How to Win Worlds Guaranteed! (Why Game Analysis is the most important step in Robotics)

I lied, reading this will not guarantee you a World Championship title. It can however, get you MUCH closer than you’d get without it. The Clickbait title was required to attract all the young and impulsive members of our wonderful community already on their 3rd DR4B Rebuild for the season that are about to trash it and copy either ALBA, 8059, the California RI3D, or whatever SingVex Leaks show up in the next week. Hello young competitor, set down your caffeine and 3/32 Allen wrench for a moment and take a seat. If you’re wondering how I’ve read your mind up to this point, that would be because I was you 3-7 years ago today. Life is funny like that.

Alright down to business, I was originally going to make a short post in this thread (not because ALBA did anything wrong, in fact, they’ve done an absolutely magnificent job, however their reveal did cause me to realize that a lot of teams are going about this season the wrong way), but it got long and off-topic, so now we find ourselves here. The most important thing you can possibly do to be successful in VRC, real engineering, life, or whatever, is to have a plan before you act. Many teams do not do this. This is not an issue unique to rookies or teams of any skill level. There are Round Robin teams who don’t do any form of proper game analysis before they start building, and while they made it on the big screen, they had a heck of a time getting there and went through a lot more trouble than might have been necessary if they had really thought things through beforehand. I would credit the Design Convergence that has plagued our competition from year to year to a lack of proper game analysis and forethought before the average team starts building. Now for the single most important thing anyone will ever give you during your time in VEX:



This is literally the man who designed the game telling you how to analyze the game and design your robot, which will then play the game he designed. Watch the entire thing and take notes. Do this multiple times.

If you still don’t believe me, let me give you these links:

http://www.simbotics.org/resources/strategy/strategic-design
https://www.team254.com/documents/fundamentals-of-game-analysis-and-strategic-design/

These would be two of the greatest teams in FRC history telling you how they go about building robots. Surprisingly, it’s the exact same process…

Before we apply all this wonderful knowledge to VEX In the Zone in a moment, I need to rant about something (as if this post isn’t one giant rant already ;p). The word “Meta” annoys me to no end. After Starstruck, a game with a painful amount of Design Convergence, there seems to be this prevailing idea that there is one perfect design for any VEX competition season. Sometimes this is true, but most of the time it isn’t. A lot of teams may end up with one design and thus believe because everyone has it that it must be superior, but I would argue the reverse, that it ends up being superior only because everyone has it. More teams trying to perfect one idea and bouncing improvements off each other is going to lead to a more finely-tuned robot in the end, but it certainly doesn’t make that specific robot design superior to any other. It only means the other options haven’t been explored as thoroughly. I’ll remind everyone that the World Champion Alliance for Nothing But Net had a robot with a double flywheel, a robot with a single flywheel, and a robot with a linear puncher and stack bite combination, all three designs bringing something different to the alliance and covering each others weaknesses. There was no specific robot design that was inherently better than the others, robot performance was mostly dependent on how much work had been put in. The Skyrise World Champion Alliance saw two double-reverse linkages and a scissor lift, two pneumatic needles and a hook conveyor, two Skyrise builders and a cube bot. Three very different robots. Starstruck was literally 90% rear-dumping clawbots, yet two of the World Champions were forward dumpers.

Time to apply this to the present. I’ll get super specific, and then pull back to a more broad perspective. This robot already stacks 7 high using only its chain bar swing arm with no traditional lift system (%between%, thanks to ALBA for inspiring this post with their reveal). It also has two motors currently unused. If you mounted the swing arm on a vertical rack and pinion (that would only take two motors and be literally the simplest lift design possible, how convenient), you’d pick up a little less than 18" of lifting. Each additional Cone adds a little less than 3" in height. ~18" / ~ 3" = ~6 Cones. Now you can build a 13 stack. If you build four of those stacks and push them all into the 5 point Zone, no goal lifting required, you are almost mathematically unbeatable, even if you lose auto and all four highest stack bonuses. If we can get one of those four Stacks into the 10pt Zone and park a robot, we win every single time. No matter what.

Scoring Breakdown:
Four Stacks of 13 is 52 Cones. That leaves the opponent with 28 Cones to score.
104 - 56
We now push three goals into the 5pt zone, and one into the 10pt zone. They put one in the 20, two in the 10, and one in the 5.
129 - 101
Assume they won Auto.
129 - 111
Now it gets close. They need to capture all the Highest Stack Bonuses to win while expending minimal resources, so they put a single cone on their 20pt Goal and their Stationary Goal. One of their 10pt Goals is completely empty. We have a 13 Stack in both the 5pt Zone and the 10pt Zone and they need 14 points to beat us. The problem is, there’s only 26 Cones for them left to score. if they split it in two 13 stacks, no one gets the remaining bonuses and we win. If they want to win either bonus, they must concede the other to us, they cancel out, and we win. The last ditch effort is to forget the 5pt Zone Highest Stack and pick up the extra 5 points for moving it to the 10pt Zone. This gives us the Highest Stack for the 5pt Zone and the opponent the other three with an extra 5 points for the last goal.
134 - 131
All they have left is to park. We park one robot.
136 - 135
With four Stacks of 13, three in our 5pt Zone and one in the 10pt Zone, and a single parking bonus, we cannot lose. This entire strategy also negates the fact that we can actually build 7 Cones on to the Stationary Goal without having to go above a 13 tall Mobile Goal stack, giving us an even greater lead.

In reality, it’s going to be impossible to capture 39/52 contested Cones against a good opponent, meaning you won’t be able to build those four full 13 stacks and our unbeatable strategy is no longer full-proof. But we’ve proven that you never need to go higher than 13 Cones, and you can do that without an overtly complex double-reverse linkage. In fact, you could do it right now with a swing arm and some linear sliders. The 13 Cone limit is the magic number, once you can go that high, you are as tall as you’ll ever need to be*. If you believe the DR4B is really the way to accomplish that, be my guest, but I think a lot of teams will find that they can build well within their means AND be extremely competitive at the same time. If you take a moment to firmly define the constraints for your team’s robot design and determine EXACTLY what you need to do to be successful, you’ll save yourself a lot of hassle. There are plenty of other ways to formulate an unbeatable scenario based off different goals and robot abilities. My main constraint was not building higher than 13 to avoid building an unnecessarily tall lift and my analysis was based around using almost exclusively Cone Points to prove this magical height of 13 cones. Do I think the swing arm on linear slides is the world champion design? Maybe, maybe not. Swing arms have an excessively large range of motion which will tend to increase the time it takes to score Cones as they must travel farther, however I URGE all the teams about to begin their journey in the VRC to design around a strategy, instead of building your strategy around a design. The constraints define the robot, not the other way around.

Good luck to you all. Maybe I’d post more often if I was capable of writing under a thousand words XD.

*I’ve given a pretty long explanation about why 13 Cones is the magic number and why we’d never need to go higher than that. However, keep in mind VEX reserves the right to change the number of Cones and Mobile Goals on April 5th, just a few weeks before the World Championship. Do your own game analysis, don’t just use mine. And watch the videos. There’s a lot more to analyzing a game than just what I wrote about. This post explores a single section of Defining the Problem, which is only a single piece of the Design Process.

*I wrote this at like 4 AM because I am insane. If I missed some math somewhere, let me know and I’ll fix it / reanalyze it. My main points stand.

1 Like

Wow…

ThAnKs F0r tHe NEw MeTA MAn!!!

But no really this is gonna be a huge help for a lot of people i appreciate it.

You spit facts max

Good timing and well said. Thanks Maxx. I still owe you big time. Just let me or Sam know if there is anything you guys need. I’ll try to make it happen.

@400X

I think you forgot about the match loads :stuck_out_tongue:

His analysis assumed 80 cones, which is correct. Are you talking about when he said 39/52 contested cones (i.e. not matchloads) or an I missing something?

Personally, I’m building a DR4B because I think there will be cones added for Worlds and I want to be prepared.

I totally thought that their were 80 cones on the field + the match loads for a second XDDDD brain fart.

Anyways i do see one problem with this and that is time. To do this by yourself you have to score one cone every ~2.0 seconds (Not including auton) which i see as impossible and to do it with a partner you have to do one in ~4.0 seconds (Not including auton). which i still don’t see cycle times being this fast. I think the next best approach with this style of robot is 1, heavy cone defense, and 2 scoring mobile goals in higher zones. I do like the idea of going 13 high with a simple and light robot though.

Also I really hope they do something cool like add more objects near worlds, maybe even a new way to score it would switch the game up a lot.

I have already made a double reverse 4 bar but this thread is making me question my approach to the game.

I could see 4-second cycle times, to be honest. That’s why I believe more cones will be added at Worlds.

Note also that you need to get 52 cones before your opponents get 28. That seems unlikely to me.

I’m not trying to dissuade anyone from building a DR4B, but in my opinion, one of the best parts of VEX is the ability build upon past designs and change your designs throughout the season. Building for a game that may not even happen several months down the road seems like a poor use of time to me. When you could be building and optimizing for the current game, you are building towards a potential game. I’m not going to go into this, but ~15 cones is when the stacks start becoming too unstable. Also, while the GDC reserves the right to add/remove cones before worlds, I really do not see much space on the field for the addition of cones. Additionally, changing one of the most fundamental aspects of the game a week or two before the world championship also seems like a bad idea, potentially ruining numerous teams’ autons, strategies etc.

I am pretty sure there are already robots that have 2 second cycle times for cones right? Or am I missing something?

Why does an elevator lift need 4-6 motors? It’s lighter than a DR4B by far, because it only lifts 2 slides ,the claw and a cone instead of 14 c channels, a claw and a cone. Also, most DR4Bs built in the past used 4-6 motors, not 2.

Honestly, a post on this forum hasn’t made me smile this much since that time someone asked if they could tape boxes to their shoes. There’s a lot of great advice in here and I suggest you read it and absorb it. No need for me to repeat anything, since 400X has done such a great job of spelling it out clearly and providing easy to digest examples.

2 Likes

Maybe for Matchloads.

That was my favorite post ever on the forum, so funny.

If there is a robot that can hold a constant 2 second cycle time for 1:45 then please share. I don’t think it is possible to do that.

I 100% agree with the sentiment from 400X. Not sure I completely agree with the game analysis, but they are definitely correct about how to approach analyzing the game and thinking things through before jumping in.

Karthik is right, I haven’t seen a thread on the VF this productive in a long time, I look forward to reading the discussion here.

Not for 1:45, for certain. I do think teams could approach or meet this time for matchloads, however (I hope to!). Although, in a way, it probably won’t matter how fast we all do matchloads, because we will be done with field cones and have nothing else to do after the matchloads are done but park, or maybe move things around.

Obviously asking silly questions on the forum is the first step to winning worlds. Also, you never answered the follow up question :frowning:

This post is good, it sums up what most of the top teams will be doing at this point in the season. If you don’t know how you are planning on winning right from the start, you probably aren’t going to win. I think this shows why having the “best” robot sometime isn’t enough, it’s so important how you use your robot as well.

It was a perfectly reasonable question when you’re as tall as you were back then. [https://vexforum.com/t/answered-high-shoes/18730/1)

In relation to the thread: yes. Actually looking at game analysis first is a huge area that teams miss out and I’m glad it’s been mentioned. If you have a good plan, you can design and build to that plan. If you have a good robot, but no plan, you may get lucky, but over all you’re not going to be one of those top tier teams.

Yes i agree match loads will be done at about that rate. I don’t think that the field will get cleared very often though, It still requires an insane cycle time and with the cones spread out and other robots playing it will get increasingly harder as more cones go away. I also do not think that a 4 second cycle time can be achieved for this amount of time. Of course the speed of match loads will give you a bit more time to do more field scoring. The only cases i see this being possible is if you get lucky and end up being able to maneuver in an extremely optimized path. The way i see it is with picking up mobile goals, releasing them, and all the driving (and some possible defense on you) all on top of the ~4 seconds cycle time makes this idea seem a bit far fetched. Just my thoughts though.

This game is going to be interesting, and i do believe a modified version of this idea can be made to win like @400X was talking about, and this is exactly why i think that they posted this.

I would love to see someone make this cycle time work because it seems right now like a very impressive feat to accomplish, if even possible.