Chief Delphi

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)
-   Rules/Strategy (http://www.chiefdelphi.com/forums/forumdisplay.php?f=6)
-   -   The Ultimate Game-Breaker Robot: 2012 Edition (http://www.chiefdelphi.com/forums/showthread.php?t=102697)

jason701802 15-02-2012 16:32

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by artdutra04 (Post 1127455)
Should we build robots based upon the intent of the GDC or should we build robots based only upon the rules written by the GDC?

Or do we build robots upon what we think are the GDCs intended rules. It's a small distinction, but a significant one.
Although there is no way to say for sure what the GDCs intent is, one could argue that 179's bot goes against the GDCs intent of the game because the robot is not on top of the bridge when it is balanced. Following the intent of the rules would seem to allow 179's bot because it is fully supported by the top of the bridge, but wouldn't allow a troll 'bot, whereas following the rules exactly (before the update) would allow a troll 'bot. Hopefully this is a good enough example to show the distinction between following the GDCs intent for the game, their intent for the rules, and the rules as they are written.

To me, the first option is kind of boring because it limits some of the greatest ideas (179 this year, 469 in 2010...), and the last option is only going to make the teams who didn't notice a loophole jealous of the teams who did.

Duke461 15-02-2012 17:52

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by Daniel_LaFleur (Post 1127414)

I wonder ... now that the ball shield isn't part of the bridge ... can I somehow use that area to keep my opponents off of thier bridge?

G25
Quote:

Robots may not contact or otherwise interfere with the opposing Alliance Bridge.
Violation: Technical-Foul. If the act of Balancing is interfered with, also a Red Card and the Bridge will be counted as Balanced with the maximum number of Robots possible for that Match.

Brandon_L 15-02-2012 23:05

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
In my unimportant honest opinion:

Whoever built a trollbot: You took a risk knowing that the rule could get overturned by the GDC at pretty much any time. Obviously you would hope it wouldn't, but it has, and you took that risk. Its quite obvious what they intended you to do with the bridges. Deal with it, just as you would have to in the real world.

To the GDC: Next time, don't wait until week 6 to pull something like this. It clears up what you intended the bridges to be used for/as, but this should have happened at maybe week 3 at the latest.

To both: The Q&A is a unspecific mess, thanks to the "We wont comment on exact robot design" rule that I heard somewhere from the GDC. Sometimes, you have to see an example to see where the poster is coming from. This probably would not have helped much in this case, as only a "What makes up the bridge?" question was asked, but if the question was more specific they most likely would have flat out given you a "No, you can't do that" on the spot. As long as the poster doesn't supply detailed drawings and such, it really should be allowed.

My 2c

Tristan Lall 16-02-2012 00:16

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by Andrew Schreiber (Post 1127151)
They were perfectly clear. Twice. The bridge was defined as everything in a certain drawing. The rule says supported. To me balanced and supported do not mean on top of. As an engineer my job is to be clear in unambiguous in the language I use.

Quote:

Originally Posted by wireties (Post 1127161)
For all of the future engineers reading this thread - you give me an answer like that quoted above in an interview and I will NOT hire you. Time is money and I will assume you are likely to waste both.

Quote:

Originally Posted by Andrew Schreiber (Post 1127166)
Then it is a good thing I don't intend on interviewing at a place that clearly has very bad and narrow definitions of words and doesn't evaluate all options for how to solve their clients needs.

Quote:

Originally Posted by wireties (Post 1127180)
If you needed a job, (with all due respect) does it matter what you think or intend?

So this side conversation confuses me.

If one needed to hire some engineers, doesn't it make sense to hire at least one who is willing to point out when the spec might be ambiguous, or when the customer's expectations might not correspond with the contract language? By arguing the point, that's exactly what Andrew is expressing to you.

Is it cheaper to hire him now (and waste a little time/money now and then), or to wait, and hire a lawyer later?

Bill_B 16-02-2012 00:31

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
I can forsee a new adjunct activity for FIRSTers that maybe goes along with Fantasy FIRST (FF) or maybe it will be a sidebar in that popular game. It involves assigning a probability to each and every rule in the 2013 manuals concerning whether it will be changed during the season by team update or Q&A. Tiebreaker points for those guessing the update or distance into the season at which the rule is altered. Nolo contendere applies for obvious spelling errors.

Of course, there will have to be a point in week 1 when the parimutuel window closes but I think there's a good chance it will have the effect of EVERYONE taking a good hard look at the rule book before much metal gets cut or even before too many CAD electrons are spun into fruitless webs. I made a post (somewhere) about my observation of teenagers' reaction to hearing about rules they hadn't read well enough to know if they could be ignored. I will find it again when I click on my name in this post. ;) :D

EricH 16-02-2012 01:33

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by Bill_B (Post 1127815)
I can forsee a new adjunct activity for FIRSTers that maybe goes along with Fantasy FIRST (FF) or maybe it will be a sidebar in that popular game. It involves assigning a probability to each and every rule in the 2013 manuals concerning whether it will be changed during the season by team update or Q&A. Tiebreaker points for those guessing the update or distance into the season at which the rule is altered. Nolo contendere applies for obvious spelling errors.

Rules Rumble. FF (or, to be more exact, Season Long FF) is pretty busy, and often ignored by those not involved. Plus, in a normal year, we're already done drafting when Kickoff rolls around (other than MSC, MAR, and CMP) and we're going through the adds/drops right about now. This year, we've got one more draft to do...

I think it could be done. But I'd do more of a pick-em style, where you could choose whether or not to make predictions on a rule. Such as: # of Q&As, # of posts on CD discussing the rule, if there would be a change (and if so, when), and whether the change would be made by Q&A or Update. :D Points awarded by accuracy (0 for fail, 2 for full accuracy on any one prediction (within reason), 1 for being "in the ballpark, but you're in the stands part of the ballpark"). Bonus points for predicting what change will be made. No making exactly the same prediction as someone else on a given rule.

Tristan Lall 16-02-2012 01:46

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by Siri (Post 1127152)
Disclaimer: I have to say I view the FIRST challenges more as sports than engineering assignments in this regard. This is probably because I in no way see the GDC as my client--that place is reserved for my students with the GDC as mere facilitators. I really don't care about the GDC's intent after they give a ruling, but I understand that apparently leaves me in the minority. (I am in no way implying that anyone who disagrees with my sentiments doesn't view the students as their goal/client/whathaveyou. Most of the guys on the other end are absolutely excellent mentors and engineers. I'd venture they probably don't view the GDC as just facilitators, though.)

Quote:

Originally Posted by IndySam (Post 1127060)
Not if the easy way accomplishes the task set out in the specs but in know way does what the customer intended.

The GDC intended robots to be balanced on top of the bridge. Sitting on a piece of structure that is solely designed to help prevent problem with the balls getting stuck is obviously not what they wanted.

This update could have been seen coming from miles away.

Quote:

Originally Posted by wireties (Post 1127073)
Teams that proceeded to build a troll bot were taking a calculated risk. It was theirs to take. IndySam (and myself) thought the risk was very high, others in this thread thought the risk acceptable. It turns out that the risk (this year) was too high.

Identifying and mitigating technical risk is a classic engineering exercise. The older (and hopefully wiser) one gets, the more careful one gets (hopefully). I love to undertake risky novel design approaches but I NEVER do so w/o crystal clear guidance from my customer (which usually comes after a informed, precise query) and a contractual promise (more money and time) my efforts will not doom the project. The goal (professionally) is to deliver on time, under budget AND to get the customer to come back with more work.

I'm in full agreement that the risks teams take have to be weighed against a tolerance for risk, and a set of incomplete information—and I've got more to say in that regard later. But I also wanted to highlight Keith's statement about the nature and perils of risk in engineering. That's a great summary of that constraint, as it applies to a contractual relationship between engineer and client. The trouble is, like Siri noted, I don't think this is the relationship that exists in FRC.

Fundamentally, we're not being asked 'build a robot that confirms our expectations about how robots can complete certain tasks'. The fundamental intent and the fundamental constraint is merely 'here are the rules, go play and maybe learn something about the value of engineering'. That encourages creativity and allows teams to try things that would never fly in a conventionally risk-averse engineering environment. That sandbox mentality is a big part of why it's fun. (And let's be perfectly honest: if this was a real engineering task, there are all sorts of things you'd rarely do—like spending time teaching the least capable members of the team new skills, or humouring bad ideas for educational value.)

I say that because FRC is unusual, even among competitions: the game rules are freshly written every year, and there's little weight in precedent or unwritten convention. By choosing to create a new game every year, the GDC has both the enviable opportunity and the significant responsibility of creating an internally-consistent set of documents that describe all of the critical parameters of the challenge.

Worse, when they resort to cop-outs like 'sorry, that's not what we intended', or 'use engineering common sense', they're betraying the essence of any competition—that the competitors are treated equitably so that any difference between them comes down to skill, preparation and the luck of the draw. Taking the role of the capricious god who strikes down his loyal subjects is no way to administer a competition.

FRC is not about a relationship between equals like the engineer and client, both of whom are equally able to negotiate a contract on mutually agreeable terms. Instead, it's like a contract between entities of disproportionate power—like you and the government, for example. In those cases, the more powerful party bears an additional equitable burden to ensure that they are treating those who have no recourse in a reasonable way. In FRC, that means setting expectations and sticking with them, unless there's a compelling reason to do otherwise—and only then, with due consideration of the impact on the weaker parties.

This situation violates that equitable expectation, because teams should have been able to rely on FIRST's past explicit statements. I of course realize that FIRST probably made an innocent mistake—and that has mitigating value. But FIRST missed other opportunities to mitigate the damage. They could have written the rule more clearly in advance, or labelled the diagram better. They could have explained that they surveyed all the teams, and if none were using this strategy, then the issue would be moot. They could have explained that the GDC had considered all of the issues, but upon reflection, decided that this was the best course of action.

Indeed, I think it's that last one that's annoying people. Probably for eminently practical reasons, the GDC is like a machine where questions go in, and dicta come out. There's rarely any indication of what's going on inside. Did the GDC actually understand the question, but deemed it unreasonable? Did they misunderstand the question? Did they spend more than 30 s discussing what a team feels is a crucial and complex issue that doesn't easily lend itself to a fair resolution? Occasionally, we even hear that the GDC has rather heated arguments over points of principle—is the ruling the result of level-headed conversation, or a shouting match? Knowing these things isn't crucial, but it is beneficial, because it establishes a frame of reference for the team that was just told their robot is pretty much useless. It also dampens criticism, because the explanation crowds out alternative theories like incompetence or ulterior motives—theories that can spread unchecked when the only response amounts to 'it's week 6: now the rule means this'.

The other factor feeding this discontent is the fact that the GDC had three perfect opportunities to get the definition right: once when they wrote the rules, and twice when teams asked for clarification. All three times it was an (understandable) oversight, but that doesn't change the fact that teams would desperately like to believe they can trust what the GDC has to say.

And that brings me to risk. Having been involved with this competition for a very long time, in many capacities, I know a lot about how the organization operates, and how GDC decisions are implemented in the field. There's a substantial amount of variability (most of which is undesirable), and this makes for a lot of risk, when it comes to non-mainstream designs. Will your inspector pass you or fail you, based on their understanding of the rules, and their willingness to tolerate equitable deviations from the rules under extenuating circumstances? The same goes for the referees. This has to figure into your risk estimates, if you're trying anything even remotely controversial.

Sometimes you make out well—like 190 in 2008, who managed to convince the officials (at one regional) that the rules referred to a robot-based co-ordinate system, rather than a field-based one. And sometimes you don't—like 1519 in 2008, who complied with all of the rules, but were dealt a mortal blow as a result of a dogmatic interpretation of the undefined term "robot".

Another aspect of risk is based on the likelihood that the GDC itself will change the meaning or interpretation of something that a team relied upon. Again, in my long experience, this happens all the time in FRC. The examples are too numerous to count, but sufficient to say, all too often they occur late in the build season or even during the competition season, when the impact on affected teams is most severe.

Now, I know about these aspects of risk, and because of that, I make contingency plans when I'm trying something controversial. I wasn't exactly surprised by this ruling. But the biggest problem with FIRST-induced risk isn't its effect on me, but rather its effect on less-well-prepared teams. If something isn't clear to a well-informed rookie (who has read the rules, but has no special insight into what the GDC is, much less today's definition of "engineering common sense"), it's not clear enough.


Quote:

Originally Posted by wireties (Post 1127115)
We have to supplement the rules with some common sense. it is not possible for the GDC (or anyone else) to be perfectly clear. The same is true in the real world, there is no such thing as a perfect contract. At some point both parties have to be reasonable. When the rules say SUPPORTED by the bridge (further reinforced by the points awarded by BALANCING), teams should start with the assumption that the robot is on top of the bridge. Any other interpretation is NOT common sense, carries considerable risk (doesn't mean its wrong) and should NOT be the basis for a design (especially with only 6 weeks) w/o crystal clear clarification from the GDC.

Quote:

Originally Posted by Bill_B (Post 1127116)
Does anyone else see the similarity between the "bridge problem" and the use of the term "vertices" for the frame? Some of the first words from my team members upon hearing the vertex-8-inch rule was "what if the robot is oval?" (having no vertices). Forget about how you might go about making 3/4" plywood bend around an oval, they were simply exploring what was meant by the given terms. The GDC has always been able to make their intention clear by specific wording to the desired effect. It is becoming apparent that the attempt to turn the rule book into some sort of street sign (word starvation) is doomed to failure. Particularly in a high-school context in which members are ALWAYS and constantly looking for imprecise statements and the way around rules. There's no furor about the vertex thing because no one is making a round or oval robot. Or are they? :yikes:

I'm no fan of "common sense" as a substitute for precision, where precision is beneficial.

We can say that it's common sense that driving under the bridge is not balancing it. But that presupposes a whole bunch of assumptions about what things are, and what terms mean. Since those terms are not all clearly and precisely defined in the rules (much less in common parlance), those assumptions are inherently open to debate. That doesn't make it not common sense—it merely demonstrates that nobody has a monopoly on common sense. If I disagree, based on my own self-consistent and reasonable interpretation, there isn't really much you can say to prove me wrong. We just disagree, because we have different worldviews.

For that matter, isn't it also common sense to observe that a robot can be round (ever see a Roomba)? If common sense is so important, why did the GDC fail to employ it in that case? Could it also be true that common sense is not always obvious, even to our esteemed rule-writers? If they can't get it straight all the time (and why would we expect that of them?), why should we? Holding every team to their brand of common sense is untenable—and apparently unreasonable if they can't manage it themselves. Actually, the history of the rounded robot question is even more sordid than that. In 2009 FIRST actually said that a curved bumper was composed of an infinite number of corners, and thus each needed to be protected individually. (Again, that's something that a rookie wouldn't know, and therefore couldn't easily rely upon to assess the likely actions of the GDC.)

They've said other nonsensical things, like plywood isn't plywood if you made it from plies of wood, or that the frame perimeter is necessarily a polygon (it's not: polygons are planar, the frame perimeter need not be). Give me a while, and I can go on and on—but the point is that I can say with some authority that many GDC interpretations do not reflect my understanding of common sense, and that they probably set off red flags for others as well. And as Bill noted, it should be pretty obvious that the average teenager has a different attitude toward common sense than people the age of their parents (like some GDC members). It's stuff like this that makes me adamantly opposed to the use of common sense as a crutch for bad specifications.

wireties 16-02-2012 02:28

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by Tristan Lall (Post 1127801)
Is it cheaper to hire him now (and waste a little time/money now and then), or to wait, and hire a lawyer later?

It is cheaper to hire engineers who are both innovative and practical and who communicate effectively with the customer. These traits are not mutually exclusive. It does not matter (in the real world or in FIRST) if a novel design does not meet requirements, does not please the client, is over budget and/or is over schedule.

No doubt the teams that started a troll bot are pretty upset but it was a risky design decision. No amount of rule parsing, pre or post GDC, is going to change that. So how about we all learn from this and pose less obtuse questions to the GDC next year?

wireties 16-02-2012 03:00

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by Tristan Lall (Post 1127861)
... the GDC has both the enviable opportunity and the significant responsibility of creating an internally-consistent set of documents that describe all of the critical parameters of the challenge.

This is where you miss the point. It is NOT strictly possible for the GDC to create a foolproof "set of documents". There will always be some wayward engineer asking what "is" is (or thinking you can balance a robot under a bridge) . Common sense is an essential component. This is how the real world works.

Quote:

Originally Posted by Tristan Lall (Post 1127861)
FRC is not about a relationship between equals like the engineer and client, both of whom are equally able to negotiate a contract on mutually agreeable terms. Instead, it's like a contract between entities of disproportionate power—like you and the government, for example.

The engineer and customer/client are not on equal footing. The party paying the bills has final say over all matters.

And FIRST is nothing like "you and the government". FIRST is voluntary, one is compelled to do nothing.

Quote:

Originally Posted by Tristan Lall (Post 1127861)
Did the GDC actually understand the question, but deemed it unreasonable? Did they misunderstand the question?

It is my understanding that a direct query was not made (in an effort not to reveal the strategy). If I am mistaken, I apologize. If a team posed a question like "Assume team XYZ makes a robot with a 7" profile and parks it on the lexan under the bridge, will it count in the same manner as a robot on the top of the bridge at the end of the game?" and the GDC said that is OK then I also would be pissed. But that did not happen, did it? Everyone was being a little too clever and got stung.

I have designed some innovative control systems in my career. I have to work hard at not straying so far outside the box that I can't see the box. And I can honestly say that I would not have derived the trolling strategy from the GDC queries. Even now I can't believe anyone would think this a valid strategy - what on earth were they thinking?

Quote:

Originally Posted by Tristan Lall (Post 1127861)
I'm no fan of "common sense" as a substitute for precision, where precision is beneficial.

Your mistake is thinking common sense and precise thinking are mutually exclusive.

Quote:

Originally Posted by Tristan Lall (Post 1127861)
We just disagree, because we have different worldviews.

Wrong - we disagree because we have lived it. We have made similar mistakes and learned from them. The mentors in the thread (who do not agree with you) are not expressing opinions but justified conclusions based on decades of experience. Is that to count for nothing?

Quote:

Originally Posted by Tristan Lall (Post 1127861)
And as Bill noted, it should be pretty obvious that the average teenager has a different attitude toward common sense than people the age of their parents (like some GDC members). It's stuff like this that makes me adamantly opposed to the use of common sense as a crutch for bad specifications.

This is not a dumbed-down game for high school students. It is for students who want to step up and do college level work and/or entry level professional work. The whole purpose of FIRST is to introduce extraordinary students to a real world-ish engineering exercise (mentored by practicing engineers) and prompt them to consider to study engineering, physics or computer science.

Brandon_L 16-02-2012 03:09

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by wireties (Post 1127879)
I have designed some innovative control systems in my career. I have to work hard at not straying so far outside the box that I can't see the box. And I can honestly say that I would not have derived the trolling strategy from the GDC queries. Even now I can't believe anyone would think this a valid strategy - what on earth were they thinking?

This. I can't agree more.

The purpose of the bridge is pretty obvious. to balance. not to find loopholes to score more points. And I also completely agree, if someone flatout asked if it was legal, and they said yes, then overturned it, there would be a reason to be angry. But what is going on here is completely ridiculous. If you can figure out how to make a robot that small and still score 3 pointers you can suck it up and put a bridge manipulator on your robot.

lemiant 16-02-2012 03:19

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
The purpose of the bridge is what the GDC said it was, nothing more, nothing less.
And of course they didn't post a clear Q&A. The public nature of the Q&A would make that require revealing their whole strategy. They had the GDC state regulations which made them perfectly legal according to the rules and that should have been enough.

wireties 16-02-2012 03:35

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by lemiant (Post 1127883)
The purpose of the bridge is what the GDC said it was, nothing more, nothing less.
And of course they didn't post a clear Q&A. The public nature of the Q&A would make that require revealing their whole strategy. They had the GDC state regulations which made them perfectly legal according to the rules and that should have been enough.

And how did that approach work out for "them"? Everyone makes mistakes, the real question is do we learn from it. Evidently you are not there yet.

My first boss had a saying - "don't make big mistakes and don't make little mistakes twice". Trolling-gate was a big mistake that hopefully will not get repeated.

lemiant 16-02-2012 03:38

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Spotlight took me to this:
http://www.chiefdelphi.com/forums/sh...ad.php?t=18309
What would you have said?

wireties 16-02-2012 03:53

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
Quote:

Originally Posted by lemiant (Post 1127885)
Spotlight took me to this:
http://www.chiefdelphi.com/forums/sh...ad.php?t=18309
What would you have said?

sorry - was before my time

DjScribbles 16-02-2012 10:35

Re: The Ultimate Game-Breaker Robot: 2012 Edition
 
I haven't followed this whole thread (read the beginning), so forgive me for being a bit out of context but...

It seems to me there should be an alternative way of submitting these types of questions to the GDC when it involves a unique strategy that you don't want to share with the world if it's legal.

The reason the Q&A question was so ambiguous is likely because teams didn't want to tip their hands; but if they could ask the question and privately get approval that it is within the intent of the game, and if the strategy is denied then the private Q&A is made public so other teams are made aware of the ruling and subsequent clarification to the rules.

I don't think world class teams should be forced to gamble their success on whether or not something is legal. If you decide a strategy probably isn't really "the intended game design", but it turns out to be legal, you'll be kicking yourself when you get schooled by someone who took that risk and "got away with it".

To me it seems the only arguments for not having a privacy option to the Q&A would be to minimize the amount of work required and to protect the "not comment on a specific design" clause.


All times are GMT -5. The time now is 20:13.

Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi