I would agree for the most part. I would tone down on the 100% non-random bit. That’s not really true. Take a look at your W-W-W-W-W-L-W (called team A below) v. your W-W-W-W-W-W-L (called team B below) example. Below, team C is the undisputed best team, and team D is definitely #4. These four teams go W-W-W-W-W and have two matches left. They are randomly paired against each other. Team A gets randomly assigned Team C and loses. Team B gets randomly assigned Team D and wins. Team A then beats team D and Team B loses to Team C. Had the random draw been the other way, Team A would have beaten Team D before losing to Team C, and Team B would go the other way. So why is Team B ahead of Team C as you note above? Random draw.
Again, I pretty much agree. I’m not a fan of how things are distributed as statistically it is not designed to get all the robots to the top, and I feel the better robots should get further in a good selection process. It’s just that this system isn’t as perfect as you seem to imply with the 100% bit.
I would agree for the most part. I would tone down on the 100% non-random bit. That’s not really true. Take a look at your W-W-W-W-W-L-W (called team A below) v. your W-W-W-W-W-W-L (called team B below) example. Below, team C is the undisputed best team, and team D is definitely #4. These four teams go W-W-W-W-W and have two matches left. They are randomly paired against each other. Team A gets randomly assigned Team C and loses. Team B gets randomly assigned Team D and wins. Team A then beats team D and Team B loses to Team C. Had the random draw been the other way, Team A would have beaten Team D before losing to Team C, and Team B would go the other way. So why is Team B ahead of Team C as you note above? Random draw.
Again, I pretty much agree. I’m not a fan of how things are distributed as statistically it is not designed to get all the robots to the top, and I feel the better robots should get further in a good selection process. It’s just that this system isn’t as perfect as you seem to imply with the 100% bit.
A 1v1 would be difficult to implement in a game that is already established for 2v2 and would double the amount of matches that need to be played. This method with 2v2 would be better. Also, the changes to alliance selection and elimination matches are completely unnecessary.
Well, you claim that your 1v1 Swiss method reduces the number of matches that need to be played, but the way I see it is that if you make the matches 1v1 instead of 2v2, you will have to play twice as many matches. Let’s assume that there are 28 teams at a tournament with each team playing 6 matches. That means with the current system, 42 matches will be played. If you use the Swiss method in the same setup, 84 matches will be played. Even if you play 5 matches instead of 6, you still have to play 70 matches. Do you see the problem? Besides this, I think that it would be cool to implement, as the current system does have its flaws.
Edit: Just realized that @Bryan R already mentioned this problem in an earlier post.
I don’t like the idea of 1v1. VEX is inherently 2v2, and we’ve always been able to plan around this. This means that defensive robots are viable in qualification matches, although less so than elimination matches. In a 1v1 structure, they are not.
Interesting ideas. It would be fun to see a scrimmage try the 1v1 or 2v2 system you suggested. Although the ranking discrepancy has existed every year, I think Starstruck exacerbated the discrepancy with the zero-sum scoring. It is yet to be seen how things will go with In the Zone.
With that kind of disparity, it seems pretty random still. I would prefer to pair a 4-3 team and a 3-4 team against a 4-3 team and a 3-4 team than 7-0 0-7 with 7-0 0-7.
I would agree that keeping it 2v2 is important as that’s what it’s been. I would also agree with @Aponthis that pairings would be best made between comparable teams.
This is a very interesting post and a good thought for anyone who feels failed by the current qualification system, I agree with the sentiment entirely but there are a few flaws.
Do you want to crown the best alliance?
The best single team?
Or the best two teams and their choice of 2nd pick?
I feel like the data used in this post to justify the need for an alternative system is slightly biased. Starstruck at high level was more luck based than previous games were, and small mistakes at critical points could easily swing the match in one direction. Analysis of further games is needed to fully justify this. Just because in SS, Arts 4 beat 1 in their semi-final it doesn’t mean Arts 4 had the best robot/ team and Arts 1 did not for that division. It only would be if both teams had played flawless games, which is nigh impossible. Also Starstruck had an over reliance on alliance partners, I.E: if something went wrong it was very hard to carry your partner, this contrasts with previous games where it was easier to carry bad alliance partners.
The way Swiss pairings are generated mean that it is very hard to plan ahead, and combined with the accurate ranks at the end, Swiss places a lower emphasis on good scouting. Which in my opinion should be one of the most important bits of a tournament.
While I really like the Swiss system, after seeing the implementation of it in some TCG’s like Magic among others. There are a few things that I think may not transfer well over to the VEX environment.
Swiss is really good with establishing pairings of similarly ranked groups, but only AFTER round 1. In TCG’s you see players get into a bit of a rut if they lose round one and get forced into the so called losers bracket. Which is sometimes harder for them than if they had won the first round, and had to go against the better players.
So Swiss places an increased emphasis on round 1, which in VEX doesn’t benefit teams who usually experience small issues early in competitions and only fully run at their peak performance until mid to late in the competition. While you can argue these aren’t the best teams, and all issues should be ironed out for crucial events. You frequently see the best teams experiencing this early on and get recognized by the top seeds and picked early in alliance selection.
In TCG’s you frequently see players “dropping” half way through a tournament because there is no point in continuing on with the record they have, it is completely impossible for them to make it into the elimination bracket. While VEX has alliance selection, if there are Swiss rankings surely only the top 24 teams will be chosen? So teams may conclude like TCG players there is no point in trying following their 3rd or 4th game at large tournaments. Do we really want to see this happen in VRC?
I still think Swiss still has plenty of merits over the current system. But Swiss in VRC would be by no means perfect, however I would be keen to see someone hold a fairly large unofficial tournament and try to use Swiss for qualifications.
This argument is a recurring one. In every recurrence, the value of a new team randomly pairing with an experienced team is ignored. This Swiss system intends to arrive at separating the worst and best teams. It is a great concept for a friendly competitive environment where all participants want to have a match at their level. It is a poor concept in an environment where the principle goal is educational… elevating the new teams, expanding the number of people participating and learning.
From a practical standpoint, what does a decent local event look like under a Swiss system? Unless you exclude teams from the outset, ~50% of the field is new, ~40% has 2-3 years experience, and ~10% are 4+ year veterans. The upper end may not have enough participants to really fill out a Swiss concept unless you repeatedly group the top 4 to 6 teams in different combinations.
And +1 to starstruck being an outlier in imbalanced alliances being a problem. Starstruck was zero sum with a small number of playing objects and minimal opponent interaction. At even a low level of play, you had to rely on your partner more than other vex games. In years past, is was very possible to win 1v2, and that is likely the case again this year.
I agree with Doug on the educational standpoint. From my perspective here in Tennessee, it seems to be that it’s one thing to be in the top 8 for one tournament; it’s another thing to be in the top 8 at pretty much every tournament that you go to. I think that is the real measure of skill: how consistently a team is getting into the top 8 (or top 4, for that matter) at the varying tournaments the team goes to. With the current system, your robot should be built to hold up against the wide range of possible partner and opponent robot combinations. There are teams that definitely can do that, including here in Tennessee. I also like the current system for a related reason; it’s quite forgiving on new teams, which is what we should be focusing on, given the educational aspect of VEX robotics. Of course it still should mostly be the best of the best competing against each other at a tournament. My view of how the competition ladder should work is that while there should be mostly the best competing against each other for top dog(s), there should be an aspect where high-level teams give new teams a sip of being a high-level teams, and let the new teams that get a sip know that if they want more where that came from, they have up their game. This sip should happen through high-level teams “riding along” a new teams to Worlds, competing against new teams, etc. Competitions should create incentives for new teams to improve while still set up so that teams are generally ranked according to ability. That, I believe, helps cancel out the discrepancies of the current system and makes competitions fair for every one. Why you think FIRST has the Rookie All-Star Award, wild card qualifications, and the championship qualification lottery for unqualified teams?
So, STEM concepts are learned in the classroom during brainstorming, building, programming, etc. Competition is primarily about teaching “teamwork, leadership, communications, and more.” IMO, 1v1 is incompatible with these goals, but any kind of 2v2 (including Swiss) would suffice. Regardless of the match schedule, you still have to work together with your alliance partner.
Regarding new teams, it’s a mixed bag either way. Under the current system, sometimes weaker teams get carried by stronger teams, but other times they get repeatedly crushed by stronger teams because of an unlucky match schedule. Under a Swiss system, weaker teams might win more often with more balanced competition, but they may not get carried as often.
Maybe the current system could be used at local events, and a Swiss system could be used for state/national/world championships. Most teams participating at a championship are likely to be at least somewhat experienced (they qualified!), and the stakes are higher because of qualification spots.
We’ve been wanting a better system for getting rankings for alliance selection since forever, and the only change recf/vex has made (at least that I’m aware of) is one that intentionally made the rankings less reflective of who has the best robots and instead (much like the rules update) promoted the way recf/vex wants the game to be played.
What I’m not sold on is your definition of “best robot”. Without doing a system like you described you don’t know factually who is the “best” and any observation or name dropping It’s just like, your opinion.
I like the Swiss system and it’s something @Robo_Eng_13 has proposed in the past. I think it would not only give the best teams more accurate rankings it will also give the newer less seasoned teams a better experience. We need a system where teams are ranked and competing with other teams of similar development. When we have tournaments with participants ranging from hyper-competitive to “happy to move in tele-op” we need to strike some sort of balance in the Q matches. Those Hyper-competitive teams need more than the one or two hard matches that they will play all day (usually in the finals or semis) so that they can practice for what matches will feel like at a higher level, and those new teams that sort of resemble a complete robot need a chance to compete with other newer teams and maybe not get curb stomped every match.
As someone that’s been in the competitive side of this and now on the educating side of it, this is what I want to get from a system like this. Just generally a more positive experience for all evolved. The good teams will still do well and perhaps the “best” robot will finally be vindicated (/s), but what I really want to see is a balance struck where everyone for the most part is competing with their equals. That’s where I think your argument falls short, if you want to see a change, the argument needs to be made on the foundation that it will provide a more educational experience. As an adult and a coach I don’t care if that team that the internet decided should win lost in the quarter finals for some obscure reason, at the end of the day I want to know that the people that paid money and traveled across the state to be in the building each learned something that will make them better, and that’s what I think a swiss system would provide.
What I’d like to see is the freedom for EP’s to implement this type of thing freely and sort of test it out. If this system is truly better then it would then become the norm naturally. If it’s not, then we haven’t introduced Chaos into our Regionals and World championship.
The issue with the Swiss system is that the match schedule is dynamic for the qualifications rounds. I have spoken with people from the RECF about this in the past and they said this is something they would never do (dynamic match schedule). It is too hard on the event partner and the teams.
Propose something that doesn’t affect the match schedule and you might have a chance.
TRSP (opponents wp-teammates wp) as a way to judge the strength of your opponents vs the strength of your alliance member would be an excellent replacement for sp.
The other big thing is doing local events with 8 qualification matches per team. Each match goes a long way towards sorting out the randomness.
I think you are not giving developing (“bad”) teams credit for understanding the nature of their wins. Of course they know they were carried. Of course they aspire to build a better robot. Of course they learned that they need to get better. They, too, can learn about 3 cube needles - don’t sell them short.
I really dislike the “good” team and “bad” team distinctions. I have seen teams that have robots of questionable build quality make it the distance with a combination of luck, being good drivers, or having done thorough scouting. In my opinion if a team wins a tournament they aren’t a bad team they’ve obviously done something right to get them to that spot.
I’ve heard a lot of students assume that the skills robot is the natural champion and that any other opinion is logically incorrect for reasons, and that robot couldn’t win a tournament to save its life. Tournament wins are anything but an exact science and getting on the podium takes a lot more than a good robot.
I’ve never seen a team win a competition and just sit on their hands, we paid very close attention this year and the teams that were winning were constantly innovating this assumption is a pretty weak statement to base an argument on.
I’ve never seen a robot win a tournament purely on luck, I did note it was a combination of factors that lead teams to win.
This is thinking only in black and white, being too “absolute”. In Indiana one of the most respected teams in the state lost about 90% of the Q matches at state, ranked 60ish got picked first by the first seed team and made for tough competition in the finals. Had you just taken a snap shot of their day without taking the rest of the season into count by your definition they would have been a “bad” team.
With that said you can’t just take data from a season as an indicator for success either. we had a team qualify for state under the 1483A plate, we signed them up and then that team dissolved and the 1483A robot showed up to state as a proto bot. On the flip side the B team that was completely destroyed in the first three competitions showed up mid season under new management and won the state tournament. So no I don’t agree with your assessment. The very best teams still make mistakes and the seemingly awful teams will often times surprise you, you are right that there can be a clear pattern, but just assuming your theory is law is a unscientific approach to this scenario.
Don’t get me wrong scouting for eliminations is extremely important, the 1st and second picks can make or break an alliance. But if you don’t scout well in qualifications you may not actually get in a good position to make that dream alliance. To get into that position you need to be lucky, or have a plan. Sure if you are a supremely confident top tier team with a robot that can max out the game by itself in 1:45, then sure you can go for the strategy of turning up to the field, and telling your partner to just stay out of your way and you’ll take care of that match. But even so you may encounter problems.
Ideally you should before a match scout out teams you’re against and then your partner. To win you need a plan. If you’re formulating that plan on the queuing tables or on the field then you are probably doing something wrong.
Sure most of the time your plan may just be something intuitive and simple like: “throw the stars over”. But if you count on doing that every game and turning up to the field with an alliance partner who is on the same page as you then you will be caught out on occasions. A mediocre alliance who works together and plans ahead could have a good chance of beating a top team who is just winging it, and has no idea of their alliance/ opposition’s strengths and weaknesses.
When at a competition you should be scouting throughout and planning for every match. To plan you need to scout well for qualifications, and of course this can lead into who you pick for alliance selection.
Testing would be needed of course. However I could see many teething issues for EP’s and Volunteers running a Swiss event. You give teams very little opportunity to plan with short notice of rounds and it would be a nightmare for judges to talk to teams/ interview for awards.
I’ve never seen a team to win a tournament, or for that matter go to one and not think there is something they could do better. No matter how good people are they will always find something the could have done better, most people will choose to rectify that. Even the lucky teams @Coffee talks about will know that just because they were good enough to win this time, everyone else will be improving. So they know that they need to, too.
