Picking from top 8 seeds

[nobrakes8]

Originally Posted by rick.oliver

While it may afford two other teams the opportunity to compete in the Final Elimination matches, it may also deny a team with a more competitive robot a spot in the Championship event.

This comment is an awesome perspective on the issue I didn't consider so props!

I'm still not a fan of super alliances after week 1 and if the teams involved have previously qualified for the championship (in regards to getting qualified for the championship and during the championship I'm all about survival of the fittest). I say this because if the two or three best teams compete against each other that brings in strategy into play instead of domination --and I feel like if you want to be the best you need to beat the best. Plus your alliance picks on those 1,2 and 3 seeds matter for strategy and all sorts of reasons not just to have them ride a great team's coattails.

[hiyou102]

I think that the 2nd pick is still a good way of making weaker alliances more effective. We still want a system that rewards teams who do well rather than putting a barrier that makes it so less good teams qualify for worlds.

[themccannman]

Curie division last year, 254 was having robot malfunctions through most of quals which ended up leaving them ranked in the bottom 30 teams. During their last couple qual matches they got their robot working again and started shooting the lights out scoring nearly 100 points each match but not going for cooperition points which did 2 things:

it allowed them to shoot for the entire match instead of balancing to demonstrate their offensive ability, and it caused teams they played against to lose coopertition points which increased the likely hood of 341 staying the #1 seed (which was the team they wanted to be picked by).

If you change the ranking system around this is much more likely to happen, if a team wants to be picked by the #1 seed they will throw matches so that they don't make the top 8 and can get picked. This system would discourage teams from playing well to rank highly and would not solve the problem of "super alliances" as the good robots who aren't #1 seed would purposefully rank out of the top 8 to get picked by the #1 seed.

The current system encourages teams to play their best throughout qualification matches except in some very rare cases (e.g. in the previous case if 254 had a match against 341 and wanted to keep 341 as the #1 seed). The other important fact to note is that the seeding is typically not accurate whatsoever at telling you which teams are the best teams, if it did everyone would ditch their scouting systems and just make picks based on seeding order.

[Tom Line]

Originally Posted by themccannman

Curie division last year, 254 was having robot malfunctions through most of quals which ended up leaving them ranked in the bottom 30 teams. During their last couple qual matches they got their robot working again and started shooting the lights out scoring nearly 100 points each match but not going for cooperition points which did 2 things:

it allowed them to shoot for the entire match instead of balancing to demonstrate their offensive ability, and it caused teams they played against to lose coopertition points which increased the likely hood of 341 staying the #1 seed (which was the team they wanted to be picked by).

Hmmm. That can be a fairly touchy accusation, because by implication it is suggested that they were costing their own alliance partners points as well. Do you have something concrete that would back that up?

I think a lot of the discussion here simply highlights why everyone should move to a district system if possible. Basing your attendence at worlds on one competition would be incredibly stressful. Instead, a system like Michigan's where a solid team just has to play well and not win a single event rewards consistent performers.

[Gregor]

Originally Posted by Tom Line

Hmmm. That can be a fairly touchy accusation, because by implication it is suggested that they were costing their own alliance partners points as well. Do you have something concrete that would back that up?

I think a lot of the discussion here simply highlights why everyone should move to a district system if possible. Basing your attendence at worlds on one competition would be incredibly stressful. Instead, a system like Michigan's where a solid team just has to play well and not win a single event rewards consistent performers.

If it were true, their alliance partners were free to go to the coop bridge. No harm no foul.

[Nuttyman54]

Originally Posted by Tom Line

Hmmm. That can be a fairly touchy accusation, because by implication it is suggested that they were costing their own alliance partners points as well. Do you have something concrete that would back that up?

They weren't actively denying anyone coop points or preventing their partners from attempting it. They just weren't doing the coop bridge themselves. It was a simple case of doing what was in their best interest (show off their robots capabilities), under the condition that they knew they could not rank high enough to be picking.

[Pat Fairbank]

Originally Posted by themccannman

Curie division last year, 254 was having robot malfunctions through most of quals which ended up leaving them ranked in the bottom 30 teams. During their last couple qual matches they got their robot working again and started shooting the lights out scoring nearly 100 points each match but not going for cooperition points which did 2 things:

it allowed them to shoot for the entire match instead of balancing to demonstrate their offensive ability, and it caused teams they played against to lose coopertition points which increased the likely hood of 341 staying the #1 seed (which was the team they wanted to be picked by).

True, and the main reason why the ranking system was so fundamentally flawed (from a game theory perspective) last year. If it came to a point where you were guaranteed not to seed, you had absolutely zero incentive to participate in co-op balancing (unless trying to show off your balancing ability to potential alliance captains). It was a much better use of 254's time to practice and demonstrate our ball scoring ability, so that's what we did (while not discouraging our partners from co-op balancing if they so chose).

[ErvinI]

IIRC, the top 8 ranked teams at Buckeye each captained their own alliances. There were a lot of high caliber shooters (706, 2252, 2016, 2834 etc.) that had various problems in qualifications, yet were fully functional by noon on Saturday. They ended up becoming picks instead of captains.

Just shows how rankings don't truly reflect the capabilities of a team, even in a year without coop points.

[themccannman]

Originally Posted by Pat Fairbank

True, and the main reason why the ranking system was so fundamentally flawed (from a game theory perspective) last year. If it came to a point where you were guaranteed not to seed, you had absolutely zero incentive to participate in co-op balancing (unless trying to show off your balancing ability to potential alliance captains). It was a much better use of 254's time to practice and demonstrate our ball scoring ability, so that's what we did (while not discouraging our partners from co-op balancing if they so chose).

I absolutely agree that the ranking system and coopertition points were very flawed last year which is why I hope they never bring it back since it just causes problems exactly like the one you guys encountered. You guys made the correct strategic choice and I would have made the same, if I was not in a position to make top 8 then coopertition points would be my lowest priority. I'm merely pointing out how much more flawed a different elimination round system could be.

[themccannman]

Originally Posted by Tom Line

Hmmm. That can be a fairly touchy accusation, because by implication it is suggested that they were costing their own alliance partners points as well. Do you have something concrete that would back that up?

I never said that they were denying anyone coop points, however, by not actively going for a bridge balance you have reduced the probability that coop points would be scored that match. If your robot is the best one on your alliance at balancing you are not actively denying a team coop points but you are reducing the probability that coop points are scored. You're also forcing your teammates to not score a double balance if they want coop points which makes them choose if they want to score coop points and put themselves at risk of not winning the match, or if they want to score points to win the match but not get coop points. Logically speaking they weren't directly causing anyone to lose coop points, in fact I would have made the same decision they did, however, from a statistical standpoint they were costing both alliances coop points.

[Carl C]

I would like to offer an antedote here. Basically, there can be more than one "super alliance." For example, at the 2013 North Cololina Regional, the top ten ranked teams each selected each other, which created a cluser of five "super alliances." Though it worked for some, two of these alliances were out the first round, including the second seed. My point? The first pick is not everything and I believe that captains should be able to pick the best eligible robot.

[Donut]

Originally Posted by Tom Line

I think a lot of the discussion here simply highlights why everyone should move to a district system if possible. Basing your attendence at worlds on one competition would be incredibly stressful. Instead, a system like Michigan's where a solid team just has to play well and not win a single event rewards consistent performers.

I like the district system a lot because it gives incentive to try to win every match rather than game the rankings by sandbagging, and it makes it even easier to qualify for the Championship without winning an event than the new wildcard rules do (though the wildcard rules are definitely a good stop gap until everyone is in districts). As a plus when everyone has moved to the District system the robot quality at Championships will see an uptick as no teams will qualify by being carried to an event win by a "super alliance" (I'm aware this doesn't happen at all or even most events, but it does happen).

On the note of budgets, I don't see it being practical (or a good idea) for 2 reasons:
1) Everyone already has the same budget limitation for the robot ($4000).
2) If you look at general team budgets you unfairly penalize teams that have to travel far due to their location (being in Iowa we have a budget of $15000 just to attend 1 regional with no practice bot or anything else fancy).

[DampRobot]

Originally Posted by Tom Line

Hmmm. That can be a fairly touchy accusation, because by implication it is suggested that they were costing their own alliance partners points as well. Do you have something concrete that would back that up?

I can vouch for themccannman. Something like this did go on according to a reputable friend. I believe the motivation was to maximize their OPR in order to improve their stats.

[Cory]

Originally Posted by DampRobot

I can vouch for themccannman. Something like this did go on according to a reputable friend. I believe the motivation was to maximize their OPR in order to improve their stats.

We don't care about OPR.

We absolutely told all our partners (and opponents) we would not co-op balance at a certain point at champs when it was clear we had no chance at seeding, as Pat said in this thread. Our operator became our base driver two weeks before the event and the operator at Champs was brought onto the drive team at the same time. We needed our drivers to gain confidence in the robot and get as much experience at scoring from the key as possible. Spending the extra 45s scoring made a huge difference in their performance as the event progressed.

There was no nefarious intent to deny our partners or opponents co-op points though. We simply needed to maximize our own performance and our attractiveness to potential alliance captains, which we successfully did.

[DampRobot]

Originally Posted by Cory

There was no nefarious intent to deny our partners or opponents co-op points though. We simply needed to maximize our own performance and our attractiveness to potential alliance captains, which we successfully did.

Fair enough; perhaps I was mistaken. Seems like a very reasonable and successful strategy. Just saying that at times, strategic decisions are made to appeal to scouts, like you chose to do. Certainly nothing wrong with it.

To return to the original thread topic... A lot of threads like this tend to pop up around this time in the season. It all seems to stem from the belief that elims are essentially a formality at the vast majority of regionals, because the two best teams are basically assured their blue banner (barring any catastrophe). I've certainly had similar thoughts of my own in the past, and there's even a fair degree of truth of it. It's pretty telling that I was able to call the two winning teams of SVR with a large degree of certainty Friday night.

Would you rather have some one other than the best two teams win the regional? This is what the proposal seems to achieve. As much as it sucks to be a team that feels like it has no chance of winning a regional, it would suck even more to be a team with one of the two best robots and not win the regional. While every team works hard just to put a robot on the field, top teams work harder to put the best robot on the field possible.

Lets not forget that upsets do happen. Davis was won by the fifth alliance, and the first alliance with the two most stacked robots at the regional in terms of OPR was eliminated in the quarterfinals! The current system does a pretty good job of getting the best teams to the finals while allowing upsets to happen. Like democracy... It's the worst system except for everything else.