3 Mogo Autonomous Cagebot - Post Worlds Reveal

Specs:
*12 motor turbo drive, geared externally 5:3
*Fully passive caging

My sincerest apologies to the 10 teams who had the misfortune of being partnered with us, 8452A, in their quals. This is what our robot was supposed to do.

Obviously, I can’t let our 2-8-0 worlds record speak for this bot, so I’ll have to do some explaining myself. There were a few differences between practice and worlds that killed our worlds run, which I’ll explain in a bit, but I’m posting this to show our strategy that was, imo, well thought out and poorly executed, and I’m honestly surprised it wasn’t more popular at worlds.

Our theory of the game:
Suppose that all 3 robots on the field who are not us are the standard DR4B/chainbar robot, capable of scoring 20 cones into the 10 point zone (the highest I’ve seen claimed is 21). Further suppose that our autonomous works correctly, and grabs 3 mogos, ending in a place on the field which isn’t in the way to any robot and ceasing to move for the rest of the match. The opposing (I’ll call them red) teams can score 20 cones on the mogo, and 15 on the stationary goal. They will put the mogo in the 10 point zone, because they cannot put a stack 15 or higher in the 20 point zone. Thus, they will have 10 mogo points, 10 high stack points (10 pt zone + stat goal), 70 cone points, 10 auto bonus points, and 4 parking points, coming to 104 total.

Our partner therefore needs to score all 4 mogos (50 points), get the 20 point high stack (5 points), score 24 cones (48 points), and park (2 points) to get 105 points and win. This may seem like it favors red, as our partner needs to score 24 cones in the time it takes red to score 20 and 15 cones (as I assume each red robot will take one goal), but our partner has no need to stack high. They can make 4 stacks of 6, and therefore face much less risk of tipping, spend less time lifting up and down, and in general, need to worry less. Red needs to play a perfect game - with so few goals, a tiny mistake (like irretrievably scoring their mogo in the 20 pt zone in auto) could easily cost them the game.

Alternatively, a red team plays defense, forcing our partner to score the same 24 cones, while being defended, in the time it takes the red scoring bot to score 35 cones. This is more subjective, but in general, I think this situation heavily favors blue.

Of course, our autonomous can be disrupted. In our tests, we were clear of the opposing tile in ~4 seconds, and clear of a counter-defense autonomous in ~6 seconds. I estimate that if those times were 2 and 4 seconds, we’d be untouchable. This autonomous can be sped up - in particular, there’s a 1.6 second delay after we grab the first mogo for our wheels to hit the ground again, which we estimate could be cut in half. I’m sure there’s other small optimizations we could have employed, too - if we had had more time to work.

So why didn’t it work?
*We have a partner: In order to drive straight towards the opposing mogo, we needed to start in the center of the starting bar, which left no room for our alliance partner to drive perpendicular to us. Thus, they were forced to drive parallel to us in auto, where we would knock cones into their path. Rather than do this, we positioned off center, causing us to miss.
*It’s really fast: This robot was incredibly hard to drive, and because we put so much time into working on auto, we had no driver practice. Thus, we would often stall the drive at the beginning of driver, leaving us useless. Alternatively, we’d get a cone stuck under us, propping us up and keeping us from touching the ground, putting us in the same situation.
*Anti-static spray: Our tiles are 4+ years old, and have far less friction than the brand new worlds tiles covered in anti-static spray. This is a huge deal when your robot is supposed to drag 12 additional pounds around.
*Misc: Our function to drive straight stopped working around our 4th match, and never started working again. Our antitips (yes, our robot tips over when it drives full speed) flipped out unevenly, causing us to turn and run into the starting bar. 3 of our matches, our robot didn’t move because of a corrupted download from RobotC (which has never happened to us before; anyone else have this issue?). One match, we turned on before plugging in, causing driver control to run, changing a global variable which inverted drive motor power.

So in the end, we went 2-8-0, with our autonomous working a grand total of 0 times. I’d personally nominate us for the “Worst in Division” award. If nothing else, though, I’d bet considerable money that we had the fastest HS robot at worlds. I actually can’t take too much credit for the work that went into this robot or the theory behind it (for whatever credit there is to take), much of the work was done by @jwwood13 and @manceslaughter. If you have questions, they are probably more qualified to answer them than I am.

It’s a disappointing Worlds run to end my Vex career, to be sure, but I’ve had a blast along the way. With how long this post is already, I won’t write up a humongous speech like I did for my 1000th post Thank You, but I do want to say that the VEX Robotics and the VEX community has massively changed my life, almost entirely for the better (I’m still wondering where all my sleep went). I wasn’t going to do VEXU, but with the changes released in the product reveal, well, we’ll see. I’ll be attending Duke in the fall, does anyone know about the Vex scene in North Carolina?

I guess to sum this up, thank you for reading this far. I hope you enjoyed; I certainly did.

I theorycrafted this design, but didn’t go with it. Why? Well, I expected that one scoring team on my alliance, my partner, to be heavily defended if we employed this. Can I really expect them to score more than a couple of mobile goals with maybe a dozen cones on them? No? Then it’s hard to justify possibly losing the autonomous bonus, and a robot. Also, I didn’t trust that it would work with anti-static spray on the foam tiles. That seems to be a justified concern, considering how I had to change my normal, offensive autonomous routines when I reached the practice fields on Thursday. Suddenly the mobile goals did not move when I drove into them with my mobile goal lift. This actually helped our consistency, but I had to adjust a few things (it no longer pushed the first cone backward when the drove “through” the mobile goal to ensure a good grab). Also, our drive could only push two mobile goals at once, and about the same speed that it previously pushed four. Moreover, this strategy seems very vulnerable to a well-placed charged autonomous disrupting its access to one or two mobile goals.

Overall, a valiant effort, but we already discussed why I did not go with this strategy. I just posted it here so future generations can get an idea of how to think about defensive strategies. They either have to work well with offense so that you are wasting more of the opponent’s time than yours (e.g. bumping someone doing the autoloader or stacking in general while getting a mobile goal, or zoning opponents out of a corner while stacking) or deny your opponents any chance of winning, assuming you have a passable partner. If going for the latter, like this strategy, you have to sure it will work 99% of the time, at least.

As far as I know there are no VEXU teams in North Carolina at the moment, however seniors from our school have plans to start VEXU teams at both UNCG and NCState University, but nobody that I know of has plans for VEXU at Duke.