Firstly, I’m not representing my team here, just my personal thoughts and frustrations.
I’m just curious as to how teams detirmine their picks and what dropped teams in priority in the eyes of alliances and scouting teams.
We finished as the 14th seed with the 6th highest OPR on the field, with 5-7 gear cycling, 100% climb over our state event and our qualifiers, and a tiebreak shooter, many of us went into alliance selection assuming that we would be a first pick by one of the shooting robots on Darwin
Come to find out, we did not go in the first round, or the second round, but instead were passed up until 4th pick by the 4th seed captain, at which point we were benched for the entire elimination series without playing any matches.
My team figures it’s because we’re mecanum drive, but frankly I don’t think that we’d be disqualified for that, especially when our numbers were so good.
Scouting and alliance selection is REALLY important. Old biases and preferences are hard to change. Maybe with Team Titanium going to the finals using mecanum, some minds will be swayed.
The truth is that a vast majority of teams who use mecanum do it poorly and alliance captains don’t want to take a risk when there is so much on the table.
Here’s the process 1058 and 67 went through generating our 3rd/4th picklist:
1: Remove top 8 and obvious first round picks (only a couple).
2: Any team below a 75% climb rate was immediately removed with some exceptions.
3: We then sorted by gear throughput in teleop.
4: Teams with floor pickups were bumped higher on our list than active place/passive place HP.
5: Then we began to examine drivebases, defense, and ball shooting capability, where the picklist got more dynamic. This led us to 1640 and 2137 - two solid gear runners, one with a floor pickup and one with active place HP, but both had good tiebreak shooting and even side peg + shoot autos. In addition they both had defensive capability (tank drive/swerve respectively) and demonstrated it well, which was essential in a strategy where you had to force your opponents down to 3 rotors to win.
One thing to point out about the Darwin field is that there were a ton of teams that were strong at gears and there wasn’t much to differentiate between them.
That was another point we came up with. We may have been doing 1-2 gears more than many of the teams, but many were very similar. It didn’t help that our Average gears was lowered simply because we didn’t need to place 7-8 a match to 4 rotor since, most of the time, our alliance partners were better.
My team was explicitly told by the highest seeded alliance at our regional that if everything with our robot had been equal minus the mecanum wheels, we would have been the preferred 2nd pick. We were the preferred pick of the 2nd seeded alliance, but the first seed got the final say of course.
Before our team decided to use a holonomic drive this year, I did some research on CD and found that some teams will ignore your robot if you have one. I thought the claims were exaggerated a bit, and perhaps they are. However, I can attest to the fact that using them will decrease your chance of getting picked.
Unless you’re a powerhouse team (1986 is a good example) or defense is non-existent like 2015, do yourself a favor and don’t bother with holonomic drive trains in the future.
I’ve seen a few posts about this so I’ll outline how we made our picks in Houston because I know some teams who ranked 14-20th may be scratching their heads with how they got left out.
The big difference in scouting at championship than a regional is that teams already have preexisting relationships from earlier in the season so rank is irrelevant if you’re not a captain. We had 5417 and 5960 in our division, who we had won a regional with and came 2 seconds from a second banner with respectively so those were already two teams we felt awesome about working with again before arriving in Houston (if we were lucky enough to pick teams again). Film also helped a lot to put into perspective how good other teams are (PNW was pretty awesome this year and their district championship was harder than our championship division to be blunt).
We paid no attention to actual ranking and created out pick list strictly on strategy. In Houston we needed a team that had ability to shoot pretty well AND score a lot of gears if we couldn’t be with 1323, 16 or 2046. We also needed to find a team that had a reliable auto mode for gears. There was also one team that was pretty good who the drive team vetoed because of how they acted when they played with us (arrogant, not willing to listen to us and the other partner, being stubborn about their pilot even though our team and the other team clearly had the faster gear cycles, etc…).
Once 16 and 5104 put up the 4 ranking point match and dropped us to fifth place we thought we were screwed on getting an elite shooter. The preexisting relationship came in because we found out from a scout that 4587’s shooter was working and we knew them well from the Alamo regional and knew they had an awesome robot stuck ranked somewhere in the 40’s. Our data also suggested that 4587 was one of the best gear cyclers (i think 5th or 6th in total gears) so that helped confirm we weren’t making an emotional decision by picking them. Our next pick 4468 was based purely on data (outlined below) and our 5960 pick is because we picked them in Alamo, they improved their robot a lot to hit right side auto reliably, and frankly we just trust and like them. So again, the preexisting relationship with two of our partners played a role in why we knew about them and why we picked them.
For anybody who wants to get into the weeds about the scouting process here is what we did and I’ll be happy to talk via PM if you want to know more about the data aspect:
At all competitions this year (including championship) we had six scouts each match watching a single team for every match and we tracked: successful auto gear delivery, number of fuel points in auto, estimated fuel in tele-op, number of gear cycles, and climb success. We took that data and summarized all the teams to know who the best shooters, best auto-scorers, best cyclers, and calculate climb success rates for all teams and used that as guidance for our pick list. After we decided our top 5-6 teams who had the strategic match for us we simply eliminated all teams with less than 8 climbs, and then by order of total tele-op cycles with a bit of discussion on teams that could score in auto mode (we struggled and/or weren’t given a chance to line up center much in quals so that mattered to us a lot). We use software called Altyerx so do the summarization of our raw data very quickly for this part of our scouting process.
Try not to confuse holonomic drivetrains with low-traction holonomic drivetrains. Things like swerve and butterfly, octocanum, etc can be hugely valuable not-so-low-traction holonomic drivetrains depending on the game. The basic mecanum is simpler (cheaper, easier, lighter), though.
If it helps, the year before we made Einstein the first time (both with holonomic swerve), we ranked exactly 13th but went unpicked at both Worlds andIRI (and we completely deserved not to be).
As a team, we would first start off talking about what we wanted in a robot. Some of the things we talked about included the climbing percentage, gears per game, and Auton starting position. We also looked at teams we saw at other regionals and who knew us. At worlds this year, we did something a bit different. This was because the seeding was already chosen so we took the top 15 or so teams and took what would best complement us and them being a 3rd or 4th bot. This was nice because some captains could shoot, while others could get gears which allowed us to see what would best fit with each other.
After qualification matches we knew we would not be a captain, however we prepared for alliance selection as if we would be. We always have data and a plan ready to support whoever the alliance captain was. We never assume the alliance captain will have a lot of data.
We had a first sort by fuel, and had a ranking list based on auto KPA and tele KPA. Because there were only a few really good fuel scorers we this was not a long list and our focus was on climbing and gears.
We had a second sort that was climb % and then gears. We set a threshold for minimum climb percent, and then sorted by auto gear and total gear. We also look at trends over the two days - there is a difference between 5/10 that is every other match and 5/10 that is 0/5 day 1 and 5/5 day 2.
We also collect input from the pit crew and drive team relative to attitude, ease of working together, etc that can move a team up or down. And as others have said, previous success with a team is a factor.
We completely ignore ranking when making selection plans.
This is the exact process we followed with an enphasis on the auto gear. Our 3rd was an absolute steal because people are afraid to embrace silly wheels
Huh…pretty much what we did, too. And we thought our third pick was a steal as well; what a coincidence. It’s almost like we were in the same room…
We also looked at some video to try to determine how many of the gear numbers were less than the robot’s capability because an alliance reached 4 rotors well before the end of a match. It would have been helpful if we had a time machine for our scouting app programmers to be able to go back and record the match time each time a gear was hung.
For reference: on Daly, 862 Lightning Robotics (16th pick) had 90% side auton, which is clearly what Stryke Force needed. Although it was 1676’s instead of 862’s in F2, they would have lost without it.
I’m curious how other teams determine their pick lists in general, as we are always looking for ways to improve our (currently very chaotic) process.
Our process tends to be similar to the following:
Attempt to determine what other alliances might look like
Determine a strategy that has a chance to beat those alliances
Rank teams in a some sort of order to build this alliance
These steps usually don’t occur in any sort of near or orderly fashion, which tends to make our scouting meeting drag on for awhile. I have found this process to be much easier the higher up in the selection order we find ourselves as there are fewer unknowns to take in to consideration. (Not to mention the inherent unknown of the rankings during regional competitions, and the fact that ~1/4 of the matches aren’t played until Saturday morning)
The biggest problem we find during attempting to pick alliance partners has been something like: "We could do strategy X if we get Team A, but if we don’t get Team A, we need to execute strategy Y (which requires a very different alliance composition). I’m curious as to how team handle this situation, as the possible flowchart grows out of hand pretty rapidly (especially when picking from the 6-7-8 slots).
All of these situations become even more complex if you’re trying to advise an alliance captain after having been selected.
Just know that, unless you’re going to be an alliance captain, your rank probably doesn’t matter. Also, in a game with such nonlinear scoring (see: gears, climbs) OPR is generally a pretty misguided metric. I’m sure having a mecanum drive hurt you on some peoples’ lists, but if you drove it well it ultimately helped you. Understand this- if you drive a mecanum drive well, people assume it’s swerve and therefore assume you’re good. If you drive a swerve drive poorly, people assume it’s a mecanum and therefore they assume you’re bad. Especially in a game where zone defense was prolific and outright cross-field pushing matches were rare, mecanum wasn’t the liability it was in say, 2014.
My advice- scout yourself. Compare your numbers to other teams at your event, and you’ll always figure out why you weren’t prioritized the way you thought you’d be. It can really sting when you feel like you’ve been underappreciated in the draft, but it’s important to recover quickly, identify ways to improve, and move forward. Morale is a huge deal, especially when you pack a bunch of high schoolers into a high-stress, low-sleep environment for 5 months. Never let yourself dwell on perceived shortcomings.
I would like to start off by saying I disagree with the stigma behind mecanum drive. I can speak for our team in saying that the game determines the drivetrain, for us that happened to be 8 wheel mecanum . I’m not sure how well it is known but our 3 alliance members we played with all the way through Einstein, consisted of 2 robots on mecanums and 1 on omnis.
Some of the most common questions I received in the pits (outside of our shooter) was “Why did you choose to use mecanums?” The simple answer is because they fit the style of play we needed and the benefits far outweighed the downsides. It is all dependent upon the game as to whether or not it should be a concern when choosing alliances.
We use your list except put your first bullet last. We define what the winning alliance is likely to look like in terms of capability to define our strategy, then rank teams that fill in our biggest need (our weakness usually). We also put out team numbers that the drive team liked or wanted to veto as a pick. That was taken into consideration especially when trying to find a second pick when there was little distinction between team X and Y. At that point we do play out what the other alliances will likely look like.
Being ranked fifth at championship we had a Plan A and Plan B. Our Plan A worked out where we picked up a good shooter; but we had a completely different strategy ready to go.
We also talked about scorching the field. We felt our robot and 1296’s robot were equal and the most dominate gear scorers in the top 8. Although we weren’t picked by team 16, we figured we were under consideration by them. We did a what-if where 1296 could have tried to pick us to try and block the 16/4206 alliance (thinking we would decline them). But, we had discussed how that would be the equivalent of scorching the field and give us our best chance to win and we’d have accepted and turn Turing into chaos.
I’d much rather a swerve drive for defense than other low traction holonomic drivetrains. Especially 1640’s CVT swerve - you guys have a great process for iterating and improving it and my teammates and I were thoroughly interested reading the wiki about it.
So I decided to go back and watch your matches because I was also curious after thinking about this more. I will preface my thoughts by saying I agree you were picked lower than you should’ve been.
The first thing I noticed was your auto gears. I may have the numbers slightly off due to just doing it in my head but it looked like you went 4/7 on center gears and 2/3 on side gears (and the 2 were late in the day on Friday). Also the way you held your gear in autonomous made it so you would basically need your pilot up there for consistency, which many alliances won’t want (they want their captain and top pick pilots up there). Correct me if I’m wrong on the pilot thing. The general thought on this is the team doing the center gear needs to be hitting 100% of the time, as the side gears are obviously the less accurate ones for everyone.
Another thing I noticed was a couple of times you had defense played against you in open field, you struggled a little to get around it (see matches against 4939 and 537). And there were also times you used your mecanum strafing in open field instead of turning and using your much faster normal direction speed.
One last thing I noticed was you weren’t hitting 5-7 in teleop, it was definitely a consistent 4-5. Which is still really good, but there were many teams in that area. Also ignore rank and OPR, they don’t really matter when it comes to really evaluating teams. The tiebreak shooter was certainly a plus in quals, but with most alliances in playoffs it was a non factor (either you were trying to outshoot dedicated shooters or you were focusing solely on gears).
Overall I agree with your assessment that you should’ve been picked higher due to your consistency with both gears and climbing, but I’m just trying to give you some more insights as to why teams may have passed over you. Keep in mind many teams at the championship (229 included) can’t afford to bring enough people for quality scouting and struggle to make good picks. I hope this helps, feel free to reply or PM me if you want to talk more.
Team 5801 was in the same boat in Curie. We finished Qualifications ranked 13th with what we thought were very strong numbers and didn’t even end up as a backup (3rd pick).
By the numbers:
Finished 13th (lost last match by 2 points which would have tied us for 8th)
12th in OPR
8th in total match points
Successfully climbed in all 10 matches - 3rd in overall Touchpad score
Average match score of 331.6 with only 2 matches below 300
Completed 4 rotors in 7-10 matches (this lowered our average gear delivery to somewhere between 4-5 per match since we simply didn’t need more)
6-10 Delivering gears in auto (5 from the sides & 1 from the center). 2 misses were IR beam break issues and the pilot not pulling the gear up fast enough (this was fixed after match 3) and another because of an alliance partner hitting us.
Our drive train was a fast 6 wheel, w/2 speed ball shifters and an amazing pilot who played very strong defense. We were only pushed backwards twice - once because we were being pushed by 2 bots and the other because we were floating on fuel. ) It seemed like we caught the eye of the announcer on Curie as we had many mentions during the matches.
Our season was strong, finishing on the winning alliance at the GKCR where we knocked out 1986 in the Semi’s. No I don’t think we could’ve done it again at Champs - we love those guys and they’re simply amazing!!
At North Star in MN, we finished 3rd in Quals and chose a local team not in the top 8 with a big fan base and a rookie team. We weren’t really concerned with winning since we already won a regional, and students thought it would make a fun alliance.
I didn’t really know what to tell the students after we didn’t get chosen for an alliance. We had a great season and even without making the playoffs I’m incredibly proud of the way the team came together! Who knows what next year holds but in our 2nd season, two trips to Champs isn’t bad.
If anyone has any tips for helping us get noticed next season we’d love to hear them. Thanks to all the great teams we met and hung out with over the weekend, we had a great time!
