Cheesecake: How far is too far?

[AdamHeard]

Originally Posted by Mr. Van

I believe you are right about this. This puts struggling teams who actually attempt to accomplish the game challenge at a competitive disadvantage compared to those teams who anticipate "cheesecaking" and simply provide a "cheesecake pan".

Is this what we want?

- Mr. Van

I don't think it is, and the rules should be modified to rule this out (while fully maintaining the ability to help others with gearboxes, etc...).

That being said, if it's legal we will likely do it.

[Jasmine Zhou]

Originally Posted by Rachel Lim

Would cheesecaking robots from 2014 help alliances? Yes. Why wasn't it done then? Was the idea really not thought of until this year?

I guess a better question would be:
If we replayed Aerial Assist, would we see cheesecaking to the extent that we saw it in Recycle Rush?

Cheesecaking in 2014 (in the form of modifying/adding mechanisms to alliance members) definitely happened, at least from 1678. Mike mentioned this in the original uproar when he popularized the term cheesecake. http://www.chiefdelphi.com/forums/sh...1&postcount=81
By the time we left the Sacramento regional (and possibly earlier, my memory is already fuzzy on this), we went to competitions knowing we'd be adding passive mechanisms to many robots in qualifications in order to maximize assist points in an attempt to seed first.

The key word there is probably "qualifications". We didn't pick robots for eliminations specifically for the ease with which we could add an assist mechanism to them- the second pick was often defense and driving, not how easy it would be to modify their robot. Because defense was still a thing.
I think that the cheesecake-ability affecting alliance selections seems to be what people are objecting to, not cheesecaking teams throughout the event.

I think we'd also see more cheesecake passed around if we replayed Aerial Assist in the current climate, just because more people have realized that it's an option.

[George Nishimura]

2014 is interesting because there was a pretty low-barrier for 3rd robots. Kitbot + intake was the minimum viable robot.

Even then there was still some 'cheesecaking'. Didn't 254/2056 modify 865 at Waterloo?

[Chris is me]

I think it's really important to remember that 2014 was also an anomalous year, where including all three robots in every instance of scoring was crucial for success. Even in other games where scoring was the best option, you could have a weaker third robot.

I think the better question to ask would be if we replayed 2013, what extent would we see cheesecaking? I would argue that we would see almost none with the exception of 84" blockers. These usually weren't pre built, and if they were the blocker was still extremely simple / something that could be done on the spot.

I think discouraging the *general* practice of captains and first picks working with the second pick to modify and improve their robot would be doing a disservice. Quite honestly these situations have been some of the most inspiring parts of FRC for many of us. I get that maybe we shouldn't be letting teams compete with brand new robots for elims in the future, but I am FAR more concerned about an overly broad rule restricting teams from doing other tasks than I am about that. I'm even more concerned about a vaguely worded rule being interpreted by event staff as a ban on robot modifications of any sort.

Originally Posted by Chris is me

I think the better question to ask would be if we replayed 2013, what extent would we see cheesecaking? I would argue that we would see almost none with the exception of 84" blockers. These usually weren't pre built.

I would make sure to bring simple 10 pt. hanging devices that could be attached to an alliance partner who may not have one (or as good of one). While you're right that most tall blockers were made at competition, I would prepare my team with either the ability to make an effective one quickly or have an effective design brought with me. Some games can be played all by yourself, and others require all alliance partners to be functioning at their best. If it's to your advantage to have all of your alliance partners able to contribute as many points possible to the alliance (which could be vital to winning some of the close matches for W-L-T games, or for average scoring ranking like this year), doing anything you can to gain points for your alliance or deny your opponents points would be helpful.

[Kevin Sheridan]

Originally Posted by George Nishimura

2014 is interesting because there was a pretty low-barrier for 3rd robots. Kitbot + intake was the minimum viable robot.

Even then there was still some 'cheesecaking'. Didn't 254/2056 modify 865 at Waterloo?

We cheesecaked 2135 at CVR last year by adding some pex tubing to make it easier to inbound through their robot. We "cheesecaked" 865 by zip-tieing their intakes in place to make a ball tunnel for inbounding. Both these robots competed at champs with our modifications still on their robots.

We were also preparing to cheesecake a goalie pole onto our 3rd or 4th robot at champs but the pole was never completed (we had it in our pits though) and it was unnecessary once we knew we were picking 469.

[Citrus Dad]

I think the concern this year is over the level of cheesecake that we saw. I will reiterate a point made by several people: This is entirely in response to this year's game and rules. The best way to limit the amount of cheesecake is create a game in which the least experienced teams can make an effective contribution to the game at all levels. I and other have pointed out that the GDC did not accomplish this objective in this year's game, and statistical analysis of the OPR distribution supports this contention. This was obvious to us on the day of Kick-off; the GDC should have seen that as well. (I would like to know the composition of the GDC, but I urge that it have a number of non-engineers involved in game development.)

I've pointed out in earlier threads that an absolute prohibition would have prevented us from working with two rookie teams on Newton last year that greatly improved their games in qualifications. FIRST wants strong interaction across teams at the Championship (see the justification for going to 2 Championships.) Why remove one of the frequently used means of facilitating that interaction? Remember that competition really is only a secondary objective of FIRST; this is not the NFL. Promoting cultural change is the primary focus.

[Rachel Lim]

I guess I should have said that in 2014, I didn't see cheesecaking define the game or the strategy the way I've seen it happen in 2015. I don't think modifying robots to make them more competitive, either in quals or elims, was the major issue this year. I believe (from what I've heard) that it was that those teams were basically just their cheesecake, so being an easy platter was more desirable than trying to contribute individually. Then again, it could have just simply been that I had nothing else to compare 2014 to until this year.

It seems that we needed Recycle Rush to really figure out just how large a role cheesecaking can play.
It seems that we needed the modified response to Q461 to figure out that maybe the original response had its advantages.

I still believe that if we replayed Recycle Rush with the original Q461 response (or something similar), teams would find a way around it because cheesecaking would still be necessary. There would be some other debate about this. I also believe that if we replayed another game where it wasn't as necessary, we would see it, but it wouldn't be as defining a factor.


This has been an interesting discussion, and one I've definitely learned from. If I was to rewrite my first post, it'd probably be:

Design a game that requires cheesecaking to win, put some rules in to try and prevent it, and it will still happen.
Design a game that doesn't require cheesecaking to win, put some rules in to try and prevent it, and it won't happen much.
Cheesecaking here used to refer to the need to put an entire mechanism on to win or something to that extent, not to smaller modifications

Rules define how far a strategy can be taken, but the game defines what strategy is needed.

[Citrus Dad]

Originally Posted by Rachel Lim

Rules define how far a strategy can be taken, but the game defines what strategy is needed.

Well put.

[GreyingJay]

Great, so we'll see you all at regionals next year with our 2016 robot, "Tabula Rasa". Here's to the shortest build season ever!

[tcjinaz]

Originally Posted by Tungrus

...create any number of rules...someone will find a loop hole.

The team with the best lawyers wins?

The teams with the most money will attract the best lawyers.

[zsnow]

There is a fundamental problem with attempting to define an explicit boundary between what is good and what is bad. For every clearly enforceable rule, there is some reasonable exception where most would agree the rule should not apply. This is true whether it is varying degrees of cheesecaking, or really any polemical question (even as far as questions of morality and ethics). Trying to define an unambiguous line which "thou shall not cross" is futile.

Cheesecaking in its current form appears to be more beneficial than otherwise. The GDC was wise in not trying to come up with some unequivocal ruleset to govern such complex issue that can only be judged on a case-by-case basis.

These views are my own and not necessarily those of my team. I'll go prepare the stake...

[Abhishek R]

Like Rachel has pointed out, I have to agree it's the game design that drives the extent of "cheesecaking" we saw this year.

In 2014, cheesecaking would be limited to assisting devices. These will not inherently make or break the game for an alliance; they aren't extremely necessary to be competitive at even the highest levels of play.

In 2013, cheesecaking would be limited to full-court blockers, or maybe more advanced, a 10 point hang. Again, neither of these were necessary at even the highest levels of play.

In either case, the cheesecaked robot still had to execute its role pretty well. In 2015, you install the system, they get the cans (which absolutely are necessary in order to win), and then they're done for the match. Furthermore, they are taken in autonomous, so the team itself doesn't really do a whole lot in terms of playing their role for the match. A higher resource team is going to be able to build a faster and better cheesecake than many teams can build for their own robot's system, and there's no way to win if you can't keep up. I think this is where a lot of the animosity towards the cheesecake in 2015 comes from.

[Qbot2640]

Originally Posted by Wayne TenBrink

In my opinion, "too far" is when a robot shows up for playoffs that wasn't there for qualification rounds, and nobody else had a chance to select it.

In my opinion, this is a big part of it...but the bigger part is when a team makes some kind of deal with an "alliance captaining team" that removes other, arguably more qualified teams, from consideration. Alliance selection should be based upon the performance of a team throughout qualification rounds...not upon what kind of "backroom" deal can be made. In this "now infamous" case, no one has mentioned the other 44 teams on Curie who all started packing up after alliance selection ended. One of them lost an opportunity.

[Citrus Dad]

Originally Posted by GreyingJay

Great, so we'll see you all at regionals next year with our 2016 robot, "Tabula Rasa". Here's to the shortest build season ever!

So long as the GDC doesn't make the same fundamental mistake that they made this year, we shouldn't see anything like what occurred this year.

Originally Posted by Qbot2640

In my opinion, this is a big part of it...but the bigger part is when a team makes some kind of deal with an "alliance captaining team" that removes other, arguably more qualified teams, from consideration. Alliance selection should be based upon the performance of a team throughout qualification rounds...not upon what kind of "backroom" deal can be made. In this "now infamous" case, no one has mentioned the other 44 teams on Curie who all started packing up after alliance selection ended. One of them lost an opportunity.

In general alliance captains do not choose teams in order of their qualification rankings, or even with regard to the rankings at all. The alliances are put to together through a combination of strategic (which team is a bigger threat?) and functional (which team has best kiss passer or trusser?) assessment. Please don't expect that your qualification ranking will be rewarded with a high alliance pick unless you're an alliance captain.