and anyone else who over-performed) for subsequent races based upon popular vote. The vote was fun, but runners then started running strategically, often saving their best for the last race to avoid penalty. So then everyone was watching everyone else looking for the bandits. I think this greatly improved the experience in an unexpected way.

so a reality show version of orienteering then

Well, with head cameras it could be Big Brother, and with voting it ticks another reality TV box. What about the unusual contests? A course entirely through a swamp? Moving controls? Team score event in which the first to a control takes it down? Controls hung three metres off the ground, requiring a human pyramid to punch? Maps as puzzle pieces that need to be assembled first? Sounds like a great holiday party event.

Snail: I would really consider it for one of the days. It isn't that hard to organize. And remember you have a club 2 hours to your west with many experienced officials.

Somewhere through the thread I think the people posting these excellent and fun ideas lost track of the fact that you will be hosting the Eastern Canadian Championships and thereby are probably constrained by the options you have.
If the rules say you should have a sprint, middle distance and long then so be it.
Given the constrained area then the best idea above then is the two loop race for the long. In effect then instead of setting two middle distances you set three - one for the middle distance champs and two the next day for the long, run consecutively without a break in between. Controls can be used more than once on a course if care is taken to make sure they are approached from different directions each time.

Anna, Gord and any other Canadian event planners who might be reading, if you feel that your event would be better using some other format than sprint, middle, long, please reach out to the Technical Committee (who are tasked with sanctioning) and we'd be happy to discuss the options with you.

This discussion in priceless: brainstorming, and humor, and voice of reason.. thank you guys! And by the way:mark your calendars: July 29-30, Morin Heights, Québec, 2 days of Orienteering fun followed by Junior (Sass Peepre) Camp!

It is too bad you find yourself short of terrain in Morin Heights. For a number of years it was the center of the Canadian orienteering universe as Viking had large maps on each side of Jackson Road, another immediately on the south side of town. Azimuth had two more maps nearby. Montreal OC had maps of a scout camp and a Y camp not far away as well.
I have seen new AR/ Rogaine maps of the area for miles around with amazing contour detail.
I'm in for that weekend. for sure.

Hey no kidding Gord! Some of the terrain is still there, despite big losses to new developments and private ownership of the land. Rogaines are still very much alive and exiting, due to dedication of the last man standing. And no, we are not short of terrain, we are short of money to map it. I hope renewed Ramblers will get it together, and reconquer Laurentians for Orienteering.

If you need motivation to do new mapping, you may have a headstart on many of us with Quebec Open Data.

https://groups.google.com/forum/#!topic/lastools/B...

"Province of Quebec has announced that the 200,000 square km of existing LiDAR will become Open Data with free download for several derived products:
* 1m DTM
* 2m slope map
* 1m hillshade
* and another forestry product.
Furthermore MFFP Quebec plans full coverage on the entire provice of by 2022 by aquiring 70,000 square km a year of the remaining 300,000 square km at a density of 1.5 and 2.5 pulses per square meter depending on the dominant forest type."

No kidding Bryan that Quebec Open Data has been like opening up the candy store for us kids.
For example five years ago I produced a small map on a private property using the available 10 metre contours. It was very laborious making contour corrections and frankly I just concentrated on the hill tops and valleys. This week Andrew Cornett sent me two cut outs of a larger area including those properties containing the 1m contours (with every fifth metre highlighted in purple) and a colour coded map of the forest cover.
I have compared the contours of the two maps. The new Quebec product is mouth-watering in its detail.
Guys proficient in Kartapaullatin could probably be turning out Quebec orienteering maps on a daily basis.

Open (vector) topo data is great, but there is nothing comparable to raw LiDAR point clouds when you want to quickly make great orienteering maps: You can always draw in roads and houses from aerial photos but for contour detail and vegetation there is no substitute for nice/dense LAZ files.

Open (vector) topo data is great

Yes, it is. And when you're subsisting on the mapping equivalent of starvation rations (30m-90m resolution DTM, anyone?) you'll happily eat all the open data you can find.

I'd be very interested to know the Slo-Co calculation. Any chance to get information on it?

In Calgary we ran a 10-week urbran race series, aimed primarily at the older juniors. For the first nine weeks every runner was given a handicap (our version of the Slo-Co) that changed depending on the performance. For the 10th race we did a reverse chase. As suggested above, all the really slow people began in a mass start ahead of the rest.

I think it was really fun for participants. We could have improved a few things, such as:
* perhaps a couple of butterfly loops in the reverse chase so the chasers had to be a bit more on their toes, rather than just chasing down those in front
* better data. The two winners clearly won because of way-too-high handicaps. This will obviously be more accurate for those two next season, but perhaps we'll have another couple with bad handicaps
* manual override when the handicapping fails - kind of like the "voting" suggestion above.

We had been concerned about 'gaming' the handicap to do better in the Reverse Chase. But in fact the weekly incentive to improve your handicap seemed to play a bigger part in peoples' motivations, and they seemed to play it true. The bigger problem is dealing with injury or other (legitimate) reasons that people might not give 100% in a race.

AZ: The results will never be exact but getting close is the key. We have done 3 so far and 2 of them turned out great. #2 not so well because of miscalculation of finish time. The course designer will run the course ahead of time to give a factor for the formula. The time that it failed, we did not do that and only guessed. If you have too short of a course, the fastest will never have a chance to catch the people with the early starts. If you make it too long then the fast ones will always catch up.
Bksk1lz came up with the formula but not sure he is on here that much so I will encourage him to join the conversation. On your note about trying a few things. The only thing we have done is to make the course go past the start halfway through the race to get an idea of the progress. More than one of those would be an excellent idea.

The Bendigo Ironman uses a database of sprint event results. From these I calculate a kilometre rate (for 2 kilometre courses) using events with similar terrain and navigation standard. Those for whom there is not a good number of results are then adjusted according to the performance of similar runners. This is always necessary because of people coming out of injury or improving rapidly in junior age classes. From there its an arithmetically simple matter of working backwards to get the start times. Take the expected finish time of the slowest runner as zero time and add time representing the difference between the slowest runners running time and the expected running time of each other runner. When I started I used the results of normal sprint events and found that the fastest runners rarely caught the earlier starters. This is because the chasing format makes relatively little difference to the speed of the faster runners, but can make a big difference to the speed of the slower runners. As my database included more and more results from chasing start events, the handicap accuracy improved. If I get a half the runners finishing within a minute I am very happy. Most of the time its less than this.

@AZ So we just do one Chase, at the end of a series of normally-scored sprints. For any given race, a runner's Slowness Coefficient is their time divided by the winners time. So the winner gets a SloCo of 1. Lets say his time was 10 minutes, then someone who did 11 minutes gets a SloCo of 1.1.

At the conclusion of the race series, we take the median* of each runner's SloCo results to get the series SloCo for that runner.

Then for the Chase race, we have the meet director estimate the time he would need to run the course. (This part if the process is probably most susceptible to error). To get the predicted time for each runner, their SloCo is multiplied by the ratio of meet director SloCo over meet director Estimated course time. For example if the meet director's SloCo is 2 and he estimates 20 minutes to run the course, runner A with SloCo = 1.5 is predicted to take 15 min, runner B with SloCo = 1.2 is predicted to take 12 min, etc.

From there, we handicap everyone's start times so they will all arrive at the finish line at the same time if they perform as predicted.

*Using the median instead of mean helps negate the effect of outliers - getting lost, injured, etc.

The more data, the more accurate the SloCo is, and estimating the time to run the Chase course is tricky. But all in all, it's worked surprisingly well. In the best case, there was a group of 10 or 15 all finishing within a minute of each other. The winner was about a minute ahead of the group, and there were some stragglers.

Thanks for the details - that's great to know.

Our handicapping was a lot more complex and probably less accurate ;-) I tried to model ours on golf handicapping. It 'worked' like this:

* in each race we set a "Par TPK". You guys didn't have to do this, except for the final chase. Having the "Par TPK" is intended to give everyone a target for each race.
* everyone had a handicap going in - either based on previous results, or in the case of newcomers their initial handicap was whatever would have given them the "Par TPK" in their first race.
* in each race, everyone got a "X-factor" - which is what handicap they would have needed to reach "Par TPK" for that race.
* after each race, everyone's handicap was adjusted to be 80% of their previous handicap plus 20% of their X-factor.

I'm going to have to think a lot to figure out if this extra work was worth anything ;-)

Why it might have been worth the trouble:
* everyone had an updated handicap after each race. This provided them incentive / motivation to come out each week and lower their handicap.
* results for every race could be compared "straight up" or with the "handicap time". Each has its own merit.
* runners had a target that was not dependent on other runners - only on the Par TPK and their own handicap

Still...

AZ It is an incentive to lower their handicap only until they realize that to win the whole schlemiel they need to raise their handicap. I'm always after my wife to play a few poor rounds of golf so that her handicap will go up and we can have a better chance in mixed tournaments. No luck there. She thinks I'm joking.

Ah - but golf handicapping defends against that (and like your wife, I'm pretty sure you're joking) by including only the best 10 of your most recent 20 results. I intend to add that feature (perhaps best 3 of the last 6) in order to reduce the likelihood of such cheating and also to even out scores for people that get injured / don't try hard / are getting better quickly.

The other way we defend against this is making the final prize pretty measly ;-) This way the incentive of self-improvement for nine races is much higher than ultimate victory.

In our scheme, in any case, I figure it is so rigged in favour of new-comers (who are likely improving faster than their handicap is reflecting that improvement) that if an old timer wins they will be met with withering scorn. So in fact winning the final race for old timers is a bit of a non-inentive ;-)

@AZ yes our method requires the winning time to be consistent for each race in the series. If you have a world-class guy show up once or twice, everyone's sloco for those races will be inflated. However, for us it's usually been a valid assumption that the winning time is consistent from race to race

I have a python script to calculate the slocos from IOF XML results files here: https://github.com/bksk1lz/SloCoCalc It's still a bit rough (though functional!) so I'm working on cleaning it up. If inclined, you could put some more intricate handicapping calculations in there.