The cheesecake runaway

[Mr. Van]

Some questions about "cheesecake":

Team A has a MECHANISM brought in with the intent of giving it to their alliance partner. What happens when another team asks if they can have it?

Team A begins to work with and "cheesecake" Team B early on (as was apparently the case with the teams being most frequently discussed here). What happens when Team B is selected by another team before Team A gets the chance to? Does the "cheesecake" remain with Team B?

What happens when there are multiple teams presenting themselves as "cheesecake pans"?

What happens when a weak team is allied with Team A early on and they ask "can you cheesecake us for this round?" Is it appropriate for Team A to say "we won't give you this MECHANISM, we're reserving this for our 3rd pick"?

When might this sort of thing start? Could Team A begin working with Team B during the build season and have Team B bring an "inferior" robot to compete (with the intention of being "hidden in the rankings") and then reveal the "cheesecaked" version after alliance selection?

Just some questions...

- Mr. Van
Coach, Robodox

[Kevin Sevcik]

Count me with Dr. Joe as being opposed to this level of cheesecaking. It's troubling to me that this sort of thing could encourage teams to build and/or strip down robots specifically to act as platforms for another team's mechanism. Somehow I don't think "Design and build for 6 weeks, then scrap it all to spend 24 hours integrating someone else's design" is quite what's intended for this program.

As has been pointed out here already, what outside of practicality is preventing a team from bagging an entire robot and then cheesecaking it onto another team's RoboRIO during Division elims? I don't think you can game design around that level of cheesecake. Yes, the receiving team's drivers wouldn't be as experienced, but you could mitigate that by picking the quickest learners and giving them practice time on your official bot before the alliance pairings. If Team [Famous Number] was bruiting about they had a practice robot to cheesecake onto a 2nd/3rd pick, I'm pretty sure after this season you'd have at least a handful of teams interested in the offer. Now that we know stripping down to a ultralight kitbot drivetrain is acceptable cheesecaking, I'm not sure what's preventing this last step.

[Doug Frisk]

Originally Posted by Skyehawk

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.

2512 did much better in quals at Duluth and Peoria, enough that they were top seeds and ended up finalists in both events.

One of the things they did was an adjustment to the robot to make it a more reliable six stacker which was capable of doing 2 six stacks with noodled totes. Prior to champs they were mostly making capped 5 stacks.

That seemed to have affected their auton reliability. They scored an average of 17.8 in quals in Duluth and 19.3 in Peoria. On Hopper, they averaged only 9.2.

Their coopertition was also significantly down compared to the regionals.

I think Wave (2826) knew that 2512 was better than the stats showed and saw it fit into a team strategy. Removing 2512's 45% effective 20 point auton and adding 987s can burglers left them with an extra can or two and an extra 2 six stacks.

Watch the matches and you can see them always play that very consistent role, after the can wars they pick up a green container and trundle to the drivers station 3 corner, build a six stack with a noodled container (42 points) and a second 6 stack (12 points) with the remaining totes (unless Wave needed more to finish their 3 6 stacks). That had them contributing a reliable 54 points to each match in finals and doing an excellent job of staying out of 987's way.

Meanwhile 2826 did their 28 point auton plus 3 42 point stacks for 164 points and 987 would make 2 to 3 six stacks from the landfill with typically unnoodled containers as containers were available for 70 to 100 points per match.

If everything had gone perfectly in auton and teleop, they could have had 7 capped 6 stacks, 4 (maybe 5) with noodles which would have taken them to 312 or 320.

They worked very effectively as a 3 robot team and that's why they are the only ones who managed 290.

[Citrus Dad]

Originally Posted by George Nishimura

It's interesting to me that the World Champions in general never needed to rely on a cheesecaked robot (I believe that's true). They just had an incredible steal of a third robot (1671).

We had prepared 5012 in similar fashion as we had 5529, 5458 and 5027 in our three regionals (with 118's mechanism this time), but 1671 turned out to be so valuable beyond what we imagined that 5012 never made it on the field. That might have been different if the 148/1114/900 coalition had made finals.

[Citrus Dad]

I don't think cheesecake will be a long-term trend in most games. The last 3 years we've played on Einstein with 862, 1640 and now 1671 as the third member of our alliances. Other than putting a noodle blocker on the back of 862 I don't think we'd imagine cheesecaking any of these outstanding robots (that could have been first picks on some alliances.)

[NPeshimam]

Cheesecaking this year felt a lot like the modifications made to autonomous routines in 2014, such as reading hand gestures from a computer camera. It was not explicitly illegal, and allowed teams to take advantage of a lack of specificity in the rules to improve robot function. This year FIRST recognized the ingenuity behind those modifications, but banned them so that autonomous routines that had no human interaction would be implemented. If they really feel that cheesecaking is taking away from the spirit of the competition it will be addressed in the rules next year. Until then it seems wrong to call out teams that take full advantage of the rules and resources at their disposal to do everything in their power to win.

[Steven Smith]

Originally Posted by NPeshimam

Cheesecaking this year felt a lot like the modifications made to autonomous routines in 2014, such as reading hand gestures from a computer camera. It was not explicitly illegal, and allowed teams to take advantage of a lack of specificity in the rules to improve robot function. This year FIRST recognized the ingenuity behind those modifications, but banned them so that autonomous routines that had no human interaction would be implemented. If they really feel that cheesecaking is taking away from the spirit of the competition it will be addressed in the rules next year. Until then it seems wrong to call out teams that take full advantage of the rules and resources at their disposal to do everything in their power to win.

I think the majority opinion is not to criticize any particular person/team/etc for either cheese-caking, being cheese-caked, etc... but a concern about the next "arms race" where a team with sufficient resources could potentially bring their own alliance to a tournament by building complimentary mechanisms/robots and handing them to 1st/2nd picks.

If the rules allow it, teams would (and should) pursue it. Now in 12-14 weeks, you are working even harder... not just to make your own robot, but to design cheesecakes around stripped down robots.

A rule (or game design) that lessened the strategic benefit of cheese-caking, while leaving the option to help other teams out would likely be well received. One idea I floated on the post a while back regarding the Q/A banning all help, was to potentially allow unlimited cheese-caking through X time (Friday quals at a regional?), and only limited (2-5lbs?) changes after alliance selection. If a well meaning team wants to help another team overhaul their robot, they could and should... but it shouldn't provide an overwhelming strategic advantage to the giver. It should be able to be scouted, or perhaps posted at the inspection station... so the gift has made the team inherently more valuable to other alliances. To me, there is a distinct difference between cheesecaking a team to help them, and cheesecaking a team to help myself.

[marshall]

Originally Posted by Steven Smith

I think the majority opinion is not to criticize any particular person/team/etc for either cheese-caking, being cheese-caked, etc... but a concern about the next "arms race" where a team with sufficient resources could potentially bring their own alliance to a tournament by building complimentary mechanisms/robots and handing them to 1st/2nd picks.

If the rules allow it, teams would (and should) pursue it. Now in 12-14 weeks, you are working even harder... not just to make your own robot, but to design cheesecakes around stripped down robots.

A rule (or game design) that lessened the strategic benefit of cheese-caking, while leaving the option to help other teams out would likely be well received. One idea I floated on the post a while back regarding the Q/A banning all help, was to potentially allow unlimited cheese-caking through X time (Friday quals at a regional?), and only limited (2-5lbs?) changes after alliance selection. If a well meaning team wants to help another team overhaul their robot, they could and should... but it shouldn't provide an overwhelming strategic advantage to the giver.

Also an enlightened viewpoint and I think maybe a reasonable suggestion. The timeframe is a good way to handle it.

[aashay2035]

At Seneca (Mar Week 4 Event) there was a team 5640 (Pegasus) who needed help to improve their robot. We and 87 scraped up parts to make them a ramp. On Sunday they were not there just before their first match. So we were planning to go on the field with their robot and no one on there team. But they came on time just 2 mins before their match. If we were allowed to go on the field with someone elses robot then cheesecaking has to be allowed.

[Mr. Van]

How about:

The Withholding Allowance (or perhaps anything brought into the competition venue) is limited to Spare or Upgrade MECHANISMS/COMPONENTS for the team's (not alliance's) robot. MECHANISMS/COMPONTENTS brought for the specific intent of being applied to other team's robots are prohibited. Teams are encouraged to bring COTS items to help/upgrade ALL teams at the event. Modifications to any team's robot belong to that team.

Is that a reasonable place to start from?

- Mr. Van
Coach, Robodox

[lcoreyl]

Originally Posted by Mr. Van

How about:

The Withholding Allowance (or perhaps anything brought into the competition venue) is limited to Spare or Upgrade MECHANISMS/COMPONENTS for the team's (not alliance's) robot. MECHANISMS/COMPONTENTS brought for the specific intent of being applied to other team's robots are prohibited. Teams are encouraged to bring COTS items to help/upgrade ALL teams at the event. Modifications to any team's robot belong to that team.

Is that a reasonable place to start from?

- Mr. Van
Coach, Robodox

No. It goes back to the rule that lead to the thread where the cheesecake term came from, which was then reversed because FIRST was made aware of how it disallowed so many things that are good for FIRST.

At this point I've read this and the 900 champs thread, and all I've seen is this:

Team Cheesecake: "as a recipient of cheesecake I can tell you this was extremely inspirational for our team. Our kids loved being a part of this alliance"

Team lactose/gluten/fat free cake: "This made me feel icky regardless of the fact the team it happened to didn't feel icky. This wasn't the intent of FIRST. Also the other teams that could do things and didn't get drafted felt icky. This should be outlawed!"

Unless someone has a better argument than how seeing this makes you "feel", I think the evidence is pretty overwhelming that only good came of this.

[Lil' Lavery]

Originally Posted by lcoreyl

No. It goes back to the rule that lead to the thread where the cheesecake term came from, which was then reversed because FIRST was made aware of how it disallowed so many things that are good for FIRST.

At this point I've read this and the 900 champs thread, and all I've seen is this:

Team Cheesecake: "as a recipient of cheesecake I can tell you this was extremely inspirational for our team. Our kids loved being a part of this alliance"

Team lactose/gluten/fat free cake: "This made me feel icky regardless of the fact the team it happened to didn't feel icky. This wasn't the intent of FIRST. Also the other teams that could do things and didn't get drafted felt icky. This should be outlawed!"

Unless someone has a better argument than how seeing this makes you "feel", I think the evidence is pretty overwhelming that only good came of this.

First off, the positives your listing are also teams talking about how they feel.

Secondly, the unfortunate reality is that when the source of inspiration comes from working closely with an elite team and/or deep runs in the eliminations, unfortunately that's a zero sum equation. For each team that was picked to cheesecake, a team was left off an elimination alliance. In many cases that team was a very capable team that had a competitive season, but was left off in favor of a bot that presented a better blank slate to add parts to. In other words, they were left off for a less capable team. It's pretty easy to see how being passed over for a less capable team could be disappointing. Disappointment can very well lead to a lack of inspiration.

Does the the benefit of the cheesecaking team outweigh the disappointment of the teams passed over in alliance selection? It's certainly not an easy call to make, but it's not one that should be written off as some trivial matter that can be summed up by air quotes and sarcastic titles.

[Abhishek R]

I think the reasoning behind cheesecake is extremely simple.

If you want to win, you need those cans off the step, or you are virtually guaranteed a loss.

If you want those cans off the step, you need to be faster to them than your opponent.

If the available teams to your alliance during your draft are not capable of grabbing the cans faster than the opposing alliance, you will not win (assuming they have the necessary stacking power).

You now have two choices:

1. Go with a slower or no can grabber, resulting in a lack of crucial game pieces, in turn, a loss.

2. "Cheesecake" a team with a faster mechanism, ensuring your competitiveness.

This is a direct result of the game design, so I'm not sure what people were really expecting. If you wanted to win, and there were no teams with quick canburglars available or left in the draft, you either cheesecaked or lost. It's a chokehold that several teams saw from Kickoff, and we are now seeing the results.

[asid61]

Originally Posted by lcoreyl

No. It goes back to the rule that lead to the thread where the cheesecake term came from, which was then reversed because FIRST was made aware of how it disallowed so many things that are good for FIRST.

At this point I've read this and the 900 champs thread, and all I've seen is this:

Team Cheesecake: "as a recipient of cheesecake I can tell you this was extremely inspirational for our team. Our kids loved being a part of this alliance"

Team lactose/gluten/fat free cake: "This made me feel icky regardless of the fact the team it happened to didn't feel icky. This wasn't the intent of FIRST. Also the other teams that could do things and didn't get drafted felt icky. This should be outlawed!"

Unless someone has a better argument than how seeing this makes you "feel", I think the evidence is pretty overwhelming that only good came of this.

Actually, I think people's interpretations of the Q&A went far beyond the intention. They might add a revised version of the rule with a blue box to state that an assembled gearbox or a wheel with holes in it from riveted tread is perfectly acceptable. The wording was off, but the intent was there.
Right now, it is legal to in effect bag a second robot and throw it on the field. Given how long manufacturing takes (particularly machining identical parts or sheet metal punching), this is not a tall order at all.
If a team gets cheesecaked, then obviously they were okay with it. They allowed their robot to be cheesecaked, and so they accepted it.

[TheOtherGuy]

Originally Posted by lcoreyl

No. It goes back to the rule that lead to the thread where the cheesecake term came from, which was then reversed because FIRST was made aware of how it disallowed so many things that are good for FIRST.

*snip*

Unless someone has a better argument than how seeing this makes you "feel", I think the evidence is pretty overwhelming that only good came of this.

I thought the original issue that inspired "cheesecake*" was zero items could be shared between teams?

How about that rule about what constitutes a robot - R1, I believe... can that be modified to limit full-blown redesigns? Or perhaps a weight limit on shared parts?

I had a small rebuttal written up regarding the zero-sum game, but Sean beat me to it more elegantly.. Remember that FRC is 3,000 teams, and the opinions expressed here do not necessarily reflect those of the masses.

*Can we get a cheesecake smilie to go with the dead horse?