The cheesecake runaway

Bagged mechanisms don’t count towards the withholding allowance. Thus bagging two robots would still be within the rules. Even if a team bagged two robots they would still be able to bring fabricated items which would count against their withholding allowance.

You’re ignoring the fact that this ruling was made after the end of build. Cheesecake was effectively limited to exactly those items. The question is what the ruling is going to be next year when a team bags cheesecake with their robot and then wants to put it on another robot. All the arguments in favor of cheesecake are still going to apply to shrinkwrapped cheesecake, so why would the gdc rule against it if we weren’t discussing the implications now?

Not sure if this would hold up. We used two bags for our robot this year at our first event with our robot in the first bag and in the second we had two of our spare upright tubes (2" square aluminum tube .063in wall 72" long) along with a few smaller miscellaneous items. All in all it weighed under 10lbs but we put it in the bag in case our witholding needed to be larger than its current 15lbs that we brought in.

As we left the event the inspectors saw it differently so we went back to one bag and used these parts as our witholding allowance since we didn’t use much from event to event.

Pretty sure your inspectors were wrong. The withholding allowance rule says you have to bag all robot elements, including those used in alternate configurations of the robot. Admin section 5 also says you’re allowed up to 2 bags in case you need to disassemble your robot. So I don’t see a problem with two bags, and I don’t see a problem with bagged robot elements. If we’re nitpicking that the bagged parts aren’t actually part of the robot because they’re spares, that seems perverse and unnecessary.

This comes up all too often. I didn’t agree with it either but it came down to who’s interpretation of the rule would be followed at which point we followed the instructions of the LRI and didn’t go about it during the following events.

Fair point. Now that you mention it, I’m not sure why teams are allowed to bag as much as they want on Stop Build Day. Would there be opposition to limiting this to 150 lb? 180? Can’t imagine many teams bagged more than that.

I think part of the solution can come from the definition of COTS and what vendors provide.

Can I give an assembled gearbox to another team? If AndyMark sells it assembled with popular modifications, no problem.

I expect that this whole situation will lead to new rules that won’t be particularly popular. The game manual is full of rules that have similar origins. Many people blame this game for driving the extreme cheesecake, but teams will find uses for it in all future games.

At Championship, the extreme examples were made more feasible by: (1) the long delay between the end of qualification matches and alliance selection and (2) having a 4th bot on the alliance that could be kept off the field for modification. Item #1 could be addressed by having the final qualification matches, alliance selection, and playoff matches on the same day (not something I would favor). Item #2 could be could be addressed by going back to 3-team alliances with a backup bot list.

I like your suggestion of limiting hardware changes after qualification matches are complete.

I don’t understand why this is such a big deal. Teams shouldn’t be unable to help alliance partners due to the rules of the game.
The rules of the game need to not necessitate such extreme measures.

I read the rules on kickoff weekend and knew that things like this would happen. It wasn’t hard to foresee. Don’t make the game come down to a challenge of whether or not a team has resources. Plain and simple.

I know, I was there for both of those regionals, 2512 is one of my teams closest friends, they are a great group. I am by no means saying they had a bad robot, quite the opposite, 2512 just picked up their game enormously as soon as those can-burglers appeared. All they needed was that extra little push to reach their max potential (they were defiantly an important player in that 290 all-time score).
When I started this thread I intended it to be used to discuss how much cheesecaking effected a teams performance in a vastly positive way, and how the first community felt about extreme cheesecaking.

As a student on 1923, I am disturbed by the inaccuracies of this post. Let me shed a little light into our perspective. Yes, 1923 and 900 got cheesecaked, but you can be hella sure that students and mentors of 1923 and 900 not only had a say in the cheesecake, but contributed significantly to the design and the build of them.

If you look at the video that 900 posted with the GoPro feed, you can see that two FIRST teams are working harmoniously together to do what this is all about: build robots and amaze people. Teams like 1114 and 148 give you a model to look up to and gave us inspiration to do better.

As human player, i could not believe the amount of help and courtesy I was shown when 1114 and 148 came to our pits, when we competed with them, or anything. They didn’t become good by being arrogant people who thinking that their design is better, they became better by absorbing and emulating FIRST ideals.

I have been so inspired and motivated by them that I don’t even care that we didn’t win finals. Their significant contributions and help showed me that FIRST still has more to give me and that this truly is an amazing program.

I also wanted to say that the students of 1923 and 900 helped this alliance both on and off the field. When 148 came to our pits, they told us what they had planned, listened and accommodated to our needs/wants and asked courteously if we were okay being cheesecaked. As a student, I can tell you that these two teams have taught me more than I could ever hope to learn in just one season.

Just to summarize,

  1. 1923 was involved in the design process as well as the build process for any cheesecake done to our robots
  2. 148 and 1114 are the most inspirational teams I have ever worked with or even seen at an event.
  3. Cheesecaking allows FIRST teams to not only have better robot competition, but also allows teams to learn from other teams and get inspired
  4. Being graciously professional involves working with other teams to learn and succeed. I learned a ton, and succeeded.
  5. I looked around me standing in queue in Einstein, I saw a happy drive team from 900, saw my team extremely excited and crying out of joy. I think that 1114 and 148 not only donated their resources and time to two teams, but made us realize the potential that we all have.

Your assumptions only put blemishes on an otherwise perfect experience for me, and for my fellow students.


So I have a couple things about this whole cheesecaking thing, especially after going to worlds.

First, I believe there are two ways to cheesecake. One, where a team adds parts/mechanisms to another robot, but the robot still serves an important roll besides just having those parts (usually can grabbers or ramp) Or two where a team adds parts/mechanisms to another robot, but then is told to after using those mechanisms to just sit in a corner so they dont screw anything up.
Number two is where I have a problem with cheesecaking. If you pick a team and you know they are that much of a hazard to the alliance then you should have just picked a different team. Do your scouting and know which teams are reckless and topple stacks or which teams have careful drivers who are precise.

Second, for our team going into alliance selections on hopper I could pretty much guarantee, being ranked 52nd, that if we were picked we would be cheesecaked in some fashion. Being picked by 987 we knew they had can grabbers for us. And the only thing I asked was to involve our students somehow in the process. I know our team wasn’t going to do the whole thing, but with our capable students in the pits I thought we should utilize all of the available resources. We then collaborated with 987 on how to best attach the grabbers and then proceeded with the process. Overall I thought our alliance handled it all very well, and was very pleased with the balance of work being done by all the teams.

Putting aside the discussion about whether or not cheesecaking being allowed is a good or bad thing, just for a moment… I think calling for a rule to regulate or ban the practice is a huge overreaction that’s just going to result in some overzealous volunteer somewhere disqualifying some alliance for upgrading each other’s robots based on the subjective interpretation of an imperfect, knee jerk reactionary rule.

Instead, we just need to realize that there’s a reason this behavior was so strongly emphasized this year. It’s the game design. We just need the GDC to never design a game with this perfect storm of unique attributes again:

  • Extremely critical task that is difficult to accomplish
  • Strong incentive to “race” to complete this task first
  • Chokehold strategy present with successful task completion
  • No defense
  • To a lesser extent: Cluttered field with lots of areas for congestion

When you have this, you’ll have essentially mandatory cheesecaking if you want to win the world championship. While there was some upgrading in 2014 to get robots to provide assists, and in 2013 to block frisbees, all of these upgrades were fairly simple, limited, and things that teams could probably have done with all of the parts they had themselves lying around their own pits. The biggest instance of “cheesecaking” to the same scale and importance I can think of in the past was actually 2011, with the minibot race. My team went to an event where the event winner survived a scorched earth alliance selection by picking the best tube scorer in round 1 (despite lacking a minibot), and a kitbot in round 2 that they could mount their minibot and deployer to. A few alliances at the Championship played with placing the spare minibot ramp of the fastest robot on the alliance onto a second alliance robot.

If you look at the list of attributes above, you can see why it happened in 2011 as well - a chokehold strategy was present with the minibot race (unbeatable score), there was a strong incentive to do the task best / first (denies points to other alliance), and the task was extremely difficult to do well. Can we eliminate these features from our games? Then we don’t have to write some complex or subjective rule to eliminate a behavior that at least some people really don’t like. We can sidestep this discussion entirely by just not playing with these terrible game mechanics.

Looking Backward, I can’t think of a game over the past ten years that didn’t have these properties.

Extremely critical task that is difficult to accomplish - 2014, 2013, 2012, 2011, 2010, 2009, 2008, 2007, 2006, 2005
Strong incentive to “race” to complete this task first - 2014, 2013, 2012, 2011, 2010, 2009, 2008, 2007, 2006, 2005
Chokehold strategy present with successful task completion - 2013, 2012, 2011, 2010, 2009, 2007, 2005
No defense - 2013, 2012, 2011, 2005 (in the form of protected zones)
Cluttered field with lots of areas for congestion - 2014, 2013, 2012, 2011, 2010, 2009, 2008, 2007, 2006, 2005

Not clear on what you determine the “chokehold task” to be besides scoring points in 2013, 2012, 2010, 2009, minibots in 2011, and 2 super-long chains in 2007, but none of these are things that could easily be “cheesecaked” onto a robot like can grabbers this year. I can’t comment on 06/05.

Also in a large majority of these games, if 2 of the robots have a majority of the offense under control, there is something for a 3rd robot to do to keep their “offense area” less cluttered:

2014: Run interference/defense while cycle completes. Also be part of a 3 assist cycle.
2013: Run interference/defense while allies cycle, and hang at end. This game is the only one close to the level of “cheesecake” of this year at all because of plywood 10 point hangers and full-court blockers.
2012: Run interference/defense and participate in balancing.
2011: Run interference/defense, steal game pieces, funnel game pieces from midfield to scoring area to make allies more efficient. Possible “cheesecake” minibot/launcher here.
2010: Run interference/defense. Requires being able to expel balls from the far zone, but you don’t have to be the fastest, so you pick a team that you know can kick the balls out, and not try to “cheesecake” a premade kicker IMO.
2009: Pin opponents, and keep your trailer out of the way.
2008: Lap lap lap lap lap
2007: Run interference and participate in endgame

2015: Attempt to grab cans that lose you eliminations if you don’t have them, don’t knock over stacks, attempt to use any of the limited game pieces the first 2 offensive robots aren’t using.

I need more explanation of each of these because I’m not seeing how they fit the these definitions. Except for the 2011 endgame and the 2010 chokehold (that was eventually defeated), none of these are obvious. Defense played a big part in whether teams could get to safe zones and even then shots could be blocked. Cluttered and congested are two different things. I don’t think anyone would call 2014 cluttered or congested except with other robots–the nature of 3 team alliances.

I would think that the bottom line thought process for cheesecaking is this:

  • Something needs to be done that is very, very important for the success of the alliance.
  • That thing cannot be done by me because of whatever reasons. I’m too busy doing some other very important thing.
  • Therefore, I’ll help someone else to do it.

This is a subtly different thought process than the following:

  • Something needs to be done that is very, very important for the success of the alliance.
  • I could potentially do this, but I don’t know how.
  • Therefore, I’ll ask another team to help me do it.

Many of those years had a fairly wide open field (2006, 2008, 2009, 2011, 2012, 2013, 2014)
Unless you’re talking about poof balls and track balls littering the field?

I think you’re really stretching the limits of what each of these parameters describe to prove this point. To save time writing this post, I italicized all the years that I feel do NOT apply to each parameter.

Extremely critical task that is difficult to accomplish - *2014, 2013, 2012, *2011, 2010, 2009, 2008, 2007, 2006, 2005

In 2014, the only “extremely critical” task was possessing a ball to get an assist. And driving, I guess. Neither of these were very difficult, and any “cheesecaking” needed to get them done was usually pretty crude and something the team could accomplish themselves - no need for pre-built mechanisms from home to be bolted on. You certainly wouldn’t pick a kitbot to cheesecake over a capable third robot.

2013, what could you argue was extremely critical? Receiving frisbees from the human player, I guess? If you could only play defense, it wasn’t so critical where people were putting a complete shooter on your robot for you. You could argue 10 point hanging, but I would argue the passive hang was trivially easy that year to add to a robot that otherwise does nothing.

2011’s minibot race is an example of an “extremely critical” task - if your alliance didn’t have two minibots, at all but very weak matches (or with the absolute strongest tier of scorers) you were unlikely to win.

Strong incentive to “race” to complete this task first - 2014, 2013, 2012, 2011, 2010, 2009, 2008, 2007, 2006, 2005

None of those games have a “race” component, or any incentive to do something first, except 2011. If it takes a robot 4 seconds to complete a task versus 2 seconds, it’s slightly worse, yes, but it’s not that big of a deal unless the task itself is a race. The key word here is “race”.

In fact, the opposite was true in 2005 - you were trying to be the LAST team to score on each goal…

Chokehold strategy present with successful task completion - 2013, 2012, 2011, 2010, 2009, 2007, 2005

You’ve got a really strange definition of chokehold strategy. 2013 did NOT have a chokehold at all. 2012 had game piece recycling making a true chokehold almost impossible. 2009 did not have any chokehold strategy. 2007 had an element where if you scored enough tubes in the right places, you could guarantee a 60 point lift would beat any pure tube strategy. But pure tube strategies were extremely rare, and you didn’t have to modify alliance partners to lift them. 2005 did not have a chokehold that I am aware of.

2010 had a chokehold-esque strategy in the 469 type robot, but this could not be “cheesecaked” onto any old kitbot. A regular deflector that could get the balls in the same zone could be, but since that’s just a sloped flat piece of material held in the air, I’d hardly compare it to 2015 or 2011.

2011 had an achievable “unbeatable score” if you could guarantee first and second in the minibot race, very similar to 2015’s unbeatable score if you guarantee seven cans.

No defense -* 2013, 2012, *2011, 2005 (in the form of protected zones)

That’s not what I meant, I meant this as in “defense is not a task a third robot could go and do”. There was plenty of defense in every one of these years except 2005.

Cluttered field with lots of areas for congestion - 2014, 2013, 2012, 2011, 2010, 2009, 2008, 2007, 2006, 2005

Are you seriously going to argue that 2014, a completely flat and empty field spare four goals in the corners, is “cluttered”? I’d love to see your definition of an empty field then. 2006, 2009 had similarly open fields.

The point I was trying to make here is, an overwhelmingly important race-type task that must be completed to win at the top level combined with a relative lack of better things for the robot to do and a field too congested for three robots to score independently constantly means that there’s simply not a lot else to do with that third robot but cheesecake them. In 2015, there wasn’t a lot else for your third robot to do anyway, unless you got some steal of a draft pick like 1671. If there is a task that many robots can do that can still contribute to top alliances, such as fetching / feeding game pieces, playing defense, etc. there isn’t a strong incentive to cheesecake unless the task is of gargantuan importance like the minibot race.