Personally, I find the issue of “rules enforcement” to also be very frustrating at times.
This thread has kind of focused upon the specific issue of the bumpers that you mentioned, but for me (and, I speculate for you, as I think you were just using the bumpers as a benign example) the real issues of your question have to do with whether or not a team chooses to follow rules that may or may not end up being enforced at tournaments but which DO have a MAJOR impact on competitiveness. For more years than not, our team has had to specifically grapple with limiting our robot design in order to ensure rule compliance, knowing that if we didn’t limit our robot design in that way, we could have a much better robot. After making such design limitations, it is extremely frustrating (and often even unfair) to observe other robots on the field that did not make these robot design limitations but are subsequently allowed to regularly break the rules as written while gaining what is often a very significant competitive advantage from doing so.
For us, in 2007, the big rule along these lines was remaining within the robot volume requirement. In order to have an arm design which was guaranteed to always remain within the volume requirement yet still be able to reach all three levels of the rack, we went to a three-jointed arm which could be guaranteed to always comply with the rules. However, that year, about 25% to 50% of the robots we saw regularly violated the rule during matches. Our robot ended up being too complicated to really work, while simpler designs that we discarded because of robot volume rules were very successfully used by other teams, despite the fact that they regularly exceeded the robot volume but received no penalties since the rule was nearly impossible for referees to enforce. My complaint for the rule problem that year rests not primarily with the inspection committees or the referees, but for the GDC in writing rules which would end up being nearly unenforceable, or instituting game penalties which require the referees to have to make highly subjective judgment determinations multiple times every match.
In 2006, the regularly debated hard-to-enforce rule was the maximum ball velocity limitation.
In 2008, the hard-to-enforce or at least inconsistently enforced rule was the “interfering with a hurdler” rule. In order to make the right design tradeoffs, knowing whether or not such rules will be enforced is very important. If hurdlers are granted almost no protection in that zone, it makes the relative importance of running laps much more significant, making a lap-bot more competitive. However, if the hurdlers are granted essentially “no contact” protection in the hurdling zone, a lap-bot will have to be very careful to watch out for hurdlers in that zone and will thus need to expect to complete less laps during a match. The “crossing backwards over a line” penalty was similarly frustrating.
In 2009, I was VERY PLEASED that the game featured almost no “subjective penalties” – with the removal of penalties for pinning, ramming, and no common game-related penalties, I don’t recall as much debate over rules-related issues. (I may be forgetting something here; if I have, please chime in!) In general, I would like to commend the GDC for coming up with a set of game rules in 2009 that didn’t have subjective penalties or hard-to-enforce rules.
In 2010, however, there were multiple rules that were in the category of “subjective enforcement” that greatly impacted our robot design. The most significant of these, however, was G43 that “ROBOTS may POSSESS only one BALL at a time. Violation: PENALTY.” Initially, knowing that ball possession would be one of the keys to this years game, we wanted to maximize opportunities for ball possession by having a very wide ball possessing system. Such a system could accidentally or unintentionally possess more than one ball at a time. We had the debate about how strictly this rule would be interpreted – would a robot that could possess more than one ball at a time be deemed illegal? Would the “two ball possession” be given an “innocent until proven guilty beyond a reasonable doubt” perspective, or would it be the case that if the referee wasn’t sure whether or not two balls were being possessed, the team would be given a penalty – ie, was it the team’s responsibility to ensure that it was abundantly clear to the referees that the robot was not possessing two balls?
To get down to details, we weren’t sure whether we could make a ball-possession mechanism that stretched across an entire side of the robot, or did we need to make it narrow enough such that it couldn’t possibly possess two balls at once? Would a “software solution” be sufficient where we could have a split-roller from side-to-side, and turn off one side when the other side had a ball in it? We considered and discussed such options.
Then, when Team Update 2 came out, the GDC added an explanatory box to the rules stating “It is important to design your ROBOT so that it is impossible to inadvertently or intentionally POSSESS more than one BALL at a time.” Furthermore, it said, “Since referees may find it difficult to determine if additional BALLS in contact with the ROBOT are being herded or POSSESSED, it is imperative that teams avoid ambiguity.”
This made it sound to us that it was the team’s responsibility to actively demonstrate compliance with this rule – if the referee wasn’t sure whether or not multiple balls were being possessed, a penalty would be called, as it was the team’s responsibility to avoid ambiguity. As a result, we designed ball possession mechanisms which could not possibly possess two balls. We explicitly discarded simpler, more effective mechanisms which would be ambiguous as to whether or not multiple balls were being herded or possessed.
In this year’s game, having a 20-inch wide ball acquisition / possession mechanism which can accidentally or unintentionally possess two balls is a HUGE ADVANTAGE over an 8-inch wide ball acquisition / possession mechanism for which it would be “impossible to inadvertently or intentionally possess more than one ball at a time.”
Jumping to the end of the story, there were at least four robots in the elimination matches on Archimedes which regularly (an average of at least once per match) possessed more than one ball at a time. Some of these teams would immediately release the possessed balls when this happened. Others would go ahead and possess both balls. In the matches I watched on Archimedes (which was admittedly only a few matches), I saw many occurrences of multiple ball possession, at a significant game advantage, yet saw no penalties called in any of those cases. This gave a significant competitive advantage to the offending robots.
However, my frustration is not with the referees as much as with the game designers. It was clear to our team from our initial reading of the rules that G43 would be very hard to enforce, nearly impossible to enforce consistently, and that the specific degree of enforcement implemented would have a significant effect upon game play and match outcome. I would highly encourage the GDC to work very hard to implement games where the referees are not given the impossible task of having to try to enforce such rules.
I’ve rambled on a long time; maybe I should step off the soapbox now and let somebody else have a turn…
PS: G45 is another rule where the referees had to make such a judgment call. 469’s primary robot strategy would succeed / fail depending upon the interpretation of that rule. I believe that despite being under intense scrutiny, 469 stayed on the legal side of that line. However, their success points out that it is extremely important to teams designing robots to know where these regulatory lines are drawn, as the success / failure of a robot strategy is often highly dependent upon the exact rulings that will be made. Often times, there is a huge advantage in being able to come up right to the border of rule legality (without stepping over) as contrasted to instead choosing a design which stays well back from the border of rule legality. It is extremely frustrating to teams to make a design to come up right to the edge of the written rules (but no farther) and then to see other teams that stepped completely over the written rule, enjoying a significant competitive advantage because of their transgression.
