Hi Ben. I’d be happy to take a stab at your questions. My full opinions and advice about alliance building can be found in this guide @Katie_UPS and I put together:
Gettin' Picky: A Guide to Alliance Selections (Part 1: The Picklist)
Gettin' Picky, Part 2: Alliance Selection
I look for partners that will maximize our chances of achieving our goals for that competition. If we are at a regional and we have yet to qualify for Champs, that goal is a win or a wildcard. If we have qualified or are at Champs, our goal is to win.
To maximize your chances of winning, you must consider all possible ways to 1) maximize your score and 2) reduce your opponent’s score. From here, choose a strategy to go about this based on your robot’s strengths and weaknesses. Based on, make a list of needs and wants for robot abilities and characteristics for each pick.
Our robot was pretty good at both game pieces (especially hatches) and could level 3 climb, so for our first pick our list looked like:
- Able to score lots of a single kind of game piece, ideally cargo
- Able to score under defense without getting shut down
- Able to score both game pieces well in teleop
- Level 2/3 climb
- Scoring in sandstorm, ideally hatches
For second pick:
- Tank, swerve, or if mecanum, demonstrated ability to play very good defense.
- Driver skill, either demonstrated by defense or scoring
- Level 2/3 climb
- Scoring hatches in sandstorm (or cargo)
- Teleop scoring ability
If we are choosing between two similar robots, we will prefer teams who we gave a history of working well with (in quals or past events) and avoid teams that we have had issues with in the past.
It’s worth noting that we were the #1 seed or #1 pick at all our events (except Battlecry) so consistency was an incredibly high priority for our picks. If we were to be a low captain (like we were at Battlecry), consistency would become a lower priority in favor of adding more scoring potential to the alliance.
Rankings are an essentially meaningless indicator of robot ability. At Central New York, we chose the dead last ranked team, 395, as our second pick and they played fantastic defense.
However, rankings are a major factor influencing what alliances will form. Remember that your goal isn’t just to pick the best robot, your goal is to win/finals/wildcard/etc. Be mindful of the top of the rankings.
I have two relevant stories. At Robot Rumble a couple weekends ago, we considered the four best robots at the event to be
And the rankings were
We could have picked 2168, and while we believed them to be the best robot available, this allows a 20 6328 alliance to form, which would have a serious shot at beating us if something went wrong on our end or they had a better third robot. If we picked 20, we essentially guarantee that 2168 and 6328 will be on different, much weaker alliances, giving us a much easier path to the win.
This logic doesn’t always work this way. Going into Saturday of Central New York this year, 694 was #1 on our picklist, partly because they were the #2 seed behind us. However, as the day went on, 195 got a lot better, and our list looked like:
For a while, we considered picking 694 anyway, because it would split 195 and 2607 up because the rankings were:
A few teams
Many more teams
But then something big happened: 2607 slipped into rank #3. At this point, if we picked 694, 2607 would pick 195. If we picked 195, 694 would pick 2607. If we picked 2607, 694 would pick 195. So at this point, we just picked our top choice, 195.
Once we are chosen by a captain, our goal is to pick a robot that not only complements ourselves, but also our captain. We will do our best to persuade the captain about who we want, and we generally get them by explaining how they would complement our alliance, though we have had captains who were more and less collaborative during alliance selection. The only way to guarantee avoiding uncooperative alliance captains is to seed higher than them.
The optimal alliance is built on doing four things right:
- Choosing the correct strategy to build an alliance for based on your own strengths/weaknesses
- Choosing the right needs/wants for executing that strategy
- Choosing the teams that best represent those needs and wants
- Having situational awareness about the rankings and potential opponent’s/threats
Most failures during alliance selection can be traced on a failure on one of these four tasks.
If you have any more questions or want me to elaborate on any of this, feel free to ask here or DM me. I love talking about this stuff.