Thoughts on CoOpertition

[Chris is me]

Does everyone but me disagree with the second premise? I disagree with pre-arranged matches, but I thought "6v0 or not?" was pretty interesting. I guess I don't see "why" it's automatically undesirable for the loser to score for the winner but the reverse is okay.

[Leav]

Originally Posted by Chris is me

Does everyone but me disagree with the second premise? I disagree with pre-arranged matches, but I thought "6v0 or not?" was pretty interesting. I guess I don't see "why" it's automatically undesirable for the loser to score for the winner but the reverse is okay.

I see both as undesirable.

I neglected to add that in my opinion coopertition is achieving the opposite of it's goal since having the winner scoring for the loser is much more insulting than the blowout scenario.

-Leav

[Chris Hibner]

I'm going to make some arguments here in support of a coopertition seeding method. Please note that I don't necessarily agree with these arguments, I'm just presenting them here.

1) A Coopertition ranking system builds strength-of-schedule into the ranking. If team A wins 15-3 and team B wins 15-12, then team B's win should be rewarded more since their opponent's were most likely better. All wins are not created equal, and the coopertition method helps incorporate this into the rankings.

2) The Coopertition schedule allows better robots to rise to the top of the rankings. Let's say alliance A (a good alliance) plays alliance C, and alliance B (a GREAT world-beater alliance) also plays alliance C. Alliance A wins 15-4, which is a decent qualifying score. Alliance B is so good that by the time they score 15 points (same as alliance A could score), there is enough time left in the match to score 6 balls for alliance C, which significantly raises their qualifying score. If Alliance C is is good enough to do this, they should be handsomely rewarded with a good qualifying score.

3) The Coorpertition system encourages offensively designed robots. Many people in FIRST would rather see the game played with a offense/defense split around 80%/20%. The Coopertition system rewards high scores, so teams are more likely to develop offensive robots.

[David Brinza]

Does awarding the losing alliance the winner's score (without penalties) really make sense??

Somehow, having the losing alliance earn more seeding points than the winning alliance in a match just doesn't seem right. This situation was mostly eliminated in Team Update #16 by awarding a 5-point bonus for win (but not always).

[Jared Russell]

Originally Posted by David Brinza

Does awarding the losing alliance the winner's score (without penalties) really make sense??

I think so.

With the exception of Overdrive, every FIRST game has a finite number of game pieces or scoring locations (Triple Play/Rack N Roll) that the alliances are competing for in order to score for their alliance. If I have a good scoring alliance but I'm up against a great scoring alliance, my score will be lower than if I were to play a weak alliance simply because my opponent will have scored more balls/goals/spider legs/whatever for himself, leaving fewer for me. So a strong offensive robot with a nightmare schedule would seed lower than the robot's true ability would indicate.

[M. Mellott]

My main issue with Coopertition is that, similar to Stack Attack in '03, we were playing two different games: one for qualification matches, one for elimination matches. Some will say this makes strategizing more interesting, but some want to see the competition between the alliances, especially the spectators. Let's have just one set of rules...they're hard enough to follow as it is.

As for the inclusion of "strength-of-schedule" into the scoring system, this would be a good idea...if the schedules weren't completely random (or as random as an "algorithm" can make them). Sure, because of the scoring system, I'd love to play against a 469 or an 1114 every match because I know I'll get a lot of points (mostly from their efforts), but that doesn't happen. Sometimes, good teams draw the short straw and get less-than-desirable matches. In this game, it could happen with a bad alliance or with a bad opposing alliance.

Also, I really wasn't too fond of an award created for "earning" the most Coopertition points.

[Tom Line]

The Coopertition system this year was not solely there to reduce blowout scores.

Remember: FIRST is aiming to make this a game that is easier to watch. In that vein, shiny flashy offense and slam dunks are a whole lot more interesting than well-coached defensive teams.

The scoring system encouraged teams to SCORE. Either for themselves, or for the other team.

The really, really shocking part of that is just how many teams seemed to have absolutely no idea how to play the game. Even at championships there was a stunning number of teams that were STILL playing defense. It hurt both their own seeding score, their teammate's, and the other alliance. In point of fact, very few of the third picks I saw at the championships were picked for defensive reasons. They were picked as another scoring robot to complement the other teams.

I guess I've slid more into game strategy than coopertition. I'll get back on track.

In 2008, FIRST made all sorts of rules to try to promote scoring. It was a total nightmare. This year they got a bit more clever: rather than trying to punish 'bad' behavior, they tried to reward the good behavior.

It mostly worked.

There is a very simple very easy way to get rid of the 6v0 issues we saw this year. Have the losing team get THEIR score, and not the winning team's score. Now, the losing team has a reason to score for themselves. The winning team, if it's a blowout, still has motivation to help the other team.

That's my solution.

I can handle playing two different games. Professional teams do it all the time, depending on their opponent.

[Leav]

Originally Posted by Tom Line

The Coopertition system this year was not solely there to reduce blowout scores.

...

The scoring system encouraged teams to SCORE. Either for themselves, or for the other team.

...

I can handle playing two different games. Professional teams do it all the time, depending on their opponent.

This seems reasonable, if you want to create an intensive high scoring game where defense is not always a good option than I agree the seeding system this year is a good idea. but in my opinion it must be supplemented by a rule penalizing self-goals.

also in order to promote cooperation on the competition field I propose that the game be designed with an appropriate challenge, for example (in addition to those proposed earlier) here are some changes to previous game which would serve this purpose well:

1. 2007 seeding score bonus for a red-bot on a blue-bot at end of match (or visa versa ofcourse)
2. 2006: points awarded to both alliances according to minimum of robots on ramp at end-of-match

once you get the basic idea rolling, it is easy to supplament each year's game with such a scoring bonus which would have opposing teams cooperate in order to maximize their score.

-Leav

[AdamHeard]

Even if you disagreed with the removal of win-loss, you do have to admit top performing robots are much more likely to seed. Loosely looking over some events, the only outliers are top robots that weren't reliable in many matches, which explains why they were seeded lower.

I think the way to improve it for both the winners and losers is to give the losers their score. This more or less turns the ranking system into a function of your average score plus a bonus (of a constant plus the losers) if you win. I like that, you get rewarded for high score, regardless of win/loss, and you get rewarded for wins, regardless of score.

[Koko Ed]

I completely HATE the coopertition seeding system even if FIRST did "improve" it in mid season. It didn't stop the 6 vs. 0 junk.
I want the teams to play each other not the system.

[cyberjoek]

The fix, to me at least, is a somewhat simple one. Each team gets their *own* penalized score, winning team gets their opponent's un-penalized score as a coopertition bonus. I do like the idea of a coopertition game piece being built into the game and being worth lots of bonus points to both teams if everyone does what they need to.

If each team gets their own score then it encourages you to score for your team (every point scored goes into your seeding) and encourages you to keep the game close (if you win you also get your opponent's score). If FIRST doesn't want blowouts to be as harsh to seeding perhaps there's a floor of 1/2 the winning alliance's score that the losing team goes away with as seeding points.

[Steve Compton]

So much has been written, well thought out, pro and con, trying to parse GDC intent, effectiveness and outcomes. My thoughts are simple and clear to me, and there are only two:

1. If my alliance wins 19-0, my team seeds less well than if we lose 25-0. At some fundamental level, a system that allows that is grossly defective.

2. Because 2 different games are played in qualifications and eliminations, teams that are solid prospects as a strong elimination robot (for instance a great defender) has to intentionally reduce their demonstration of effectiveness in qualifications and go for high seeding points according to whatever strange scoring strategy makes sense at the moment.

I still believe than winning matches needs to matter and that we each design our robots for specified functions, and that a ranking system that evolves to devalue robust implemntation of conscious design is compromised.

[Bill_B]

Originally Posted by Steve Compton

So much has been written, well thought out, pro and con, trying to parse GDC intent, effectiveness and outcomes. My thoughts are simple and clear to me, and there are only two:

1. If my alliance wins 19-0, my team seeds less well than if we lose 25-0. At some fundamental level, a system that allows that is grossly defective.

In both cases the winning alliance has FAILED to help the other alliance have a better showing in those matches. For that failure they are not awarded seeding points purely for their victories. Most of the objectors to the coopertition seeding structure have focused almost solely on the robot performance on the qualification fields. If, before that 19-0 win, some part of the alliance had helped (cooperated to allow) the other one score better, their resulting seeding would be improved, right? That help can be in the form of strategy advice or mechanical design improvement, or both. Argue the impracticality or even impossibility of that level of help, BUT that is closer to the spirit FIRST is trying to incubate, nurture and harvest from the various teams involved.

Instead of picking apart this year's seeding method, I'd rather see some thought devoted to a way to award seeding status to those alliances or teams that can be verified as having positively helped their opponents during qualifications. If everyone in the pits opens their crates and shifts directly into eliminations mindset, we're not so far from the battlebots label being assigned to us by a public that doesn't know better. If a potential alliance partner doubts your ability to play a rock'em sock'em game, just invite them to a personal demo on the practice field to prove your "robothood." Or offer a video showing your robot's abilities. I saw a lot of flat-screen displays in the pits; plenty of opportunity for mechanical boasting via electronics. The need for demonstration during qualifications will not suffer for being moderated by seeding considerations.

[Kibaspirit]

I want the teams to play each other not the system.

I totally agree. Point blank. Do other sports say "Hey, let's score some for the other team, in the long run it'll be beneficial right?"

No. I believe I remember Dean and Woodie talking about how they wanted to creat a sports-like game, to get more people to come to the competition. But, a lot of parents that went with our team were confused and frustrated with the 6V0-type system.

It is great that they added the 5 point bonus, but those 5 points are more obnoxious than helpful in my opinion. Wouldn't it be easier for it to be just the way it was?

You win, you win. You lose, you lose.

I know Coopertition is in the spirit of FIRST, trying to level the competition- but creates that whole idea of "playing the system."

It was risky to score goals for the other team, when you weren't sure if you were actually ahead. Sometimes, the score on the screen, wasn't actually the score... at all. How can you decide whether to score for the other team, when you can't even trust that?

I just see, another award like the safety award. How loud can you yell robot? How many goals can you score for the other team?

[Travis Hoffman]

Loser gets loser's score. No more 6v0, along with more of an incentive to win.

Winner still gets W + 2L. Get rid of the 5 point bonus, because the stuff you were trying to mitigate by including that bonus is no longer in the game.

Teams are only allowed to score in their own goals. No more 6v0, and no more 2*loser's score padding during a blowout. Winning teams aren't doing this to help and "cooperte" with the loser; the loser doesn't magically become more of a capable machine as a result of this practice; winning teams are doing this only to elevate themselves in the standings, which they will continue to do if the rules permit them to do it. Please get rid of any other tra la la "coopertative" notions that claim otherwise, and close the door on this pointless practice.

But...BUT...since the 2L bonus is still in place, if the better teams want to maximize their seed scores, forcing them to only score in their own goals ALSO incentivizes getting them to help their opponents become more capable BEHIND THE CURTAIN, where true "coopertition" has been demonstrated for years. The only way to get a big 2L bonus is to make sure your opponents are competitive BEFORE THE MATCH and leave them balls to score DURING THE MATCH. What a concept.

Finally, I really like the idea of a mutual "coopertition" objective built into future games - this sounds like the mutual mission that spans the two fields in FLL.