Qualification rankings

I’d like to propose a modification (addition) to the qualification round process. I recommend all undefeated teams have a 1 match playoff (head to head). SP can still be used to seed the teams. This would make your first place team the only undefeated team at the end of qualification.

For example. At our last match there were 3 undefeated teams at the end of qualifications. 2 would then play 3 (head to head)and the winner play 1 seed.

Any support for this?

I don’t really see a need for it since it will just take time and there really isn’t a point to doing 1v1’s. Besides SP’s are good decider for tie breaking because it shows the skill of your opponents. For example at the competition i was at over the weekend we had 4 teams that were 5-1 but my team had 206 SP’s because we had an easy schedule versus a sister team that had the same record we did but had over 400 SP’s because they had a tough schedule in comparison to ours.

The only change I’d like to see is not calculating SP until after the qual rounds are complete and base it on the opponents and allies records. I really dislike the attempted manipulation of SP by scoring for the other team.

Since you have no control over your competition, I don’t see how it makes sense to penalize one team for randomly drawing weaker competition. Yes, if you beat better teams you get the advantage of seeding first in the sudden death playoffs and you only have to win one match. Others may have to win at least 2. It seems like a better way to be sure the 1st place finish goes to the best team.

Agreed…I can’t think of another competition where scoring for the other team is rewarded.

I don’t think that this is necessary. it would add a large amount of time to each competition and it would make my life a lot harder. I run the tournament software of some competitions in my area and it would be adding more features and steps to a program that would make running the events harder. When a event is setup you add your cycle time and your total run time for each planned segment. If you also had to plan for a round robin of teams that had no losses it would make the process that much more difficult. In short i think that the way VEX does it now is not perfect but i don’t think its that bad.

I’ve deleted a post from this thread for being pointless.

1 Like

Practically speaking this would be pretty easy and it wouldn’t take much time at an event.

But 1 vs 1 NbN is not the same game as NbN. Is it fair to get teams to play 1 vs 1 in the rare cases that they are tied for first? Is it something we want? It’s a much better system than making that decision based on SP (SP is just terrible) but I don’t think it’s the best system.

What I would do is use average score. That system doesn’t measure how difficult your schedule was, but the idea that SP does that is silly, and it’s not a required feature of a ranking system. The idea is to rank teams from best to about 15th; that doesn’t necessarily require tiebreakers that are based on schedule difficulty.

The drawback to using score as a tiebreaker is that it rewards a certain type of robot, i.e. one that scores a lot by itself and doesn’t require a strong partner to do well. I don’t see that as a problem, because those are the only kind of robots that can rank high in qualifications anyway. Tiebreakers between high ranking teams are the only tiebreakers that matter, because captain rankings are the only rankings that matter (And among captain rankings, it’s pretty much only the high seeded captains. More than 90% of events are won by one of the top 4 alliances). So using score as a tiebreaker just increases the resolution of the WP ranking, which is what a tiebreaker is intended to do - unlike SP, which rewards players for playing this weird subgame that always needs to be explained to spectators and that many people (including me) dislike.

If using average score, would that be your alliance’s average score or the overall average for the match? If it is the alliance average score than, potentially, that has the opposite effect than SPs. Two very good robots facing two weak robots have the opportunity to score much higher than when facing equally as strong robots. This is due to the strong alliance having the opportunity to collect and score more balls versus a weak alliance. Therefore average alliance score potentially rewards those who face weak competition. As for overall match average, that might be somewhat better as two strong alliances facing each other might score more combined points than a strong alliance facing a weak alliance. The difficulty then becomes the randomness of the pairings. A very strong undefeated robot may have been paired with average robots and lucky enough to have just won all of their matches, but not scoring a lot (as an alliance). Yet that robot could very well be the best robot at the event. I admit I am not terribly comfortable with playing the SP “game” either, but I have yet to see a reasonable solution as long as the events are not true round-robins.
That is my 2-cents worth, and I am sticking to it.

The best teams will win their match but not by much while still maintaining high scores. It’s really simple just score more than everyone else and it won’t matter. If everyone understands that SP’s are part of the game then the system works. It doesn’t work when good 1 loss teams only score for themselves and end up with half the SP’s as the rest of the 1 loss robots and because of that are ranked 10th instead of 4th. Which is why scouting is important because the 10th place robot might be the best pick for the 1st, 2nd or 3rd place alliance.

The biggest problem I have with rankings is when tournament finalists can qualify for something. There have been several cases where our quarterfinal is the hardest and the final is the easiest.
I guess a better system may be CCWM for tiebreakers in the highest WPs. That would be really dependent on the schedule, though… Maybe OPR…

This is very true; the system is biased no matter what because of alliances. However, IMO, it takes away from the game to force teams to score points for the opposing teams. First off, it kills the “high score” of a great team if they are scoring points for the opposition. It also seems to be rubbing the loss in the other team’s face: Would you consider it to be disrespectful if half of a soccer team decided to play for their opponents because they were so far ahead? Finally, SP’s force teams to devote mechanisms to scoring SP’s rather than winning. For instance, at a competition in October that my friend went to, the two tournament champions were from the same school and both had linear puncher launchers. Since they could only fire from the base tile, they each build tilting mechanisms that allowed them to shoot preloads into the opposing alliance’s net from the starting tile. In other words, these two teams were forced to waste a motor to score points for the other team! Would the competition be more enjoyable for these teams (who deserve their victories because their designs were superior to the other designs) if they were not forced to waste resources to build a mechanism that will be pointless in the finals? I think that’s a safe yes.

Overall match score will still result in teams scoring for opponents, but not as often. SP measures roughly the same thing as overall score anyway, because you’re trying to maximise losing score while keeping the winning margin constant. In practice the two systems should have much the same effect. Where you say:

that robot would also have low SP, because the team’s own score puts an upper limit on their SP score.

So I definitely think overall score is preferable to SP, because it would remove a lot of the unnecessary confusion.

Comparing overall score to alliance score is a bit tougher. For a game without a big defense or descoring aspect (such as Skyrise or NbN) overall score is probably better since it correlates both with being a stronger team and with having a harder schedule. The occasional team scoring for their opponents under that system is definitely less of a problem than having some teams do it in every single match.

It’s the games where descoring and defense are important where it’s difficult to come up with a system that works well, because there isn’t much difference between the field after two strong alliances have played each other and the field after two weak alliances have played each other. Both SP and overall score reward weak teams with hard schedules and strong teams with easy schedules, while disadvantaging strong teams with hard schedules. That’s pretty far from ideal. But alliance score also rewards strong teams with easy schedules more than strong teams with hard schedules, so it’s a question of how big the effect is in each case. The only way to work that out is to look at the historical record, which will take a while to do.

Robot skills score is sometimes proposed as a tiebreaker instead of SP, and after thinking about this some more I’m quite sympathetic to that idea. It’s schedule-independent rather than rewarding easy schedules like the other metrics sometimes do.

How do strength points work?

SP = strength of schedule points

In the ideal world, every team will be giving their best to try to win the match by as big a margin as possible.
That means, the points scored by the losing alliance will be an indication of how tough the schedule is.

But this ideal will only hold true if the winning alliance is not farming for SP, or scoring for their opponents.

I agree that total points would produce a more offensive game. If the goal is to out score your opponent, the rankings should reflect how well you did this. Sp rewards something else in many cases.

I haven’t thought through this completely, but I really like the idea of this. Is there any downside to using overall score? It seems like it would do everything SP does, without the weirdness.