I’ve always had this idea, and after hearing how accurate cross-year OPR averages are at predicting how teams will perform, I think it may be viable.
Essentially go through competition with the same alliance. Have a 3 team pit area where all 3 teams work together to create to perfect alliance, rather than the perfect robot. I imagine the alliance’s being distributed a week before competition.
Similar to sports where its really an alliance that wins the championship, not a small selection of the best players. I think it would be pretty cool to see how fixed alliance’s attacked a game vs random alliances. Also, to see how a 3 team alliance, utilized a pit would be pretty cool, you could have a lot more space if you were sharing it/tools and whatnot.
Also it would make it easier for spectators to follow what is going on as your would have 20 teams (alliances become teams) to follow as opposed to 60.
You are entitled to your own opinion, but I think this would be really unfair and it would take the fun out of the competition.
Playing with different teams every match and working together with them is half the fun. Teams that play well earn the right to play on a good alliance. If you got stuck with a bad team, the entire competition would be ruined. I wouldn’t want to pay $4000 for a regional and get stuck playing with a team that couldn’t do anything.
Working together with the same alliances would be interesting for strategy, but not interesting enough to be worth it, given the aforementioned problems.
My intention was that stronger teams would help the weaker teams, and therefore increase the overall level of play.
If I had been paired with a weaker robot, you better believe I’d strap a minibot to it and turn it into a strong robot, immediately
This year when it comes time to elims, more times then not it comes down to the 2 best minibots. Create a minibot monopoly and you win. If there were fixed alliances I think elim rounds would be a lot more interesting.
Thanks for your feedback though, I’m sure many if not most will share your viewpoint.
If implemented, IMHO this would undermine a major tenet of gracious professionalism (mutual gain through cooperation*) and increase us-versus-them mindsets (they’d always be your opponent, not a potential partner in a future match). Both of these would have major negative consequences for the FIRST environment/community.
If they are always your partner, it’s not GP to help them, it’s a more selfish looking-out-only-for-my-best-interests mindset. With rotating partners through qualification rounds and the chance that your opponents one match may be partners in a next, helping another team is not selfish as while it may benefit you, it will likely also benefit other third-party teams, thus mutual gain through cooperation.
Getting ‘stuck’ with a ‘bad team’? Yes, you may not win as many matches as otherwise, but you’d have a wonderful opportunity to have a great impact on the less experienced team by inspiring them, teaching them and raising them up a few notches during the weekend and for the next year and years to come. That’s a great challenge and stepping up to it and meeting it is a big win and accomplishment. When you see them again the next year and see how far they’ve come, you’ll feel a great sense of accomplishment having been a big part of that.
That’s what FIRST is about, more so than making sure you come out on top on Saturday afternoon.
Having said all that… I do believe the random alliances are a better model.
Team 2530 “Inconceivable”
Mentor, Drive Coach, Inspector
How does that help the weaker team, other than getting a free ride at competition?
Pair up the powerhouse teams with rookies… and you’ll end up with those Powerhouse teams rebuilding the rookie teams robots at competition. That rookie team then doesn’t get to see the fruits of their own labor. They don’t get to drive the improvement process themselves - they get it handed to them. So the following year they show up without having improved, and someone rebuilds their robot again. And again. This type of system wouldn’t encourage individual team growth.
Lots of naysayers… hmm. I think it sounds interesting, but I can see where it would complicate the events. An off-season might be a great time to play with this though.
Maybe a system could be devised to group the teams based on OPR or previous years data or something, to help keep groupings even. Rookie teams with other rookie teams, or mix it up, make each grouping 2 veterans and a rookie, something like that.
Given the rougher-than-average schedule 2815 drew at Palmetto, I would be leery of being stuck for an entire tournament with one alliance, no matter how many well-planned algorithms tried to ensure parity.
(Plus, I won’t lie, I was bummed that we didn’t get to see 1902 amongst others at all in qualifying. We wound up meeting in semifinals, but there were far too many variables to guarantee that in advance.)
I guess we’ll write this off as another one of my bad ideas :).
The idea originally came to me with regards to autonomous. I thought by having a constant alliance, you could do some really cool collaborative autonomous modes, then again, thats probably too advanced for FIRST at this time.
I think the benefits of this idea would be balanced out by a ‘meh, we’ll just get it from our alliance teammates’ attitudes. Why bring your own spares when others may have them? Why go through iteration 99 of your minibot when you could so easily borrow your partners’ every match?
Of course, I’d do it on the cruise ship if doing such alliances were the only way to have a competition on a cruise ship…
Right but its a very real world application. Right now I am working on a project between 4 or 5 different companies, each bringing in their particular expertise working to build a product. Yes it doesn’t always go smoothly but in technology based companies this is a very real experience.
I think having alliances through the build season would emphasize this more, I just don’t know how it would be done throughout the build. That being said, it could be mandatory that each team on an alliance be the expert on each part, so you could have the drivetrain expert, the manipulator expert, and the minibot expert. Also making veteran teams wait on a minibot from rookie teams, is a very real experience. I can;t tell you how often I’m waiting for designs from a design firm, keeping me from getting started on anything.
I think if it was done through build, it would create an amazing real world experience, because in the real world many project span several companies. Trusting someone to deliver for you would also emphasize earlier deadlines, and the system would get around the COTS framework.
Again, its just an idea, i just think it would produce a totally different type of game.
Another potential problem is that (excepting powerhouse teams) teams have enormously varying performance levels from year to year. It would be very difficult to group teams so that you had an even grouping based only on their past performance. For most teams, having an 80th percentile OPR or winning 90% of their matches last year means pretty much nothing about their performance this year. Maybe all their students graduated. Maybe they gained/lost a really good sponsor or mentor. Maybe they just didn’t think of a good design this year. All those factors mean it’d be really difficult to group teams so that you had a good competition and not just a one-sided massacre.
However, that problem would be solved by running it as an off-season event. Assuming that teams didn’t massively improve their robots after championships, you could probably use their best OPR or best Win% to group teams together.
I actually like this idea as an off season experiment. The major uncertainty here is not robot improvement but regression due to a driver change. This is probably not as big of an impact as on the lower tier, so here is my suggestion:
-Hold an off season competition, cap the registration at a multiple of 6 teams
-Have a selection committee identify the lower third tier of robots, using stats and elimination draft postion.
-That lower third is notified a week before the competition that they are alliance captain than they need to make a pick list. Scouting data is shared with them.
-First thing we hold a draft where those team pick their 2 partners is serpentine fashion. No inter picking, no declines.
-One Practice round per alliance.
-Round robin play between alliances to determine seeding. Depending on the size of the bracket & number of alliances, you could eliminate the lower seeds from the elim bracket. Although, instead of excluding the lower seed alliances I prefer single elim play in games for the lower seed to get into the bracket.
-Play the elim bracket out as best of 3 series (outside of play in games), and crown a champion.
I like off season experiments they are interesting beyond the same old game. It requires more work as an organizer, but it may be more fun. I also like the role reversal with those teams the didn’t make elims as an alliance captain with a week to make a pick list. It would be interesting.
I think that would be pretty awesome, and a good experiment. Maybe we can convince some of the offseasons in the northeast to take up this idea, so I can see it first hand.
I like your idea of putting the lower third in the captains position to pick their teams, this would help keep the alliances balanced (no dominating powerhouses) and could be very good for those teams to work with the ‘better’ (I use the term very loosely) teams for their personal growth.