The cheesecake runaway

I am starting this thread to discuss how cheesecaking can affect the success of an alliance to an extreme level.
Hopper Division Team 2512 (Duluth East Daredevils) had an OPR of 32.5. This is relatively low, and it showed, as 2512 finished with 52 ranking.
Then their luck turned around, picked by the number one alliance of 2826, 987, 4265 picked them. All of a sudden 2512 gets subbed in for 4265. And with cheesecaked can-burglers from 987. These grabbers are not the fastest at worlds, but they are capable of out grabbing 1114 and many others.
After all this 2512 had a VASTLY improved OPR not just due to the can-burglers, they found the sweetspot and everything clicked. So I’m starting the discussion: As far as cheesecaking is concerned, what other teams need that extra little push from other teams for their bots to reach max potential? And even more importantly, what can be deduced about the first community due to cheesecaking?

Congrats to 2512, 987, 2826, 4265 from your friends at 876!

Now that the championships are over, I definitely think that FIRST must implement rules about cheese-caking other robots. For example, The top seed alliance of the Curie Division (1114 Simbotics and 148 Robowranglers) picked two other teams (1923 and 900), and didn’t even play with them in playoffs. They faced off with just the two robots while the mech team from 1114 and 148 worked on attaching ridiculously good burglars to 1923. By the time these modifications were finished, it was the finals of Curie Division and they brought out their third team to take the division finals. After this, they set about disassembling the entire drive-base and structure of team 900, to attach four of 1114’s ridiculous harpoon guns so that they would be in the weight limit. Now while I have no problem with 1114 and 148 having great mech teams that can build this machinery, I think it is not in the spirit of FIRST to ask other teams to change their whole robot with something they prepared earlier to ensure their own victory. It does not embody gracious professionalism, does not enforce the idea that FIRST is “more than just robots”, and does not encourage problem solving skills or strategy, its more like “We are a great team so lets just basicly build two robots that make the perfect alliance and win championships”

In short, I think that there should be some ruling that says, “The robot inspected on the first day must be the same robot as enters the field”, with the job of deciding what the same robot is up to the judges. Obviously new parts and innovation must still be allowed, but not all of this crazy 4 harpoon gun tethers and stuff.

This is because of the game design of Recycle Rush, really

I agree with this, most years aren’t like this, they happen every 4 years or so; i.e. 2011 and minibots.

Yeah, this year’s game has many more aspects to help with.

You raise some interesting points. I think it would be informative to hear about the “cheesecaking experience” from the point of view of one of the receiving teams.

My bet would be that the experience that teams like these received was more inspirational than the experience they would have had either 1) making minimal contributions and/or sitting out entirely during eliminations or 2) not getting picked at all due to limited capabilities.

I think it’s also useful to frame this within the context of the teams “serving” the cheesecake - several of them are hall of fame teams who have been recognized for their outstanding programs and serve as role models in FIRST by embodying the “more than just robots” philosophy.

I can only applaud teams for their ability to work within the rules and come up with both effective and creative cheesecake recipes for Recycle Rush. 1114 and 148 had two of the best, and I wish we had been able to see the harpoon guns on the field.

That being said, I think and hope that in future seasons cheesecake will not be allowed to this extent. Teams should be chosen based on their ability and fit within an alliance…not literally the fit of a premade mechanism within their chassis. Seeing less capable teams chosen over more capable teams solely to operate a foreign subsystem disincentivizes striving for excellence for a large part of the FRC population, and instead encourages things like making flyers advertising how easy it is to commandeer their robot.

Again, I do not mean to slight any cheesecake givers or receivers, as they have acted within the rules the same way I would have (and did). But there needs to be some sort of limit. Having a game that didn’t depend so much on a single sub-1 second function would certainly help.

I feel as the concept is okay when it involves modifying the other robots in the alliance to work better with one another in a way such as adding complimentary ball intakes in Aerial Assist for easy passing or a device to allow the scoring of coopertition totes. But when it reaches the point where the recipient is pushed into scrapping their entire robot that they worked on for the whole build season for the sake of improving the chance of victory for the alliance captain, then it becomes a real issue. In this case, victory is put ahead of the values of FIRST as cheesecaked members of an alliance effectively have to work around their “benefactors” rather than with them.

You could very well say that playing the 2 v 3 match was also 148 and 1114’s biggest flaw. They didn’t qualify for the Einstein finals by missing out by 1.66 points in their average score. Which means if they had scored 5 more points in any of those 3 matches, we would have a different champion. If they made the decision to let their 3rd alliance member work on stacking from the human player station there would have been a different result. While I agree that FIRST must find a way to limit cheesecaking, cheesecaking does also present a unique element of strategy.

I have a feeling that we will never see a game that matches are decided within the first second ever again.

Everyone loves cheesecake, but when you can no longer love your bot because of cheesecaking, that’s where I draw the line.

We all know it is a competitive atmosphere out there, even if it is “more than just robots”. I want to win, you want to win, we all want to win, and some teams will go further then others. I am sure that if I found myself in the position, I would cheesecake the heck out of another bot (or my bot for that matter) if it meant a trip to Einstein.

Well I feel like in the long run this hurt 1114 and 148. You couldn’t win recycle rush with just having two bots do everything on the alliance. It is about the alliance flow. 118, 1678, 1671, and 5012 won because they had three bots contribute a relatively “even” point distribution. When each bot only has to worry about getting two capped stacks of 6 or 5 it allows them to focus on the task, not feel rushed, and focus on being consistent.

While the level of cheesecake may or may not have been too much, in the end it was about building an alliance that had the best flow without cheesecake.

I’m torn. A major part of the FIRST culture has always been help each other out. Helping a team add a new component to a robot certainly feels like GP in action. learning directly from the approach of an elite team is also likely a great experience for those on the receiving end.

On the other hand, the logical conclusion of all this is to simply build a very light kit bot with plenty of mounting options, and provide lots of raw material for your patron team. While it is doubtful that a team would set out with this goal at the start of build season, I can definitely see how a team could look at their rankings at some point during qualifications and simply decide to “cut bait.”

And, there’s the question of who keeps the resulting robot? Is the harpoon contraption 1114’s parts, or is it part of 900’s robot now? This seems very different from lending a battery.

This isn’t meant to single out 1114 or 900 in any way at all. Their’s is simply the most famous example now.

Maybe a good compromise would be a weight limit for cheese cake parts, like 20 lbs(I randomly used 20, could be more or less depending on the game or actual analysis) of cheese cake, so it is still mostly the same robot with cheese cake added and not a completely new robot.

That said, I love the cheese caking this year (and not just because it is cheese related and my teams sole purpose is to find every cheese pun ever). It really raised the competitive level for teams. I would just like to see some actual rules in place for it, before it gets to a point where teams are essentially walking in with their own alliance partner, and randomly picking a team to drive it. I would be incredibly disappointed if FIRST got rid of it though, since it does have a positive impact.

(This is my personal opinion)

To clarify what 1114 did with 900, 900 built a kitbot chassis on Friday in order to have a light enough robot to be cheescaked, this obviously made them a viable pick despite their low ranking in the subdivision. This had more to do with 900 making a decision that paid off for them than 1114 and 148 asking them to do something they may not have planned on.

Also, the majority of the cheescaking that I saw happening this year was not very elaborate, most of them being ramp bots or adding a simple mechanism for canburgling. It was not until champs where I saw more elaborate cheesecake recipes, such as 1114’s harpoons.

Which further emphasizes the point made by grstex regarding teams looking at their ranking and “cutting bait” as he puts it.

But as others have pointed out, a lot of the problems here stem from the nature of Recycle Rush. It’s a game where 2 top-tier robots can score almost all of the totes by themselves and don’t need a third robot, and where it is incredibly easy to bolt on incredibly important and impressive add ons to an otherwise average robot and take that robot from average to “the best” at something (either a canburglar or ramps for the feeder station, your choice).

In 2014 you couldn’t bolt on a manipulator to handle the ball. In 2013 you couldn’t bolt on a shooter. You could bolt on a 10-point hang device, but that wasn’t as important as the canburglars this year. And the same can be said for 2012 with regards to shooters (not possible) vs. a “stinger” (possible, but not critical). The last time we’ve had something where it was relatively easy to bolt on an extra device was 2011 with minibots, another year that had diminishing returns in scoring and the top-tier teams could do all of the “worthwhile” scoring. And that was/is one of the biggest complaints for that year. Matches were decided before they started based on the minibots.

I am hoping that FIRST realizes this when designing games in the future and stays away from these problems.

It did also allow teams to add things they did not have weight, which is pretty much what happened with all the ramps.

I think that cheesecaking to the excessive extent which was common throughout this year can be solved mostly through smart game design. Here are just a few factors off the top of my head which made recycle rush uniquely suited for cheesecaking:

  • An incredibly cluttered field, which made it hard to effectively utilize three scoring robots without getting in each other’s way
  • No defense, which necessitated that a third partner contribute to the alliance through means other than their drivetrain, generally a crucial, and very integral, element of third partners.
  • Three distinct, moderate/high difficulty autonomous tasks (20pt stack, cans 1/2, cans 3/4), all with extremely high reward, which were very hard or impossible for a single team to do more than one.
  • The ease through which canburgling could be done through an “auxillary” mechanism, rather than something more integral to the design
  • The transport configuration rules, and the extreme flexability they gave teams in these types of “auxillary” mechanisms. (see tethered ramps)
  • The extreme strategic importance of a task which didn’t immediately draw attention to itself, and was not initially focused on by a majority of teams.
  • The fact that this importance varied to an extreme degree with the level of play, to the point where designing entirely around can grabbing was likely not a smart choice for a low resource team
  • The “arms race” nature of the task, with continual dramatic redesigns being a requirement to remain competitive.
  • The fact that canburgling was autonomous and very fast, giving the cheesecaking team complete control over a number of variables they likely would not have otherwise.

All of these can easily be designed out of future games. Not necessarily saying that they’re all bad things (in fact, I quite like some of them), but all together, they created the perfect storm of cheesecake this season.

You’re all correct that the prevalence of cheesecake is not due to teams’ being un-GP, but rather to the idiotically designed game they were forced to play. They did the best engineering they could with the restrictions they were given.

The rules setup and scoring dynamics definitely need to be reevaluated for the future in terms of how all partners of various types can contribute to an alliance effectively and fairly at all levels. I’m going to try not to say more than that because my opinions are too biased to be relevant. I just want to clear 1114 and 148’s names in all this.

We approached 1114 on Friday morning and asked if we could cheescake for them, not the other way around. Our main mechanism was suffering persistent unresolvable issues, and we knew based on our schedule that we’d be bottom ten and useless for elims. We also knew the biggest weakness of the 1114 148 alliance was canburgling, and that would be their priority for a 3rd robot, which, unfortunately, they wouldn’t be able to snag the best of by their 2nd and 3rd picks. They needed cheesecake; we needed to change. So we decided on Thursday evening that we would offer up our drivetrain. When 1114 was hesitant because they could only work with something super light and tiny, we jumped on it. We were not at all pressured by them to do so; they thought we were nuts for wanting to do it. We were up Friday night cadding the new drivetrain, and raced all of Saturday to assemble it.

Our team took immense pride in the fact that we got so far out of proper analysis and sheer persistence. This was our decision; We were trying to play the hand we were dealt, and I apologize for how it came across. 1114 and 148 were nothing but professional, gracious, and AWESOME to work with (1114 has long been my favorite team in FRC).

Sorry for any typos or hasty wording; I’m on mobile and rushing to get this out because no blame should lie on the alliance captains for the “new” robot, because that was our team’s initiative and decision.

I know a few teams that are “cheesecakers” and always go around and ask teams on their pick list if they would be willing to let them modify their robot for elims before alliance selections. And I would argue that it does require solving skills and strategy because you’re literally working on the clock to make the third robot as useful as possible to the alliance. I would rather have a cheesecaked bot on the field that has a purpose than having one on the field that’s just forced to sit in the corner and stay out of the way of their alliance partners.