How Do You Determine Which Team Would Be A Good Option For An Alliance Partner During Tournaments?

[andromedagalactic4]

Hi Everyone,
I’m just curious on how you determine which team would be a good team for an Alliance during tournaments? Do you look at their stats? If so, what part of their stats do you pay attention to the most? Do you ask other teams questions about what their robot is capable of doing, and maybe practicing together with them when you have the opportunity to? Do you watch some of their matches? Do you take notes? I just want to know what your team does so that we get some ideas. In the past, our team usually got eliminated R16 or Quarterfinals, so we could use some tips.
Thanks!

[EcstaticPilot]

I usually like to look at their robot design, what organization they are from, and their ccwm all as contributing factors to make an alliance pick.

[2775Josh]

I remember Wednesday night of worlds we went back to the bnb and ate dinner and then just watched matches for like 3 hours. We made our list of 5 teams to watch for and then watched all the matches, and ranked the teams just based on what we saw in match. And I think we made the right pick. So that’s my answer: just watch their matches.

Stats are kind of useless, and so is other teams’ opinions. But if you’re losing in QF, it probably isn’t an alliance selection issue anyway.

[technik3k]

It’s all of the above: you have to delegate one of your team members to watch all games and see which is competent at gameplay driving and how well would their robot autonomous complement yours.

Then sometime before launch break choose several teams that you like and ask them if they want to do some test practice together during the break to evaluate potential alliance partnership.

Some teams will snub their noses at anyone below their perceived game level, but you will be surprised how some of the best teams will consider partnering with novice teams that have competent students and solid but not yet sophisticated robot.

At the very least if they see a young but serious team, they may play with them on the practice field and give advice on what needs to be change in the robot design, autonomous routine, or driving strategy to be a good potential alliance partner next time.

[Harik_Sodhi]

I take a note of every single robot at the competition and everything they can or cannot do, usually on an excel sheet. At worlds, I did this on day 1 before any matches. I get my other team members to get the Robot inspected etc. and I would just focus on making a table. For example, in Tipping point:
How many motors in drive?
4 bar?
How many mogos can they climb with?
etc.
That night, or just afterwards in a smaller competition, I would rank all of them from ideal to worst. Before the tournament, I would look up skills score: if they have a very good autonomous skills score, they probably have a good autonomous routine in a H2H. As well as this, you can have a look at number of wins vs Win Points and calculate AWP rates. If there are any videos of past competitions, I would watch those beforehand.
You should also have an idea of the “perfect alliance partner.” Going back to the Tipping point, we wanted a team with:
a fast auton
Ability to do rings
Ability to climb the platform with 2 mogos
Good driving so that we can stack around them
In reality, you will probably not get a team like that, but using the processes above, try to determine which team fits the description of a perfect Alliance partner best.

[andromedagalactic4]

@2775Josh

Stats are kind of useless

This seems to contradict the other responses. If I am understanding correctly, you believe that stats are useless because seeing what actually happens helps with determining the right Alliance partner? Or why do you believe that stats are useless? I just want to hear your point of view.

[dvia_98377G]

the opr, dpr, ccwm stats don’t really mean a lot, you want to alliance with the team you think you can win best with rather than just picking whoever put up the most points in quals. i would say you should make a short list based on who you already know is good and then watch those teams in match as well as talk to them for final decisions

[nah_yo]

What my team has learned to do is we review the tournament rankings and watch for teams ranked in the top 16. from there we look at their highest scoring match and rewatch it on the competition livestream, if there isn’t one i would recommend looking at the match breakdown in vex via, and get an idea of how much the team was actually contributing to the score by looking at their alliance for the match. We’ve learned that at many competitions the rankings are not always indicative of good teams, so if you want to base your selection off of rankings you need to do a little research on whether or not they are actually good.

[Asinkid]

my team goes

1. teamwork some teams don’t know how to work with other teams and are hard to work with and they usually fail criteria 2
2. driver’s skills/mentality. You can’t compensate for a bad driver. Some good drivers choke and the fall during round of 16.
3. complementing factor’s. Shaq and Kobe were the best they had complementing factors like shooting and dunking

[Worthington3799]

Personally, my region is very, very small. So we know pretty much every single team and are friends with a lot of them. So we know their organization’s capabilities and their individual team’s capabilities. That degree of knolwedge of other teams’ capabilities over the season or the year helps, but a lot of the time you can squeeze that knowledge in during the tournament. If you guys are getting eliminated in R16 or Quarterfinals, 2775Josh’s point about avoiding stats is probably the best idea. What I’ve found is that you don’t always want to pick the best performing team stats-wise. A lot of the time they can get carried by other teams that are better than they are. Pick a team you know you synergize with and can strategize with. Ignore the stats and find the team that best suits your play style and that you know you can talk to before, during, and after a match, regardless of what happens.

[Trent_13444A]

My team looks for

1. Compatibility between the robots and drivers (will we work well in an alliance)

2. Skills score / rankings (especially driver control scores)

3. Strength of schedule and win points/ranking after qualifying

Those are our main 3. Personally I find OPR, DPR, and CCWM to be unreliable so I try to avoid using those for alliance selection.

[4073B]

Region 3 = Best California Region

[aivintle]

Every team has different goals going into a competition. Some teams are aiming to win the entire thing. Others are there just to have fun, and are aiming to drive. Yet others are aiming in the middle, and are happy to get picked for alliance selection. My first piece of advice going into any robotics competition is to make a goal: what do you want to get out of the competition?

Next off: strategy. One of the top robotics strategists, Karthik Kanagasabapathy, has given his ever-changing presentation Effective FIRST Strategies many times (he is on VEXforum, but I will not invoke his presence, someone else may feel free to do so if they have questions on one of his presentations). This presentation specifically targets FRC, but many of the points of strategy advice that are given can also be applied to VEX. You could do a quick search on Youtube for one of the versions of this presentation (do note that it is a long presentation, but it is a worthwhile listen and quite entertaining).

Now onto my own opinions(0) on how to choose teams for alliance selection. First off, I would always begin thinking about alliance selection before a competition early, if there was a remote chance that I could be participating in playoffs (at smaller events where almost all teams participate in playoffs, this would mean that I would definitely plan to choose an alliance partner). Please note that this post includes a lot of information about information to collect, towards the bottom of the post there is a section detailing ways to gather/compile the information collected)

Before The Competition

Before the competition, there are some things you can do. You can look at what teams are attending the competition. If this is a more local competition, you may already know some (or a lot) of the teams attending the event. Additionally, if this event is later in the season, you may already know a good deal about the teams competing in the event. You can begin to do some research about the teams that you may not know so much about.

Some of the points to consider **before an event** (in approximate order of importance) (1):

1. Big Accomplishments this season - relatively quick data points to gather that can identify top performing teams quickly.
   a) Tournaments won
   b) Major judged awards
2. Smaller Accomplishments this season - takes more time to put together this data, but they can include more insightful knowledge. If you notice a team consistently placing top 4/8/16 at events, but aren’t winning events, you should do some thinking/digging to find out what happened, especially if you think that it may be relevant.
   a) Number of matches played
   b) Rank at each event(s)
   c) Smaller judged awards
   d) Matches won
   e) Reveal video (check youtube, forums, social media)
3. Stats - You may be able to quickly pull this data from a database/api. If you have it, it’s nice, if not it is definitely not the end of the world.
   a) Average points scored ______ (fill in the blank with the quality you are looking for in an alliance partner)
   b) Average points scored
   c) maybe some opr/ccwm (2)
   d) Where are they from (are they from kind of far away?)

Start of competition

So now you are at the event. What do you do now? First off, look at the number of teams at the event. Now, look at the teams you have information on. Now, it’s time to gather some more information that is more relevant to the current tournament/event. At this point, if you have time/team members, I would “pit scout” some of the teams that you are interested in working with.

Teams to pit scout include:

1. Teams that stood out in your pre-event scouting. That’s why you did pre-event scouting.
2. Teams that you are alliance partners with. Also talk about match strategy with them for your match with them before the match. You have (generally) played one more match with your alliance partners than with other teams if you are at your first event of the season, so you have some practice playing with them.
3. Teams that are right next to you - always good to have friendly pit neighbors, unless they are teams in your own organization, in which case you should not pit scout them or pit scout them before the event
4. Any and all other teams that catches your eye. Even if you maybe aren’t interested in partnering with them at this event for whatever reason, you still should talk to other teams, especially if they have something interesting on their robot you could learn about and apply on your own robots.

So you know which teams you want to pit scout. What is pit scouting and how do you do it? Pit scouting is when you go to each of the team’s pits (hence the name pit scouting) and ask them about their robot. Don’t forget to always introduce yourself, with your team number. Some examples of questions you may want to ask include:

Drivetrain:

1. What type of drivetrain do you use?
2. How fast is your robot? Some teams may need help with this question. A quick glance at motors/gears could help you figure this out on your own.
3. What type of wheels do you use (although you probably could figure this out with your eyes)?

Scoring (Season specific to Over Under):

1. How many cycles can your robot complete in a match
2. How high can your robot climb in a match?
3. Does your team focus on scoring triballs in the offensive zone or goal?
4. How does your robot preload/alliance station load?
5. How long does it take for your robot to score all the triballs ___?
6. How long does it take for your robot to score the triballs if we/opponent does ______?
7. Programming/driver skills scores

Other questions (I always ask question 1)

1. Tell me about your robot/is there anything else you would like me to know about your robot.
2. Can I take a quick picture of your robot? (this can help later when trying to remember which robot is which when choosing).
3. Name of your robot?
4. Can I have a pin for my pet [insert team (stuffed) animal mascot here]? (3)(4)

During the competition

During the competition, if your team is large enough, or if you can find another team to work with, you should do some match scouting. When you are doing match scouting, you are trying to gather qualitative and quantitative data on other robots so that you can make an informed decision on who to choose. Your pit and pre-scouting should guide your match scouting, but you should still consider all teams (5). The extent of your match scouting is heavily reliant on how many person(s) and how skilled the person(s) doing scouting are.

Qualitative Observations

1. Auto: Robot touches bar
2. Auto: Robot scores alliance triball
3. Auto: Robot removes triball from match load zone
4. Tele: Robot plays defense (Choose: Excellent, Good, Attempted(6), Does not play defense).
5. Tele: Both teams on an alliance are/(are not) working together to climb
6. Comments: Anything else about the robot (Did not move, broke down in the middle of match, scored a lot, had a lot of defense played on it, descored 15 triballs from an opponent goal, etc.)
7. Major rule violations

Quantitative Observations

1. Auto: Robot scores ## game pieces
2. Tele: Robot scores ## game pieces (can be challenging in a match where both robots help to score game pieces)
3. End: Climb time
4. End: Climb level
5. Find ## penalty/foul points

After Qualification Matches(7)

As qualification matches come to a close, your team should begin making a picklist, based on your team’s goals (this section will mainly assume your team is ranked within the top 16 teams). This picklist should include at least 24 teams, and could go through all of the teams at the competition. If you have good match scouting data, start working with that. Sort teams (make a picklist) to find a partner that will help you accomplish your goals. If your goal is to have more fun, choose a team that you like working with. If your goal is to score a lot of triballs to win matches, choose a team that will help you score a lot. If you want to be able to play offense (or defense) while your opponent does the other, make sure to choose a complimentary partner.

Some teams create multiple picklists, one for offense, and one for defense. I would personally recommend creating one picklist with both offense and defense teams, with the perspective that every point denied while playing defense is a point gained if offense was played instead. I would also say that teams should honestly evaluate their performance and place themselves on their picklist. This could help to point our team on how they can improve.

If you are the first seed,

you are guaranteed the first pick. You should be active in finding an alliance partner that you would like to alliance with. Talk to some teams that you would like to pick before alliance selection. If you believe that an ideal alliance partner is one of the top seeds, talk to them before alliance selection. Even if they tell you no, it doesn’t mean that you can’t pick them during alliance selection, just that they may say no. If several of the top seeded teams decline your alliance, don’t worry, as those teams will not be able to alliance together making it easier for you to take them down later in the competition. Also, do not box yourself into the top seeds in the competition. Use your data and look for upward trending teams, who may have had a rough start (auto wasn’t tuned in yet, driver’s first time at the controls) and they may continue their upward arc into playoffs.

If other teams try to talk to you about alliance selection, hear them out. They may have a really good insight that you may have missed. In addition, if you are a top seed and they propose a plan with some risk (they have a wall they haven’t used or they have been programmatically limiting their speed for quals, ask if they can show it to you. Do not try to boss around other teams — it will make you look bad, but if a potential alliance partner wants to show you something to use in playoffs, do ask first to make sure they are being truthful.

If you are not the first seed,

You should still talk to other teams. Look at the teams near the top of your picklist. Ask those teams if they would be willing to alliance with you. Have some reasons (showing them your scouting data can be a hit or miss), you are saying that they should choose you because the data that you collected says so… That said, reasons for wanting to alliance with a team can be “You are all really nice and we like playing with you in match number x” or “Our robots compliment each other because you do this well and we do this well.” Do not put other teams down while making comments like this, it makes you look bad.

If you are picking from a lower seed, and you have some data, look for teams that had some really good matches, but may not be as consistent for whatever reason. These teams may be ranked low because of their lack of consistency, but have the capability to shine in playoffs after they worked through the robot gremlins.

If you aren’t the first seed.

Data Collection

Throughout the rest of my post, there were many different mentions of collecting data. How can you do that? There are many different apps that can be used to collect data. I would break these down into several different categories.

Spreadsheets are good for storing data values such as numbers, strings (Names, locations), and links. They can serve as a place to gather data from many sources. They also can be used to sort and filter data, such as teams, by different characteristics.
Databases are spreadsheets, but additional information can be attached to each team. If you have ever seen/heard of a mind map, they are a mashup of a mind map and a spreadsheet. Databases can visualize data in different ways allowing it to be manipulated and sorted/filtered similar to a spreadsheet. In addition to storing values like a spreadsheet, databases can also contain images, videos, and documents for each team (while these can be put in another column on a spreadsheet, it tends to be messier and databases are generally built to handle mixed media more easily.

Below are some software recommendations for software that I personally have used to build scouting apps.

Spreadsheets:

• Google Sheets
  • Good if you have several team members who want to help collect scouting data
  • Some of your team members may already be familiar with this software
  • More user friendly, a quick google search can generally help you if you get lost
• Excel
  • Good if only one team member is doing all the strategy/scouting
  • Excel can be used with Google Drive to sync changes
  • The most powerful of these three spreadsheet options
  • Not as user friendly as sheets, but there is much more documentation and help available if you need it
• LibreOffic Calc
  • Open-source spreadsheets that is similar to excel
  • Free
  • Some limits in functions supported

Databases

• Notion
  • Really good interface for working on a phone
  • Also useful as a tool to make planners or calendars
  • More user-friendly than JSON files
  • Data can be imported/exported
  • Each team is given a “page” where notes can be written about the team, all scouting for a competition can be done in one database and team members can put all information about a robot (pre-scouting, pit scouting, match scouting qualitative observations) into one page. Teams can then be sorted/filtered by their characteristics, just like in a spreadsheet.
• JSON files
  • Many APIs export data in JSON
  • Readable data storage format, although far more difficult to read as the other data formats by a human when it is in it’s stock format

Other:

• Google Forms
  • Integrate Forms with Sheets to do match scouting and link it to something that can analyze the results more easily- each scout submits one form per match.
  • Easy to understand how to use
• Programming Language (Python, java, etc.)
  • Use some programming to pull data from APIs
  • Please be kind to APIs if you are going to do this, don’t bombard the API with too many requests
  • A .txt file that is tab delimited can be pasted into a google sheet nicely, which is a good and easy way to take data from the web and integrate it into a database
  • A scouting app can be made for pit scouting and/or match scouting to easily record information that can be saved/exported to a spreadsheet/database.
• Google Slides
  • I personally would use google slides or notion for pictures of robots
  • One slide per team, label the slide with the team number and add the picture while pit scouting

I hope that this post was informative and will help you make strategic decisions during alliance selection into the future.

Notes

Throughout this (long) post I have notes indicators (#) for text that didn’t need to be included above.

(0)

The opinions expressed in this post are mine, and do not necessarily reflect those of my team or fellow team members.

(1)

Gather this data before an event. If you do not have time before the event to gather this data, it isn’t the end of the world. Especially at large events (States, Sig events, worlds), it probably isn’t worth the energy to gather all the data. I would probably filter teams using the first criteria and only would look up information for the second criteria for a smaller set of the teams (20 teams). Searching this data up using an API could be another way to get the data quickly. Divide and conquer also works well - use your teammates/ask your teammates for help gathering data.

(2)

Personal note/rant: I dislike opr/ccwm because they can often end up becoming a better strength of schedule metric instead of a robot quality metric. While they can be fun to play with and may be correlated to robot performance, with limited number of matches, especially on a local level, especially earlier in the season, these numbers may turn out to only be numbers, and it is always best to, at the least, do a small “reasonability test” before blindly using them.

(3)

If you are looking for friendly teams to alliance with or teams that have lighthearted members, the “other” category questions are great to ask. In addition, asking other teams for a pin for your pet turtle rarely gets old.

(4)

This section only has a small sample of questions you could ask. Feel free to let the conversation go where it goes, especially with teams you want to alliance with. They are more likely to choose teams that they know well, and teams that they talk to for extended periods of time are teams that they will know well/remember. Don’t forget to tell them your team number, and if your team hands out pins or other similar items do bring one/some to them when you go to talk to them so they have something from your team that may help them remember you when they are trying to pick an alliance partner.

(5)

Okay, so your match scouting is mainly so that you can try to find a good alliance partner, so if a team really didn’t get along with you and you don’t want to alliance with them, then maybe don’t scout that team and mark them down as Do Not Pick (DNP).

In addition, if there is a team that is clearly struggling to drive, they should probably also be marked down as DNP.

If a team can drive, but they cannot score, you could consider having them play defense. In a game such as Over Under, most teams that can drive should be able to score (place match loads such that the robot can push it across the field) or play defense on opposing alliances. If a team clearly struggles with both tasks/are unwilling to do either task, then they probably should be DNP.

(6)

The word “attempted” here is my polite way of saying bad. If the word attempted confuses you, please just use bad, but don’t do it in earshot of other teams, lest word comes around that you are saying the team XYZ is playing bad defense.

(7)

You could start to complete the steps listed here before the end of qualification matches (and it would be a good idea to, especially if you are ranked 3-10).

[andromedagalactic4]

Thanks to everyone for their advice. I greatly appreciate it and will definitely take all of this into account as my competition season begins.

[Dave-TBR]

Most of the good information has already been said here, but I’d like to add some of my experience. Alliance selection takes different shape depending on the size of a competition. At larger competitions (more than 16 - 24 teams), @aivintle’s previous post on this thread is perfect. However, for smaller tournaments (under 24 teams), there are some different aspects to alliance selection (in my experience). Here are a few things that can happen during smaller comeptitions:

• Low level team gets carried (top of seed picks)
• High level team has a bad match schedule (low seed picks)
• Mid Level Teams can be spread across the entire seed rankings
• Elimination matches can be interesting

I’ll elaborate on all of these points eventually, im terrible at making concise forum posts.

Especially at lower level tournaments, if you are looking to win, you need to talk. A lot. During one of my spin up tournaments this year with 16 teams (idaho moment), we talked to almost every team to get strengths and weaknesses of everybody, which really helps to narrow down who you should actually be focusing on. This also carries over into match scouting. You can quickly weed out who you shouldn’t be paying attention to just by a look at a teams robot, or with a quick chat (pushbots, jankbots, herobots, etc). Match scouting higher level teams is crucial, both for potential alliance partners, but also knowing what your competition is, and what they will do in matches.

At smaller tournaments, low level teams can be carried into a high seed ranking, which can ruin a lot of potential alliance partners (due to the rules of alliance selection, if someone is asked to be a partner, and says no, they can only be a team captain). Talking to these lower level teams right after elimination matches end can be very helpful. If you make it clear that you won’t accept, it can save you from not being able to be picked from the teams you have already talked with. Please be sure to communicate this in the nicest way possible, as low level teams are usually new to vex tournaments, and we want to make sure they feel welcome in our competition. At another Spin Up tournament this year, a rollerbot got carried into #3 seed rank, and proceeded to choose all the high level teams below them, all of whom rejected. This meant that none of those high level teams could pick each other, which they had planned on doing.

High level teams can have a bad match schedule, have issues during the competition, and end up in a low seed rank. If you are in this position, you are the one that needs to do the talking. Mid level teams can be much more open to picking a lower seed team compared to a high level team. I’m guilty of not picking a high level team in a low seed rank, usually because low seed rank means consistency issues, penalties, etc. If you are a mid level team, and see a high level team in a low seed rank, talk to them, try to get an ez alliance partner.

Mid level teams are great. They can do well in matches with a good alliance partner, they are usually spread across the seed rankings, so you usually have access to pick one. Talk to the mid level teams, they’re pretty chill.

Elimination matches can have major upsets. They might be rare, but they do happen, especially at smaller tournaments. You don’t have to have the best alliance to win a tournament, you just need to win all your matches. Its generally better to pick the more consistent alliance, over the higher scoring alliance. During yet another spin up tournament, our team picked a lower seeded team, and also lower scoring, but they were incredibly consistent. Ez tournament dub.

If you’re first seed, 2nd seed isn’t always best.

Sorry for this dumpster fire of a forum post, reply if you have any specific questions.

[aivintle]

@Dave-TBR

Low level team gets carried (top of seed picks)

If you are ever a “low level” team, or feel that your robot isn’t the best, but end up in the first seed. Remember, you still earned your way there. Yes, maybe you had some really good alliance partners, but an alliance is made up of two teams, not one.

If you are ever in the position where you are picking from a top seed spot, even if teams approach you beforehand and tell you not to pick them, you still earned your spot, and if you are inclined to do so, you should pick them (especially if it would help your team do better in the elimination rounds). If they decide to decline your invitation, it is their team’s loss. There should be little to no shame in using a “burn” strategy, picking top tier teams until one accepts, giving your alliance an advantage as you are splitting up teams. Yes, the other teams may be upset, but you still earned your way to the first seed and have the right to pick the team that you would like.

*

Opinions expressed are my own.