The Compromise Elimination System- Best of Both Worlds

As we all know, there’s been intense debate over bo1 recently. I’ve argued for bo3, but I’d like to consider all sides of the argument and elaborate on something Carter has already proposed. To do this, I’m going to look at the goals of competitors, spectators, event staff, EPs, and vex itself with regards to eliminations. Paragraphs are too long, so I’ll just list each group’s top few priorities below. Here goes :stuck_out_tongue:

-Large sample size that gives teams an opportunity to win despite bad referees or vex keys
-lower barrier to entrance for new teams- no requirement for 100% consistency
-low pressure environment for good driving and clear thinking
-reasonable scouting scouting requirements for small teams and no throwing- no third alliances

-no throwing- no third alliances
-shorter events

Event staff:
-innovative fail safes for the design interview and notebook
-robots that are easy to inspect and pass inspection first time
-low pressure environment with few referee arguments
-larger sample size so referees don’t take blame if they make a mistake and end someone’s season

-short events
-many teams in eliminations
-no throwing- no third alliances
-maintaining a good reputation- not constantly arguing against competitors on forums

Vex (more accurately the GDC):
-keep EPs happy to form more competitions: Dan Mantz said at the EP Summit that team growth is exceeding event growth
-maintain loyal and happy community that keeps the program alive

If you disagree with a goal or have something to add, I’d love to hear it! However, please only comment if you’re a member of that group. I don’t pretend to have perfect insight into what the GDC or random volunteers want, and they don’t have perfect insight into what I want. This will just help to minimize clutter. Thanks everyone!

Right, back to business! If vex implemented the following compromise, I believe every single one of these goals could be met, at least to a much larger extent than they are now.

The idea, as Carter has gone into great detail about, is a 12 alliance, 24 team bo1 double elimination bracket. For anyone unfamiliar with the bracket, I’m just going to try to summarize briefly. Seeds 1-4 get by’s to QFs, and 5 plays 12, 6 plays 11, etc for a spot in the QFs. Then, this half of the bracket continues normally with bo1 rules. Every team that loses a match is moved to the loser’s bracket, where they continue to play a normal bo1 style bracket. Any team that loses in this half of the bracket is eliminated from the tournament, and the winner of the loser’s bracket (the only team to only lose 1 match) then faces the winner of the winner’s bracket (the only undefeated team) in a grand finals match. The winner’s bracket team only needs to win once to win the tournament, but the team in the loser’s bracket needs to win twice in a row to win. This gives each team an opportunity to lose 1 match on bad luck without being eliminated from the tournament and maintains 24 teams in eliminations. It also means the elimination bracket is always 23 or 24 matches long.

Going down the list of priorities:

-large sample size: each team is allowed to lose twice before being eliminated, giving them more room for error and more of an ability to recover from problems outside of their control
-no 100% consistency required for new teams: having the security blanket of being allowed to lose once goes a long way towards boosting morale, and really helps smart but inexperienced teams that might not have fixed every consistency issue
-lower pressure: being allowed to lose twice helps a lot
-no third alliances: check
-shorter events: this 23-24 match bracket takes slightly longer than a 15 match bo1 bracket, but it’s much shorter than a 30-45 match bo3 16 team bracket. The extra 9 matches assuming 3 minutes per match add just 27 minutes to the competition. Whereas a 45-match tournament would add 30 matches and thus an hour and a half to the competition. 27 minutes is quite reasonable, an hour and a half isn’t.
-cool design features to show off in the interview: even though double elimination is more forgiving than single elimination, it’s less forgiving than bo3. With bo3, you can lose a match every single series, whereas with bo1 double elimination, you can only lose 2 overall. This means the robots will still need to be built for extreme consistency, and backups and fail safes will still be critical. Check out 974X’s NBN worlds bot out as an example. They could literally shoot with half of the flywheel broken! Talk about a cool interview!
-robots that pass inspection easily: a robot with elaborate fail safes is a lot less likely to fail a software check during inspection.
-lower pressure matches that don’t put strain on referees: obviously no referee has an easy job, but with a little extra insurance, fewer teams are going to argue as much or as intensely. It’s much less frustrating to be eliminated because of 2 losses than because of 1.
-larger sample size to protect referees: check
-more teams in eliminations: this is by far the biggest drawback of my proposed system. 24 is better than 16, but it’s worse than 32. That said, many competitions don’t even have 32 teams, and no competition has 32 good ones. At worlds, the vast majority of 13th-16th seeds didn’t even survive 1 round, and only 1 alliance out of 24 below 12th made it past div quarterfinals (in any division). The 16th seed captain in technology couldn’t stack cones. My sense is going down to 16 teams in elims would be a huge issue, but keeping it at 24 instead of jumping to 32 isn’t too horrific. 24 has been the norm for years, and to my knowledge it’s never been an issue.
-keeping EPs and competitors on the same page- the bo3 system was bad for EPs because it incentivised throwing matches, the new bo1 system is bad for competitors because they can be eliminated on bad luck and have so much extra pressure. With this compromise, there is no throwing and 1 loss elimination, and it takes basically the same amount of time as the other two systems. Hopefully this would allow competitors and EPs to see eye to eye and tarnish fewer reputations on these forums.
-creating extra competitions: what EP wouldn’t want to host a competition that could be run in a reasonable amount of time and with a body of excited, loyal students incentivized to play their best instead of throwing? This system seems to benefit EPs more than any other group, they get literally everything they want except 32 teams in elims. It’s a win win win!
-loyal and happy community: I can’t see anyone throwing a fit over this system. It gives teams a chance to lose once without being eliminated, it’s not too boring to watch, it takes a reasonable amount of time, it maintains the 24 team in elimination tradition, and it seems like a really reasonable compromise that meets everyone’s main goals.

It’s a little less forgiving than bo3, it takes a little longer than bo1, and it gets a few fewer teams involved than with 16 captains. But all of these drawbacks are minimal. It seems to capture the most important attributes of each elimination system with a couple of easily overlook-able hiccups. There’s not going to be a perfect solution because EPs, the GDC, volunteers, spectators, and competitors have such different goals, but this seems like a pretty fair way to balance everyone’s preferences in a compromise that meets each group’s fundamental goals.

“You can’t always get what you want, but if you try sometimes, you just might find you get what you need.”

I just have a couple comments to make about this. The system I originally posed was one in which the teams in the first round (seeds 5-12) do not get a spot in the losers bracket if they lose. The benefit of this is it rewards teams that seed high a bit more and shortens the event a bit. However, this comes at the cost of putting a bit more pressure on referees and doesn’t give those losers a chance to redeem themselves. I don’t particularly care which system is chosen if one of these were to be the replacement elim system though. My original format was the way it was because I thought the advantages it has over the one Anomaly laid out were the ones the GDC valued more.
Edit: both of these systems give the lower seeded teams a higher chance to win a match than the current system, which is in my opinion significantly more valuable than just getting a spot against a top seed, basically a guaranteed loss. I would much rather have a 50 percent chance at earning an award that getting a participation trophy.

This seems like a good idea, but it has a fundamental flaw: complexity.

As a former competitor, an event staff member, a spectator, and someone who’s worked extremely closely with an EP to host tournaments, I simply don’t see such a complex system working in practice across the whole VRC program.

  • New teams already generally don’t understand the classic 8-alliance, 24-team Bo3 eliminations. Under the proposed system, there would be chaos as event staff (not necessarily understanding the system themselves) would have to constantly tell these teams where to go.
  • From an event staff perspective, there have been times where I (understanding the system from years of competing) have found myself helping inexperienced queuing staff with classic eliminations, but the classic bracket is at least easy to get the hang of. With the proposed bracket, inexperienced queuers don’t stand a chance.
  • Following on the last point, EPs can no longer just assign random volunteers to queuing and train them the morning of the event. Queuers now have to be experienced, adding unnecessary, additional burden to EPs.
  • Lastly, spectators will just be confused if they try to figure out how eliminations work just by watching, but I guess that was largely true with the classic system too.


Yeah, that is an interesting point, the system is a little more complicated than a single elimination bracket. That said, the complexity issue could be totally fixed by just putting the bracket and the fields where each match will take place up on the projector. It’s a lot harder to be confused when there’s a road map right there.

About volunteers being confused, just about everyone that’s competed seriously in a sport or robotics knows what double elimination is, even if they haven’t usually use it. Just keep the bracket up on the projector and announce when each team needs to queu, and there shouldn’t be a problem.

I actually think spectators would be a lot less confused with this system than with the old bo3 system since third picks made very little sense. My grandma actually asked me in 8th grade why we had put the weaker robot in at all, and I had to explain to her the rules about third picks and everyone playing one match,

But overall complexity is definitely a concern and the GDC would need to implement some basic guidelines and practices to help everyone figure it out.

There are so many events (in my region at least) that literally only have 1 projector/screen for the whole event that keeping the bracket up simply isn’t possible (match timers have to be displayed somewhere).

Also, I think TM would need a single more informative display mode before this is reasonable (elim bracket with captions for field assignments and annotations like “Queue now” and “On field”).

I definitely do agree that complexity is an issue to contend with here. However, going away from three teams should help. People get confused when the second pick robots come in because they see the matches as being in a different series. This could make the tournament harder to follow. With two alliances, spectators and competitors begin to recognize different alliances and the bracket becomes easy to follow. It is unfortunate that some comps only have one screen, but the queue areas can still be announced and the bracket should be up as long as possible.

Double elimination sounds good, however teams could throw their opening match and out perform in the lower bracket.

Personally I like double elimination format, from playing DOTA 2 and csgo lans

My personal opinion is for signature events like US open and worlds to use a better elimination system. Since those are really the events that would mean the most to competitors

Is it ever worth throwing a match in double-elimination? Everyone who loses goes into that lower bracket. Losing early means you enter the lower bracket earlier and have a greater chance of being fully eliminated. As long as you keep winning, a single loss won’t knock you out so you have safety. Even if you think you’re one of the top two and are up against the other you believe to be in the top two, if you throw the match then in theory you’ve planned to absolutely beat them twice in a row in the finals. But if you can beat all the teams well enough you’re throwing a match against the other top team, then you might as well beat them now, keep some safety in case of a glitch, and only have to beat them once in the final. Maybe there may be some situations… So, just when would you want to throw a match?

Honestly, while I would prefer this style of tournament over straight bo1, I’m not sure I’d be fully on board with this. I have a pretty decent understanding of lower brackets now, but it’s really hard for newbies. I learned about double elimination brackets from an esport I used to watch, and it was not fun trying to figure it all out when I logged on to the stream for the first time. Imagine trying to learn this new system in a hectic day of competition with breakdowns and practicing and scouting. I’d prefer a 12 team bracket with byes for the top four teams and bo1 for the bottom 8, with bo3 going on from there. That would in a range of 18-25 matches, so it wouldn’t be the shortest of options, but it would be only four games longer than the old system. Essentially, it would lose the time advantage of bo1, but it would also maintain participating team size while eliminating the third team.

Personally, I agree with 12 alliances, bye for 1-4, single elim for 8-12 with Bo3 being used for QF, SF and F. I can compromise on Bo3 only being for the Finals.

It’s similar enough to what VEX U used to do for quite a while.

As a College Basketball fan, I like their system the best.

I hadn’t considered it much, but bo1 to get to QF and then bo3 in QF SF and Finals is super similar and is probably easier for everyone to follow.

And you can tell it’s a good system because we have a student and an EP agreeing!

@Paul Copioli and @DanMantz would this be a viable option? Have the GDC considered this? Are there any obvious reasons they’re not using it?

I like this system too. personally, I have four systems that are far better than the others. They all have their advantages and disadvantages, so I’ll go over what I think these systems are and how they compare to each other.
Similarities between all systems- These things all are better than than other options in my opinion, shown by all my favorite elim choices having these.
1- 12 alliances
2- 2 team alliances

These are ordered in the order I prefer them. Note that this is not the order I think the GDC would like one system over the others.

  1. 12 Alliance Double elimination- the bottom 8 seeds play a bo1 match against another team for a spot in the next round. Losers here will not get a spot in a losers bracket.

Pros- Fairly short, tight time schedule, competitive integrity, all teams have a chance to win at least one match, rewards good qual play, helps balance bracket
Cons- Difficult for new teams to grasp, 24 teams in elims

  1. 12 alliance bo1 to bo3- the bottom 8 seeds play a bo1 match to get into the next round. All the remaining matches are bo3.

Pros- Very good for competitive teams, all teams have a chance to win one match, rewards good qual play, easy to understand
Cons- looser time schedule, medium length bracket, 24 teams in elims

  1. 12 alliance Double Elimination 2- Standard 12 team double elimination bracket- look up online if you are not familiar

Pros- Tight time schedule, competitive integrity, all teams have a chance to win at least one match, rewards good qual play, helps balance bracket
Cons: Difficult for new teams to grasp, 24 teams in elims, medium length bracket

  1. 12 alliance bo1 to bo3 2- First two rounds of bracket are played with bo1, finals and semifinals are played with bo3.

Pros- Tight time schedule, fairly short, all teams have a chance to win at least one match, rewards good qual play, easy to understand
Cons- top 4 seeds play bo1, 24 teams in elims

  1. 16 team bo1 to bo3- First two rounds are played with bo1, bo3 in semis and finals.
    Note: This bracket is in the next tier behind the four I listed before it. In my opinion, this bracket is significantly worse than the top 4, but it isn’t terrible and definitely has improvements over the current system.

Pros: Fairly short, tight time schedule, easy to understand, 32 teams in elims
Cons: Top teams play multiple bo1 matches, lower teams are unlikely to win many matches

I like the idea of using a different elimination bracket. There are obviously issues with Bo1, but Bo3 also had issues for EP’s. Different elimination brackets might make everybody happy.

I agree. There are benefits to be gained from both BO1 and BO3, and a combined bracket could do it better. I like pretty much every idea here, each of them is a better reconciliation than all out BO1 or BO3. I do love the idea of a double elim bracket but I also agree that it would be difficult for many to understand. I think @Carter explain each method well and I agree with the pros and cons of each format

I think it’s worth noting that a pro of each of the brackets Carter listed was having two team alliances. While obviously writing this in the pro category does little to differentiate them from each other, it does well to show that bo3 and 2 team alliances are not mutually exclusive.

This is correct. I just didn’t put it because it would be redundant and it was listed as what each system had in common.

Oh my bad didn’t see that. Glad to see you pointed that out.

16 alliances, lower seed chooses bo1 or bo3. Most will choose bo1 for better chance of winning by luck, but matches between closely ranked alliances will probably be voted as bo3. Rounds that dont matter go fast, rounds that do matter go slower. As someone who has been first seed champions all last season (except worlds) as well as second pick of eighth seed many times years ago, I can tell you qf1-1/2 and qf3-1/2 are not worth two games time. This probably wont happen, but it’s still an interesting idea.

I don’t love this idea, it seems hard on EPs and it would be tough to figure out exactly how long the tournament would take.

How would these systems work if you only have 16 teams and one competition field at a tournament?