Einstein 2014

Competition Rules/Strategy
1

I don’t have any other great method of getting the answers out of those involved in decisions around Einstein 2014, and I feel like if teams who played in the semifinals and specfically the finals match were willing to discuss their thoughts behind decisions made on the field in a public forum, it would be a great benefit to a lot of people. With the 2014 competition season over, I know I am left wit ha few questions.

To 1114:
When did you guys decide you were going to add on a goalie stick?

Why did you add it on (as in, was this specifically for “if we run into the Poofs on Einstein finals”, or another reason)?

How were you controlling your robot during the hybrid period?

Was part of your goalie strategy that you get quick turnaround on cycles coming out of autonomous, or was that a secondary benefit?

To 254:
Why did you stick with the given autonomous strategy after the first match after seeing how close 1114 was getting to you guys?

What caused the misses in autonomous in F2? From my viewpoint I could have sworn 1114 made bumper on bumper contact with you, but I’m not sure.

I think I know the answer to this, but why not give a ball to every team in F3 since all robots were already in the white zone?

Why not accept a possible 15 point deduction on ignoring hot goal detection to avoid the 1114 menace?

Do you think you would you have made those shots in autonomous with or without the pick 469 attempted?

To 973:

Did you intentionally leave the goalie zone, or were you trying to see how far you could push the zone before breaking it, or was it just a glitch?

That’s all I have for now. If anyone has more questions or wants to generate idle speculation before we get answers, I guess they can go in here.

1 Like
2

Instead of directly answering the question, I will let you do some math.

Leaving the zone is 50 pts. In that match the red robots had 1 ball each, 2 total. So we were preventing 20 points at most .

3

I’m not 1114, but http://www.chiefdelphi.com/forums/showthread.php?t=128639

4

G15
During AUTO,

a ROBOT starting in the white ZONE may not cross fully beyond the TRUSS
a ROBOT starting in its GOALIE ZONE must remain in contact with the carpet in its GOALIE ZONE.

Violation: FOUL. If contact with an opponent ROBOT, TECHNICAL FOUL.

G23
If a ROBOT is in contact with carpet in its GOALIE ZONE, and for only one ROBOT per ALLIANCE at a time, there is no height restriction; however, any extension or combination of extensions above 5 ft. may not extend beyond a vertical cylinder with a 6 in. diameter (see examples in Figure 3-5).

Violation: FOUL. If continuous or repeated violations, TECHNICAL FOUL.

I count 90 foul points.

5

This is exactly what our strategy team was thinking about after watching Einstein. We were wondering why Simbotics didn’t come out of the zone a bit more or all the way out with their stick down and hit the Poofs to save at least 20 points of auton.

6

I’ve actually heard second-hand that they used a kinect.

7

In there case the math is different, but I’d wager they still didn’t want to take a tech.

In our case it was obvious mistake and was not intentional.

8

Even better.

9

This is also what I saw from the stands. 1114 had a student standing to the side of the driver station in front of what looked like a Kinect.

10

Since a couple people already mentioned it, I can confirm that 1114 used a Kinect. Four teams on Einstein (1114, 254, 469, 973) had some form of driver station input during “autonomous” mode.

11

And I’ll second this confirmation as I discussed it with several of their students. They had a lot of very cool kinect options for that goal pole as well as for hot goal detection.

If any of you ever get a chance to have a really in depth view of that robot from a controls perspective I think you’ll be very educated. Their students were able to effectively explain everything they did. Kudos to them.

12

This is interesting because it is actually more complicated than that. You also have to factor in the cost of a missed auto ball.

One could consider this ‘ungracious’ or egregious (red card), but it’s technically within the rules (just as famous examples of rule loopholes):

Have I missed something?

EDIT: I think Chris Endres beat me to it.

13

As has already been stated, the actions you could have incurred could have ranged from 20-90 points on a 20 point ball. I couldn’t remember if the robot that 973 was attempting a block on had either one ball or 2 until I saw the video.

There are incidences and specific scenarios where you could leave or almost leave the goalie zone in an attempt to thwart autonomous and it count as an actual net benefit provided you do not get a red card.

However, as always, it’s a pleasure.

14

His point is it would have been idiotic for 973 to intentionally leave the zone in order to prevent ~4x less points than they would have given the other alliance by doing so.

15

I’d like to first say that 1114 having a goalie pole and hybrid auto mode was about as “shocking” to us a our having a low goal 3 ball (neither of us was at all shocked).

Sticking with the three ball auto was a calculated risk. We decided that the chance their actual pole would block the shots to be very low. The danger was all in the position of their claw. The claw came between our front ball and the low goal in F2, and was actually pinning our front intake down in F3 (intake came free in time to score the last ball).

We discussed not going hot before each match and decided that they would likely be able to legally put something (claw or robot) between us and the goal in time.

We also discussed running a two ball mode that lined up on the edge of the field and drove straight (shooting one ball on the run and one near the front of the low goal. This involved bringing 74 off the bench and having them run a 1 ball (these guys had an impressive 10/10 one ball auto in quals). I asked Adam (67) what he thought we should do. His advice was the “safer” choice, bring 74 in and everyone run from the side, still getting all the balls hot.

We thought about doing this, but the last minute change (bringing in a fresh team to Einstein) and giving our opponents the mental edge that’d forced us off of our bread and butter didn’t seem like the right thing to do.

In the practice match on Einstein, we ran 2848’s modified auto. It was programmed to run full speed into the goalie zone and deploy their brake plate. The idea being they could stop 1114, or slow them down, allowing us to gain position to fire the 3 balls into the hot goal.

It was just after sending Dan and my operator back to the field with this strategy that Adam turned to me and said “You guys are nuts. This is chaos. How do you know that they (2848) won’t crash into you?! How do you know they won’t both crash into you?!”

My response was that I trusted our (254) programming team, we’d run 2848’s blocking auto on the field already. At this point my heart was about to pop and everyone near me (Jess B, Adam H., & Adam F) probably thought I was going to pass out.

In the end I don’t know if we would have made the shots w/o 2848’s block, I know we had made and missed once before without it, and that something had to happen to try and stop 1114 from gaining position on us in F3.

16

We talked about a goalie stick during our brainstorming session after kickoff. At one point there were even discussions about goaltending being our primary strategy, as we felt it would be the easiest way to differentiate ourselves from the pack. Ultimately it ended up lower on our priority list, but it remained on our minds.

On ship weekend, one of our new junior mentors spent some time experimenting with the Kinect (see below), and we realized that we could implement a framework that would allow us to have almost complete control over our robot during the autonomous period. Since we saw that the Q&A had made using the Kinect during autonomous legal, we suddenly realized that we could do something very cool with goalie stick during auton.

We had hoped to add the goalie atick in time for Waterloo since we knew we be playing some of the best teams in the World at that event, unfortunately we weren’t able to get things done in time. So it was pushed back to a Windsor/Champs item. After Waterloo work began on the goalie stick, and we even installed the first stage of the stick in Windsor, in the spot where our infamous broken wing from Waterloo went. We were somewhat surprised that no one really questioned this random large piece of square aluminum tubing on our robot. The rest of the stick was completed after Windsor, and added to the practice robot.

Once we realized we could use the Kinect to drive the robot, we knew that a goalie stick would provide an immense amount of strategic options. All year long we saw that missed balls dramatically lowered the scoring potential of alliances, so forcing even a single miss could be devastating. We knew that plenty of teams had multiple ball autonomous modes, so we could go to the Goalie Zone and only cost us 5 points. Very minor compared to potential denial of points from both autonomous balls and delayed cycles. Also, we knew that an effective goalie stick would make us a more attractive partner to a strong team with a multiple ball autonomous mode. (Such as the Poofs)

We used a Kinect located in the alliance station. Here’s a brief description from our soon to be published design notebook.

  • The detected arm position is mapped to a specific location in the goalie zone
  • Range of the arm height is modified from between -1 and 1 to between 0 and 100 for more accurate control on drive movement, which is corrected using gyroPID and encoderPID
  • Balance control using gyro and voltage to avoid unintended turning or tipping and to stay within the 6” cylinder
  • Returns to original position if no arms are detected

Basically the Kinect gave us the ability to drive to any position within the goalie zone. Of course, we knew that many teams would simply line up against the wall as a counter move. Hence we came up with the move where we curved and drove as far out of the goalie zone as possible, and then lowered our claw. The hope here was to block any robots who needed to be right in that front corner.

It wasn’t the main reason, but it was definitely a benefit. We even did this with non goalie stick partners like 5288 in Windsor.

17

Precisely. Neither of us are new to this game, and our teams are very open with each other. (We definitely discussed our Kinect abilities and goalie plans with the Poofs during a social gathering and cultural exchange in Waterloo.)

18

This move in F3 was one of the greatest strategic maneuvers in FRC history. I remember watching the Curie Alliance set up those three robots and telling Ben (1114 Drive Coach) something like “dude, I think they’re trying to block your goalie, what’s our move?” What a mind game!

Far and away the best matches I’ve ever had the pleasure of coaching. Huge congrats to all teams involved, it was a battle to the very end!

-Mike

19

Same story here.

I feel we were the most (or close to it) effective blockers, and this was because we had a mutually beneficial arms race with them all season. Our blocker had the ability to deploy and rotate and block in front of either low goals, all because of the time they put into countering the original blocker threat.

We never ran this feature as we were “Saving” it for poofs, but in hindsight we should’ve deployed it in front of us on 1678’s 2 ball in SF 2.

Being down there talking with Adam and EJ was amazing for that play, trying to offer 254 the frame of mind that we had blocking, to help figure out 1114’s move. I advised their play in F3 was incredibly ballsy and risky, but it sure as heck worked out for them. I’m glad they didn’t listen!

20

I was watching the video Team 20 posted of F2, and it actually looks like 1114 never touched any of the balls. The first ball just shot low, and it looks like the second ball was one of the most unlucky shots this year because it went all the way through, then bounced back out.

But all of the mind games on Einstein were absolutely incredible, and using 2848 to block in F3 was one of the ballsiest moves I have EVER seen in FRC, but it payed off, and that match came down to the WIRE.