Rule 3c - Arena Drop-off

A discussion forum for proposed changes to the AWS rules (2014)

Moderators: administrator, BeligerAnt, petec

Forum rules
* Only one rule per thread. Any deviation will be moved by the moderators.
* Keep the discussions on-topic, relevant and polite. Anything else WILL be removed by the moderators.
* If you start a new thread (to discuss a different rule) quote the existing rule in the first post so everyone knows what you're talking about.
* The existing rules (version 4.2) can be found here: http://robotwars101.org/ants/rules.htm
Hogi
Posts: 1002
Joined: Sun Dec 30, 2012 6:47 pm
Location: basingstoke

Re: Rule 3c - Arena Drop-off

Post by Hogi » Fri Mar 14, 2014 4:55 pm

i agree with making arenas with less drop of areas but not none at all. pushers are not the only robots that would be made less effective by less drop offs. crushers are more grab-hold-and-dumpers at this scale, as are axes. without drop offs this tactic becomes useless. induction fan robots would also be effected. even if they caught their opponents. where would they take them? but my main point is: ant events are packed already. making fights longer would make it even more difficult to stick to a time schedual. i agree with peter's design. between 25 and 50 percent drop offs. that still makes things fair for pushing and holding weaponry, the rest of the perimetre low walls. this would lead to some pretty cool flipouts. however, to make an even fairer playing field. i propose a wall height just low enough for a servo flipper to be able to throw one over. obviously they'd still need to be high enough to stop spinners flying over them, ( except in extremely powerful hits) to balance things out a bit for all forms of weaponry. something for everyone that way. a straight compromise.
Daniel Jackson.

Team Hectic.

Many antweights

Super antweights: territorial.

Fleaweights: fleadom fighter, gaztons.

Featherweights: hectic (under construction)

DanTownend
Posts: 13
Joined: Wed Sep 25, 2013 9:00 am

Re: Rule 3c - Arena Drop-off

Post by DanTownend » Sat Mar 15, 2014 1:25 pm

I agree with the last few posts, somewhere between 25% and 50% drop off is good in my opinion. Mostly the same reasons, to encourage variety, and to make the fights more interesting. I also feel that the "lowest front" feature is getting repetitive, so I am happy to modify the rules of the arena for less drop-off.

Dan

User avatar
BeligerAnt
Posts: 2464
Joined: Wed May 15, 2002 12:00 am
Location: Brighton
Contact:

Re: Rule 3c - Arena Drop-off

Post by BeligerAnt » Sat Mar 15, 2014 9:44 pm

I offer no apology for the length of the following essay. I feel it is important to provide some objective evidence and alternative views for consideration.

Firstly I will explain how the current rules were arrived at and the intention behind them.
Secondly I shall present some evidence which questions the assertions presented as reasons to change the drop-off rule.
Thirdly I shall argue that the currently perceived problems are more to do with arena designs than the drop-off rule.
Finally I shall discuss potential changes to the rules and my opinions on how they may affect future competitions.

The current rules
3a) All battles must take place in a battle box, regardless of weapons being used in the battle.
4mm polycarbonate is the minimum recommended thickness.
3b) The arena will be a raised platform with an area of at least 30 inches (762mm) square.
3c) At least half of the edge of the arena must be unwalled, to allow robots to drop directly into the ditch that surrounds the arena. 50mm is the recommended height for arena walls.
3d) The minimum distance between the edge of the unwalled part of the arena and the battle box (and therefore the width of the ditch) is 140mm.
Rule 3a is for safety and I don't think anyone has any issues with this.
Rule 3b is discussed elsewhere, but it is important to note that it only specifies a minimum area – this is significant to the current discussion.
Rule 3c has always specified a minimum of 50% drop-off since antweights started. This is partly to ensure that pushers (with no active weapon) have a reasonable means of winning, but also many other types of robot rely on pushing an opponent out of the arena since with the exception of spinners most weapons do not scale down to antweight size and continue to do as much damage as their larger counterparts.
The recommendation for 50mm high walls came about several years ago. If I remember correctly the intention was that arenas would have a ditch all the way around with up to half the arena walled to 50mm and the remainder a plain drop-off. I think the only implementation of this was Tim's arena used for the AWS(s) held in Eindhoven. All other arenas have implemented around 50% of the perimeter as a full-height wall.
The intention of the recommendation was to make it possible for flippers/lifters to win at any point on the perimeter of the arena (albeit with some effort) whereas pushers could only win at the drop-off zones. It should also be possible for a vertical spinner to throw an opponent over the wall. With at least 50% drop-off all types of robot have a good chance to win by pushing an opponent out of the arena – whether by lifting, grabbing or just ramming.
Rule 3d was originally proposed to be 150mm, based on the maximum dimension of a robot (141mm) plus some tolerance. It was reduced to 140mm to accommodate Peter Waller's arena at the time. In practice, no robot is going to be able to span the ditch and get back into the arena; gravity takes over and the robot falls into the ditch, even if its back edge does catch on the edge of the arena. Also, the rule specifies a minimum width; arena builders may make the ditch wider if they wish.

Evidence for Change?
Those proposing changes to the rules often assert that current arena designs favour pushers to a significant extent, and that this is leading to pushers dominating the field since they are more likely to be successful.
If these assertions were true then pushers would dominate the results of recent AWS competitions. If pushers dominate the field and have an advantage in the arena then they should also dominate the results table.

Having checked the top three robots of every previous AWS (thanks to Oliver) at http://www.antweight.co.uk/winners.htm I have failed to find very many pushers listed at all. On the other hand, some drivers do tend to dominate the results, implying that it is driver skill and/or robot design/build that are the dominant factors. By robot design I do not mean building a simple pusher, rather building a well-thought-out design of choice very well.

There is no objective evidence to be found, but it is my observation that the assertion that a pusher can push any opponent in exactly the required direction is incorrect. It is only vaguely possible if the aggressor also manages to disable the opponent in some way, generally by lifting enough weight off its wheels that it loses traction.
If it were possible for pushers to win so easily then every fight involving a pusher would be over at the first push. Whilst this does happens sometimes, it depends very much on the design of the two robots involved, one having a distinct weakness against the other.

Problems with Current Arenas
This discussion is supposed to be about the arena drop-off rule but most of the argument seems to revolve around shortcomings in the implementation of the rule in current arena designs.

There is [b[nothing[/b] in the rules to say that the 50% drop-off has to consist of two complete sides of the arena. I have pointed this out several times in the past (see http://homepage.ntlworld.com/g0xan/robots/arena.htm) but do not feel I can push the point too far as I don't have the inclination to spend money on a polycarbonate battle box to house my arena. However, I am slightly disappointed that intelligent, creative people who build exceptional, imaginative robots all build arenas that are broadly the same, lacking in experimentation and variety.

There are plenty of ways of spreading the existing 50% drop-off around 3 or 4 walls which would prevent at least some of the perceived problems with current arena designs. It is possible to have 4 corner walls, or all open corners. There is nothing in the rules to say that the arena has to have 90 degree corners, nor that it has to be even vaguely square or rectangular.

I dislike current arena designs much like everyone else does. A sub-2-second fight isn't really much of a fight at all. However I believe this is more to do with current arena designs than the rules.

I sincerely believe that if there were more variation in arena design within the current rules we would not be having this discussion! This is not a criticism of arena builders, rather a simple observation.

The 'problem' with “the robot with the lowest front wins” will always be the case unless the arena surface were to be changed significantly. Antweights are light enough that they can run on two wheels and skid on the front edge. This means that a very low front is required to get under another robot, either to lift/flip or to force a loss of traction to enable pushing. I do not see that changing the amount of drop-off will have any effect on this.

Potential Rule Changes
Many people seem to be basing proposals on minimum drop-off on the assumption that the minimum arena size is increased significantly from the current 30” square. This should not be a foregone conclusion. 20% of the perimeter of a 30” square arena is only 24” or 600mm. Implementing this as 2 x 300mm or 3 x 200mm would, I believe, make it very difficult for any robot to win by pushing an opponent out of the arena, thus significantly favouring vertical spinners and pneumatic flippers (assuming some low walls). Even with a much larger arena, say 1m square, a total drop-off of only 800mm would seriously disadvantage pushers and any other robot relying on pushing an opponent out of the arena.

I do think there should be some statement of minimum gap and wall lengths, otherwise it is possible to make an arena with 90% drop-off that no robot could ever go out of! In practice it is very difficult to push another robot out through a 200mm gap. I really believe that reducing gaps to 200mm and the overall drop-off significantly below 50% would seriously tip the balance against pushers and other robots incapable of totally incapacitating an opponent or flipping them over the wall. As Dan (Hogi) has noted, many types of robots rely on pushing an opponent out of the arena to win, since several weapon types do not produce much damage at antweight size.

Many people have suggested that the maximum drop-off should be specified, most often at 50%. I would question the need for this added restriction. No-one is forced to build an arena with 100% drop-off. I suggest that such a rule is redundant, and if someone wants to build an arena with more than 50% drop-off let them. Variety in arena design is important (and currently missing); limiting the maximum drop-off does nothing to help.

In addition to a fear that the proposed changes will tip the balance too far in favour of powerful spinners and flippers, I worry that significantly reducing the amount of drop-off will lead to more stalemates requiring judges' decisions (often controversial) and consequently a greater reliance on judges and judging.

My preference would be to keep the rules unchanged and to see more variation in arena design within the current rules. If the majority prefer less drop-off then I would suggest:

No maximum drop-off limit (no-one is forced to implement it)
Minimum total drop-off 40% of perimeter
Minimum individual drop-off width 300mm
Minimum wall section length 75mm (may include one or more corners)
Recommended wall height 50mm (ditch on all sides of the arena)

Note that the last point remains a recommendation so existing arena designs are not made “illegal”.

A 40% drop-off allows at least 3” (75mm) of wall at each end of an open side on a square arena,
assuming two completely walled sides and two open sides.

Finally, some questions to consider:
Should the drop-off scale strictly with arena size? The current rule definitions do not specify arena size (only minimum) and the drop-off scales with the arena perimeter. Is this the best solution? Should the minimum drop-off be defined as a (total) distance, say 60” (50% of the minimum size arena)? Should it be defined as “x %, with a minimum of y mm”?
Gary, Team BeligerAnt

User avatar
joey_picus
Posts: 1137
Joined: Tue Jan 13, 2009 1:51 pm
Location: Lancaster, Lancashire
Contact:

Re: Rule 3c - Arena Drop-off

Post by joey_picus » Sat Mar 15, 2014 10:25 pm

Two potentially useless points:

1: I'd like to second Gary's point entirely, and add that I personally would consider it a moral duty of arena builders to provide us with a reasonably varied set of arenas to fight in - in an ideal world, that way we could ensure no one design could ever be favoured by the arena. As a suggestion - something I've pondered for a while but been unable to build due to being poor and useless - how would the rules look upon an arena that had all four outer walls completely enclosed, but a large central pit that served as the dropoff?

2: As much as I can agree with the desire of some people (specifically you Scott - no offence meant, I understand the e-sports analogy perfectly even if I don't believe that's what we should be emulating and you are a nice and reasonable person) to have a perfectly balanced competition, I am very concerned that the fun and essence of antweights could be in danger of being buried underneath too many rules, and the accessibility that sets us apart might be lost in the quest for a "better" competition.
Joey McConnell-Farber - Team Picus Telerobotics - http://picus.org.uk/ - @joey_picus
"These dreams go on when I close my eyes...every second of the night, I live another life"

User avatar
peterwaller
Posts: 3697
Joined: Fri Feb 15, 2002 12:00 am
Location: Aylesbury Bucks
Contact:

Re: Rule 3c - Arena Drop-off

Post by peterwaller » Sun Mar 16, 2014 9:51 am

Wow that was some post there Gary.
I must admit you make some very good arguments and the best of which I find impossible to argue against are the results.
What was most surprising although it shouldn't have been is that Andy and Scott two of the most outspoken critics of the drop off rule have between them won 9 out of the last 16 AWS's all with non pusher robots and some of the ones they didn't win were won by flippers or a spinner.
In the last AWS my flipper won 6 fights and my two pushers combined only won 7 so as you say the facts don't fit with my perception especially when you consider invertible pushers are easier to drive.
I do think it is a little inconsiderate of you Gary to destroy my heart felt if misconceived reason I thought my spinners were doing so badly :roll: .
In veiw of this I think a good compromise would be to use your suggested amendment below.
No maximum drop-off limit (no-one is forced to implement it)
Minimum total drop-off 40% of perimeter
Minimum individual drop-off width 300mm
Minimum wall section length 75mm (may include one or more corners)
Recommended wall height 50mm (ditch on all sides of the arena)

Hogi
Posts: 1002
Joined: Sun Dec 30, 2012 6:47 pm
Location: basingstoke

Re: Rule 3c - Arena Drop-off

Post by Hogi » Sun Mar 16, 2014 7:57 pm

being a fond believer that antweight combat is more about precise driving, equal competition and most of all fun, i too believe that more variety and creativity is needed in the arena department, and the excellent designing and precision engineering we've all seen and competed against in the field of the robots themselves are proof that there are people capable of building a new arena that'll completely revolutionise antweight combat. the arena at ant freeze was something different. it had corner drop offs as well as an electronicly lowered pit activated using a button inside the arena. i am personally happy enough with any arena design so long as it is designed in a way that any of the popular weapons can win in ( pushers, flippers, axes, crushers and spinners ) that brings me back to my point that antweight combat is about “ equal competition and most of all fun ” i don't see how winning a fight when you know the arena the fight was held in put your opponent at an unfair disadvantage can be fun...
Daniel Jackson.

Team Hectic.

Many antweights

Super antweights: territorial.

Fleaweights: fleadom fighter, gaztons.

Featherweights: hectic (under construction)

EpicentrE
Posts: 1079
Joined: Mon Jun 09, 2003 12:00 am
Location: Coventry
Contact:

Re: Rule 3c - Arena Drop-off

Post by EpicentrE » Tue Mar 18, 2014 11:45 am

An excellent and well thought-out post, Gary. I want to respond with one of my normal essays but I'm recovering from spending 3 days in bed due to a nasty virus so I'll keep it short instead;

I do not dispute that "pure" pushers have not had significantly more success than any other kind of robot. I may have thought that previously, but it clearly is not the case. I would however suggest that the act of pushing an opponent out of the arena is still the most popular way to win by an enormous margin, and so I still support anything that would A. give robots which are not designed to push the other robot out (such as spinners) more of a fighting chance, and B. increases the driving skill and intelligence of design needed to push someone out.

Regarding your comments on arenas, I agree completely, however arenas are expensive and time consuming to build, and so most of us probably opted for the most simple solution, which is a full wall on two sides and a ditch on the two other sides. This is the most simple design to build and probably why it has remained so prominent. Especially as we push for larger arenas, the costs only increase, and thus I'm not sure of the feasibility of suggesting people build more complex designs. I would love to see more arena builders experiment with different shapes and designs, but given the above I'm not sure whether suggesting that is a feasible solution to the perceived problem.

Having a ditch all the way around the arena (with low walls or not) would give a bigger advantage to vertical spinners and pneumatic flippers than there is currently. However since neither of these types of robots have seen a particularly overwhelming amount of success (since Anticide's reign of terror), I'd be happy to try such an arena design to see whether it did in fact unfairly benefit them.

Also, I feel it worth re-iterating that we ran at AWS39 with a larger arena with less drop-off, and as far as I can remember everyone I have spoken to has said that they preferred it, and that the fights were much more interesting and tactical. Whatever your opinions on balance, fairness, equality etc. surely that's worth taking into account?
Scott Fyfe-Jamieson, Captain of Epic Robotics. Champion of AWS38/41/42.
http://www.epicrobotics.co.uk

Hogi
Posts: 1002
Joined: Sun Dec 30, 2012 6:47 pm
Location: basingstoke

Re: Rule 3c - Arena Drop-off

Post by Hogi » Sun Apr 06, 2014 7:12 pm

the arena used for aws 41 was pretty good. it had a central section of low wall on each of the drop off sides but still more than enough drop off room to push an opponent out provided that a certain degree of luck/control/skill is implemented. my arena ( currently being constructed ) will likely have a very similar dropoff setup with 2 separate drop off areas at each end, separated by a section of low wall, only my arena will only have low wall, that should mean anyone wielding a verticle spinner or flipper ( pneumatic or servo driven) should be able to flip an opponent out rather than pushing. also considering reducing drop off even further by walling two of the drop off zones with very low wall ( 2cm max ) so that pushers can still eliminate an opponent that way however it will take a good hard shove rather than a slow doddle towards the edge. that should make battles less “ sumo like ” while still allowing a means of winning by pushing. it just means engineering skill/driving skill comes into play more as a factor in deciding whether or not the pushing attack is successful. i think this will make battles a bit more interesting as it makes way for a wider variety of individual circumstances decided by the actions and designs of the robots fighting rather than battles that revolve around shoving the opposing robot far enough in the general direction of the drop off side. my arena is going to be a smallish one so it means pushers have a good opportunity to get their opponents off, spinners have increased chance of hitting and causing damage, and flippers have a good opportunity to throw their opponent out. all robot designs have some way to win, it therefore depends on the design quality and driving skill demonstrated by the individual robot as to how far it gets in competition. hopefully we'll be able to debute my arena at this year's open in november. if by then the commitee is still undecided as to whether or not the dropoff rule needs to be changed or removed, we can use the event as an experiment with low walls and less drop off. hopefully a good experiment with something a bit different may bring a conclusion one way or the other.
Daniel Jackson.

Team Hectic.

Many antweights

Super antweights: territorial.

Fleaweights: fleadom fighter, gaztons.

Featherweights: hectic (under construction)

Hogi
Posts: 1002
Joined: Sun Dec 30, 2012 6:47 pm
Location: basingstoke

Re: Rule 3c - Arena Drop-off

Post by Hogi » Sun Jan 04, 2015 9:00 am

hi guys. i was talking to some american roboteers on the robotics community FB page a while ago about arenas. anyway, they said that some american ant arenas have dropoff sections that start the battle as walls and then open half way through the battle. it's just a suggestion but i think it might balance the effectiveness of actively damaging and pushing designs if we did that with some sections of our arena dropoff however i think it should only be some of the dropoff that starts the battle closed and opens half way through as if it was all of it, battles between pushers would be impossible to win until the dropoffs open. it's just an idea but it might be worth discussing. thanks for reading anyway :D
Daniel Jackson.

Team Hectic.

Many antweights

Super antweights: territorial.

Fleaweights: fleadom fighter, gaztons.

Featherweights: hectic (under construction)

Remote-Controlled Dave
Posts: 3716
Joined: Sun Apr 03, 2005 5:30 pm
Location: Antrim, Northern Ireland
Contact:

Re: Rule 3c - Arena Drop-off

Post by Remote-Controlled Dave » Sun Jan 04, 2015 3:45 pm

Peter experiemented with this a while ago. So has Dave W with his lowering pit. The problem is in the enginneering and cost of it. The attempts at it so far have all been simple and work to a point, but were prown to fail sometimes. To have one like that for an official AWS, it would need to work 100% every fight to make it fair. Not impossible, but a bit of a challenge.
Die Gracefully Robotics
Winner - AWS 39

Post Reply