Gettin' Picky: A Guide to Alliance Selections (Part 1: The Picklist)

Being a Scouting lead mentor since year 2 of our team after a rookie trip to wolds ,I’ve participated in five years worth of pick lists: (in Southern and Northern CA) Ended up 9 measly points from World Championships earned in two years. (4 and 5 points in those seasons)…it’s certainly shown to be truly a game of few points. I base most of our scouting off horse racing. Seems to work well. You simply need firepower and consistency and some luck.

My best suggestion from the Semifinal experience perspective , is to only place on your pick list “reliable” teams. In crunch time things break. I second OP’s , “ignore ranks” and trust in “scouting data” assuming you have had extensive “eyes on bots” or in other words: Trust your scouts if you didn’t overload them.

Our pick list is always 28 deep (solely collated by our lead scout student with input from myself and other scouts its our lead scout student’s list). No matter where we ourselves rank. We always have a 28 deep list to use or share on competition day 2. Elimination Alliance selection day.

I always have a conversation early in the season with this new student lead scout allowing for any questions and placing the responsibility in their hands. Giving them ownership over the pick list process.

The other aspect for your team is RANK high enough (or be selected by high enough). Almost every competition comes down to the best two teams picking each other and getting #24. Your team(alliance) is at a serious disadvantage if you can’t rank in top 4…plain truth. Only top 4 firepower seemingly can overcome #1 and #2 pairing up. With that possibility. Then you have a good chance if you assembled correctly.

So to that end someone on your “alliance” needs to absolutely be TOP 4 or plan to go home. Hopefully your team steps up to be Top 4 or you get lucky as the first pick of top 4. If you end up Top 4 yourself, then you concentrate on “difference makers” and “rare talents” always assume you will be likely be up against #1 and #2 , how do you mitigate that? That guides your choices, that and reliability. If not Top 4 share your insights with the Top 4 Captain.

Don’t overload your scouts, simplify it for them. Make them willingly perform.

Scouting a field of 60 at a regional is NOT HARD, half the teams simply cannot help you. So concentrate your own scouts on teams that can help you get past this competition (and win known matches to rank highest you can, earning the permission to select or be picked)… Look at other sports like horse racing and see how they make picks. The most critical factor I see is being in Top 4 alliance wise with a very “good class” teams. Make that happen somehow and then pick strong. Don’t over complicate it , just assemble the best alliance you can. Again this is basic, no need to over complicate it.

The single most important stat in horse racing is “class” don’t under estimate that quality. I’ve learned from experience there.

So just adding a kinda weird point on this. When I was on 1410 I think we were a minority that didn’t really have meetings to make the picklist (although that might’ve changed). The head scout would watch every match while discussing with scouts, make it alone, and then talking it over with the drive team and mentors to make sure they were good with how the teams worked together. I’m not necessarily sure this was the best method, but it worked out pretty well (see 1410 at the 2016 Colorado regional). At least when I was head scout I found that you could figure out a lot of the “they did sketchy things” or “they don’t have sensor feedback” by walking the pits or looking at robot photos.

I’ll also say during my time we were using an odd weighted z-score system that ended up working very well for 2015 and 2016 because of the way the games were designed, so that might’ve been part of it too.

Does anyone do something similar? Do you find the group discussions counterproductive? 3 hours seems like a LONG time for a scouting meeting and I feel like the more people you add the longer it gets.

33%20PM

This is an image of some data we collected on a team at the Central New York Regional. Not what I would call especially reliable, according to our data, the 8th most inconsistent to be exact. This team is 195 and we picked them and won the regional with little to no substantial robot failures on their end.

Just because a team is inconsistent, doesn’t mean they should be DNP.

8 Likes

@anon31892661, Katie and I spent several hours on this thread. I love seeing active discussion inspired by this, but the least you can do is take a couple minutes to refine your messages to the point where they intelligently communicate your opinions and ideas.

@others, calm down a bit, this the How To Pick guide, not the How To Troll guide, though stay tuned for that one.

We prepared this guide to serve as a resource to the community, and I think the discussion on this thread has the potential to be a valuable accompaniment to that resource.

21 Likes

I will refine, ask away its about assembling the best pick list, right?
I have much to offer there.

Thanks for putting so much effort into the OP-- it’s a really great resource, and I made sure to share it with my team. I can’t wait to see part 2 and 3!

11 Likes

This topic is temporarily closed for at least 4 hours due to a large number of community flags.

So I just deleted 34 posts from this thread… If this thread goes off track again, the posters responsible could end up with their accounts being temporarily suspended. As Brian requested, let’s get back to talking about how teams can improve the picklist processes.

24 Likes

This year I ended up making all of my team’s pick lists by myself. I would recommend other people to never take this path, as getting other perspectives is really important and crucial to this whole process. Also, doing things by yourself definitely makes the pick list more prone to errors and forgotten / incorrect information.

5 Likes

The pick list needs to be by someone the team trusts, that is usually those that put in the effort to scout, that means the lead scout at least on our team.

As the scouting mentor, I help that lead scout and reach the other scouts…and they in turn teach how to scout. I don’t believe in scouting by committee , rather scouting driven with input and someone responsible for the list after all inputs.

Also speaking as someone who used to scout without any direct input on the pick list, I completely disagree. I believe anyone should be able to contribute to the pick list directly if they feel inclined to do so. I have a firm belief that scouters need to be trusted and valued throughout the process so the entire team (including them) can understand their importance to team success. Denying them direct input can make them feel quite worthless, which is exactly how I felt.

4 Likes

Fair enough Erick2175, I disagree wholeheartedly, as decision by committee usually leads to inaction. Just like we usually pick a main driver etc (from collective experiences) not everyone can drive or successfully scout pick lists IMO, Both of these individuals step up and lead. As in real life from my experience,

That’s also 195. A little birdie told me they are pretty good at the whole eliminations thing.

4 Likes

We chose 195 based off the measurable data, however I do feel past performance, or “clout” can and should have a role in picking. Just not a significant one.

Did 195’s rep make you feel better? Is it measurable?

Remember this is a pick list thread for advice for teams looking for answers.

We spent a lot of time talking to people from all three teams we were considering picking: 195, 694, 2607. All three teams were pleasant to work with and had, for the purposes of this discussion, approximately equally effective scouting programs. I think we would have worked well with all three teams.

While 195’s history played a hand in how we interpreted their data, we still picked them because of their data, not because they’re historically good.

1 Like

When looking at current data past data has no effect?

I personally find that hard to believe.

Seems to me that would be confirmation at the very least of a picks worthiness.

Teams can change from event to event. One event, they may be the top bot of the district. Come event two, they didnt iterate and their drivetrain took too much of a beating, and they place 28th. DCMP, theyre back up as the top 4 for the district as they fixed their issues.

Every event is going to be different. Holding their past against them when choosing can be short sighted even from changes in results during the events, which is why 2791 chose 195 in the first place.

Once we determined 195 was the robot that best met our needs and wants, we moved them to first on our list. However, going into Saturday morning, 195 was fifth on our picklist, because we considered them the fifth best robot at the event with respect to our needs and wants based on our data collected Friday. No amount of previous blue banners was going to change their robot and its performance. It was their significantly stronger Saturday morning showing that made them the pick.

1 Like