Really thorny competition format requirement

[PeterGustafsson]

Hi!


Every once in a while, my club does events, in which groups get to try out fencing for a fee. Typical groups can be men´s clubs, bachelorette parties, company events for their employees, and the like.

Now, we have a company event coming up, and I could use some help with scheduling and figuring out a reasonable format.

The particulars:

2 pistes available, 1.5 hours competition time= 180 minutes piste time

1 hosting company, which has invited 3 of their customer companies. The hosts, 4 people, will take part in the fencing initiation, but not in the competition. Customer company #1 will have 7 people attending, while CC#2 will have 8-9, and finally CC#3 will have 4.

The paying company has requested that we will organize a format which fulfills the following criteria:
1. A winning team must be crowned
2. A winning individual must be crowned
3. The small team may not be at disadvantage.

I am reasonable sure that bouts will have to go to first to one, or at most 2 hits. Even so, this is going to be a really complex problem. Obviously, I do not have to worry about the competition format rules of the Swedish Fencing Federation, the problem is strictly about time constraint. To cut down time, I am reasonably sure that the format should require that no bouts between team members take place. Also, the format should cull fencers after at most 2 losses. I have some ideas, but I will save them for later so as to keep the brainstorming going.

What are your suggestions?


Have a nice time!

Peter Gustafsson

[KD5MDK]

You know, usually your formatting questions drive me crazy but for once you've come up with a question that seems perfectly reasonable.

I presume that the time constraint you listed is only for the competition part, and all the time spent teaching them fencing etc isn't included?

[kalivor]

My thoughts:

I'm not sure how to crown an individual champion without having team members compete. What if the two best fencers happen to both be on the same team?

With that in mind, a couple of suggestions:

Option 1: One big pool.
Option 2: Two still-big-but-not-as-big pools, winners of each pool fence each other in a final bout.

In either case, do some experimentation ahead of time so you have a good idea of the time lost to fencers hooking up on piste, etc, and then determine the time limits and score-limits of the bouts accordingly.

The individual winner should be pretty clear. Team winner is the team with

- The cumulative lowest placings of their top (2/3/4) fencers. So a team that had their top fencers finish 1st, 4th and 6th would have a score of 11.

- The lowest average ranking. (1st, 4th, 5th and 6th would result in an average placing of 4th).

Tiebreaker is the best individual ranking on the team. If still tied, move down the team list until the tie is broken.
-----
I like pools better for this sort of thing, as it allows everybody to fence the whole time, rather than having people eliminated early and dropping out.

Time-wise, two pools of 11 would each have 55 bouts. Imagining up to one minute per bout, and another two minutes (on average) to switch over fencers between bouts, that's 165 minutes, leaving 15 minutes to have the grand final between the two pool winners.

Alternatively, a pool of 20 would have 190 bouts -- 95 per piste. The results would be a bit more fair, as initial seeding would be spotty at best, but this clearly results in more of a time crunch when you have more than 16 or 17 fencers (120 and 136 bouts, respectively).

[Fiat Slug]

My $0.02:

How about if you run this as a team competition? Each person fences one bout against everybody else on the other two teams. Assuming team 2 has 9 fencers, that's 127 bouts. In order to maximize fencing time, perhaps you can have everyone fence dry so they don't waste time hooking and unhooking. Each bout can last 40 seconds or 2 touches. You can allocate another 20 seconds to switch fencers. I'd suggest that you don't stop the clock when the action stops. This uses up 127 strip-minutes and you'd have a team winner at the end.

For the individual event, ask each team to select one person themselves. These 3 people can fence each other for the individual medal. Perhaps they can fence to 2 minutes (again, without stopping the clock) or 5 touches.

Obviously, this requires a high degree of planning and coordination. I'm thinking that you almost need to have a bout schedule printed out for everyone. It might look something like:

10:00 strip #1: John from company #1 vs. Joe from company #2
10:00 strip #2: Jack from company #3 vs. Jay from company #1
10:01 strip #1: etc.

.

[rsy]

Option 1: One big pool.
Option 2: Two still-big-but-not-as-big pools, winners of each pool fence each other in a final bout.

This looks like a good solution except it creates a problem with the time requirement. 20 fencers and 2 strips means 45 bouts per strip if there are 2 pools and 95 bouts per strip if there is one pool and you double strip it. In an hour and a half, even dry fencing, you would be hard pressed to get through even 45 bouts per strip.

A possible solution is instead of 5 point bouts, make the bouts best out of 5 (that sounds better than calling them 3 point bouts). Dry fencing, using best out of 5 points and 2 pools of 10 you can get through all of the bouts in 1-1/2 hours and then the the rest of kalivor's suggestion is perfect.

-r

[shlepzig]

If you use strictly single touch bouts you should be able to do a 18-19 person pool. I think the following will work. I have a sneaky suspicion that unexpected math might give a team with a larger or smaller membership an advantage.

Best team has highest wins/team member average (bouts against same team members will cancel out this is where I think some sneaky math might happen).

Best individual has the highest number of wins. Fence off for ties, which are disregarded in the team standings.

-Shlep'
(PS I did a quick simulation on Excel and it looks like it will work)

[K O'N]

Hi!


Every once in a while, my club does events, in which groups get to try out fencing for a fee. Typical groups can be men´s clubs, bachelorette parties, company events for their employees, and the like.

Now, we have a company event coming up, and I could use some help with scheduling and figuring out a reasonable format.

The particulars:

2 pistes available, 1.5 hours competition time= 180 minutes piste time

1 hosting company, which has invited 3 of their customer companies. The hosts, 4 people, will take part in the fencing initiation, but not in the competition. Customer company #1 will have 7 people attending, while CC#2 will have 8-9, and finally CC#3 will have 4.

The paying company has requested that we will organize a format which fulfills the following criteria:
1. A winning team must be crowned
2. A winning individual must be crowned
3. The small team may not be at disadvantage.

I am reasonable sure that bouts will have to go to first to one, or at most 2 hits. Even so, this is going to be a really complex problem. Obviously, I do not have to worry about the competition format rules of the Swedish Fencing Federation, the problem is strictly about time constraint. To cut down time, I am reasonably sure that the format should require that no bouts between team members take place. Also, the format should cull fencers after at most 2 losses. I have some ideas, but I will save them for later so as to keep the brainstorming going.

What are your suggestions?


Have a nice time!

Peter Gustafsson

If you have the equipment I would fence the whole thing electric epee. Beginners love epee, it's objective and easily understood.

Would it be acceptable to split up the teams to CC1A, CC1B, etc, of 3 or 4 each?

If so, consider this:

*****************

You have five or six teams of three or four each. Do a pool of team bouts, which is 15 bouts (assuming 6 teams). You can double strip the pool, so you need time for 7 bouts end to end.

For each bout, you don't have time to go to 45, and that's likely too much fencing anyway. If you fence to multiples of 3 instead of 5 you save some time, and the important thing is for each fencer to get on strip once during each bout, not that they each fence everyone on the opposing team. Indeed, if I never fence my rival Jim I have some interest in inviting him out to the club next week to take lessons and eventually fence. Leave them wanting more, eh?

So, have each team carry a cup with the team member's names on slips of paper. The referee randomly pulls a slip from each bowl and announces the first bout. They fence to two minutes or three touches. Then he pulls the next two, another two minutes or the six touches. If this is a 3 vs 3 bout the last bout is already determined. Allowing 30 seconds for hook and unhook, since Swedish people are very efficient, that's seven minutes per 3 vs 3 bout. A 4 vs 4 bout would be to 12 rather than 9, and would take 9.5 minutes.

A 3 vs 4 bout is to 12 also. At the beginning of the last match the referee puts all the slips from the 3 team back in the bowl and selects one randomly to fence the last member of the 4 team.

Even assuming every bout goes to 12, the total fencing time is under 70 minutes. You'll have to not dawdle between bouts and in recording scores, but I think the whole thing will easily come in under 90 minutes.

The winner of the team event is the winner of the pool, decided in the usual way. The individual winner is harder since even if you keep track of each fencer's data there would be many 'ties'. A simple answer which also respects the character of team fencing is to make the individual winner the fencer with the highest indicator. This requires you to keep track of each fencer's TS and TR as well as team scores, but Swedish people are very efficient, so I have every confidence in your being able to do this.

*****************

If it's not acceptable to split teams I think you have to fence something similar with bigger teams. That could be arranged, but I don't like it, I can't see any way around the fencers on the smaller teams having to fence a lot more than those on the bigger teams. You'll have someone on a smaller team who doesn't really want to fence this much, and someone on a bigger team who's champing at the bit to fence but doesn't get to, it's a recipe for a certain amount of dissatisfaction.

K O'N

[larkascending]

Best team has highest wins/team member average (bouts against same team members will cancel out this is where I think some sneaky math might happen).

Intuitively, I think you are right.

Each bout between team members will drag their team average towards 50%. Whether this advantages or disadvantages them depends on whether the team is doing better or worse than 50% against their opponents.
Since the big team will not only have more intra-team bouts than the other teams but also have fewer inter-team bouts with which they can drag their average up or down, so this could have a pretty big effect.

I'd crunch some sample numbers but I have to do my homework before fencing and helping with set-up for our tournament.

[KD5MDK]

If you have the equipment I would fence the whole thing electric epee. Beginners love epee, it's objective and easily understood.

This is Sweden, they don't admit other choices exist.

[K O'N]

This is Sweden, they don't admit other choices exist.

Sensible country.

K O'N

[InFerrumVeritas]

Here's what I'd do:
Round robin bouts to 2, electric epee. Team-mates don't fence. Teams are picked by competitors (let them decide who's captain then pick one by one if you need to determine teams, like we used to do in PE classes). Keep track of which individual has the most wins. They are the champion. If there is a tie (even between team mates) have them fence off.

[PeterGustafsson]

Hi!

If you have the equipment I would fence the whole thing electric epee. Beginners love epee, it's objective and easily understood.

Would it be acceptable to split up the teams to CC1A, CC1B, etc, of 3 or 4 each?

No, each CC is to compete as one team.

Regarding weapons: previously, we have often done these events according to a hybrid weapon rules set. We use foil weapons, epee concept of priority, and sabre target. The foils limit the amount of pain they can inflict on each other, the lack of RoW rules makes it easy to understand, and I do NOT want those guys to have weapons anywhere each others legs or groins. Once we have weeded out the "pirates" in the beginning, we have often let the final or L4 fence electric epee.

If it's not acceptable to split teams I think you have to fence something similar with bigger teams. That could be arranged, but I don't like it, I can't see any way around the fencers on the smaller teams having to fence a lot more than those on the bigger teams. You'll have someone on a smaller team who doesn't really want to fence this much, and someone on a bigger team who's champing at the bit to fence but doesn't get to, it's a recipe for a certain amount of dissatisfaction.
K O'N

Given that not all possible pairings between non-teammates will be fenced, I am sure that one can design a format which will ensure roughly equal amount of fencing for members of different teams.



Have a nice time!

Peter Gustafsson

[PeterGustafsson]

You know, usually your formatting questions drive me crazy but for once you've come up with a question that seems perfectly reasonable.

I presume that the time constraint you listed is only for the competition part, and all the time spent teaching them fencing etc isn't included?

Yes, the time constraint is for the competition only, teaching them etc. is done in the preceding half hour.


Have a nice time!

Peter Gustafsson

[PeterGustafsson]

Hi!

My thoughts:

I'm not sure how to crown an individual champion without having team members compete. What if the two best fencers happen to both be on the same team?

Thanks for pointing that out. Let´s see what the chance is that the two best fencers are on the same team: (4/20)#(3/19)+(7/20)#(6/19)+(9/20)#(9/19)=
(4#3+7#6+9#8)/(20#19)=(12+42+72)/380=0.332. So, there is an about 1 in 3 chance that both the best fencers are on the same team, big enough so that it has to be taken into consideration.

I think that we will solve this by having the fencers fence members of other tteams as long as it is possible, but when a fencers no longer has any non-teammates to fence against and yet has not been eliminated, then they can start fencing their own teammates.

With that in mind, a couple of suggestions:

Option 1: One big pool.
Option 2: Two still-big-but-not-as-big pools, winners of each pool fence each other in a final bout.

I do not think that big poules are good here. They would necessitate a lot of number crunching in the middle of the fencing, breaking tempo and using precious time. Also, there would a a lot of bouts in which one or both fencers have very little to fence for, since they already know that they will not make the grand finale.

In either case, do some experimentation ahead of time so you have a good idea of the time lost to fencers hooking up on piste, etc, and then determine the time limits and score-limits of the bouts accordingly.

Points well taken.

Team winner is the team with

- The cumulative lowest placings of their top (2/3/4) fencers. So a team that had their top fencers finish 1st, 4th and 6th would have a score of 11.

This method gives a considerable advantage to the larger teams, since their 3rd out of 7 - or 9 - fencer will in all likelyhood be a better fencer than the one who is 3rd out of 4 on the small team.

I like pools better for this sort of thing, as it allows everybody to fence the whole time, rather than having people eliminated early and dropping out.

I am of a different opinion, since these are not seasoned fencers. Non-fencers fare better if they get as much help as possible in tracking what they are doing, and that is improved if the bouts of a specific fencer are bunched together. Poules do not build up to a grand finale, with teammates cheering and all.

Thanks for the input!


Have a nice time!

Peter Gustafsson

[PeterGustafsson]

Hi!

My $0.02:

How about if you run this as a team competition? Each person fences one bout against everybody else on the other two teams.

That would be 16 bouts for those on the small team, quite a few. When I did a nine-woman competition last fencing tryout time, I used a shellsort format which gave each fencer 4 5-point bouts, regardless of where they ended up. That seemed taxing to them.

Assuming team 2 has 9 fencers, that's 127 bouts. In order to maximize fencing time, perhaps you can have everyone fence dry so they don't waste time hooking and unhooking. Each bout can last 40 seconds or 2 touches. You can allocate another 20 seconds to switch fencers. I'd suggest that you don't stop the clock when the action stops. This uses up 127 strip-minutes and you'd have a team winner at the end.

The suggestions of dry fencing, less than 5 points, and short bouts are good, and we will probably use them. Filling in a reduced 20-poule diagram is less enticing, the fencers will constantly want to know how they are faring, and we would have to dedicate a club member to diagram work.

For the individual event, ask each team to select one person themselves. These 3 people can fence each other for the individual medal. Perhaps they can fence to 2 minutes (again, without stopping the clock) or 5 touches.

In order to conserve time, I would prefer a format in which all bouts count towards both team and individual results simultaneously.

Obviously, this requires a high degree of planning and coordination. I'm thinking that you almost need to have a bout schedule printed out for everyone. It might look something like:

10:00 strip #1: John from company #1 vs. Joe from company #2
10:00 strip #2: Jack from company #3 vs. Jay from company #1
10:01 strip #1: etc.

.

Yes, this would require a lot of coordination during fencing. Since these people are not trained fencers, there is a high probability that something goes awry during the fencing, and the time schedule breaks down. If there is a holdup on one piste, one can arrive at a situation in which one fencer is called for to fence at both pistes simultaneously.



Have a nice time!

Peter Gustafsson

[PeterGustafsson]

Hi!

If you use strictly single touch bouts you should be able to do a 18-19 person pool. I think the following will work. I have a sneaky suspicion that unexpected math might give a team with a larger or smaller membership an advantage.

Best team has highest wins/team member average (bouts against same team members will cancel out this is where I think some sneaky math might happen).

Best individual has the highest number of wins. Fence off for ties, which are disregarded in the team standings.

-Shlep'
(PS I did a quick simulation on Excel and it looks like it will work)

The two bigger teams will have win/bout ratios closer to 0.5, since they would have so many intra-team results to count, which give them both a loss and a win. The obvious solution to this is to only count bouts in which non-team members are paired in the team event.

A good fencer in one of the larger teams would have a great advantage in the individual event under your format, since he has more teammates - possibly his underlings in his day job - to get thrown bouts from. That can be counteracted, but it would require quite some number crunching.


Have a nice time!

Peter Gustafsson

[PeterGustafsson]

Hi!


Here is what I floated to my clubmates yesterday. What problems, if any do you see?
------
1. All fencers are given random ranking within their team.
2. Each team is given one slot per fencer in a 20-man initial random ranking. The slots are chosen so that the average ranking of all teams are as close to eachother as is possible.
3. Each fencer is put into one of his team slots.
4. We now have a balanced random starting ranking. All this can and should be done before the competition starts.
5. The fencer in slot#20 chooses an opponent. This opponent may not be teammate. They fence, according to the hydrid rules listed in a previous post,on piste 1. If the fencer in slot#20 wins, they switch slots. Otherwise, they retain their slots. They have 90 seconds running time, or first to 5. With 30 seconds preparation, that would be 2 minutes per bout.
6. The lowest ranked fencer not fencing on piste 1 chooses an opponent, and fences him on piste 2. Other rules as for piste 1.
7. All fencers have their losses and wins recorded on their slot card. Weak fencers sink to the bottom, better fencers rise. As soon as a fencer has accumulated 3 losses they may not challenge an opponent anymore. (They might be challenged from below, but that would be an unusual situation.) No fencer can under any circumstance challenge an opponent below him in the then-standing ranking, nor can he challenge a fencer currently fencing on the other piste. This way, the piste are decoupled from each other, and a holdup on one will not affect what happens on the other. Fencers can only challenge a teammate if they have bouted against all still active non-teammates, that will only come into play during the end of the competition.
8. As soon as a fencer has at least 3 losses and has no fencers under him in the ranking with fewer losses, he is done for the day and retains his current as his final placement.
9. Once only 4 fencers are left, they fence a traditional L4 with electric epees to 5, 3 minutes of effective time.
10. For the team event, all bout results between two given teams are recycled. Intra-team bouts are discounted. The team whose members record more bout wins than losses against a given opposing team wins, as a team against that team. With only three teams, one can have a sitiation in which they all win one team bout. Then individual bout win-loss differential is used as tiebreaker. If still tied, they have one deciding bout, with both bout participants chosen by the opposing team captains.

This format limits the traffic, since fencers will tend to stay at the piste once they lose. This cuts down on time spent moving and unhooking, should we decide to go electric. It also builds momentum, as the "better" fencers get moving in the end. It guarantees that all fencers will have at least 3 bouts. It requires very little counting, only count to 3 until the team results are to be calculated. Since fencers choose their own opponents, they can not blame me for uneven pools (that has happened before!), nor can they drop bouts to help teammates until the very end.

If each fencer needs 3 bouts to get eliminated, and each bout takes 2 minutes, and the first stage means reducing a field of 20 to one of 4, the total amount of piste time used for the first stage is 3#2#(20-4)=96 piste minutes. Given that I have 180 piste-minutes available, that gives me a reassuring time left for the semifinal stage, and several messups.


Have a nice time!

Peter Gustafsson

[PeterGustafsson]

Hi!


The event has taken place, and we used the challenge format that I described in the former post, with some really minor changes. That worked well, the company employees grasped the system right away. Due to illnesses and a pregnancy, they were only 15 of them, not 19. Reducing the 15 to 4 took 80 minutes on two pistes, for a total of 33 bouts. That gives an average time per bout of 4.8 minutes, but we did not use stopwatches for the bouts and challenger choice took somewhat longer time than expected.


Have a nice time!

Peter Gustafsson