Rationale

For the research in the SSL to be transferrable to the final full - size robots vs. human league, we need to fix the length scales in the SSL to make the AI developments in the SSL transferrable to real soccer. For a reference, here are some scales from human soccer:

Footprint of a soccer player, determined as the largest dimension of the ground-projected convex polygon of the player:

~0.5m when standing still

~1m when running

~1.5m when diving.

Let us assume a 1m projected length for the player for the subsequent calculations, since that’s the dimension when running, the most common state.

The dimensions of human soccer field have been taken from Law 1 of the FIFA 2013-2014 laws:

http://www.fifa.com/mm/document/footballdevelopment/refereeing/81/42/36/log2013en_neutral.pdf

Length of field: 100 to 110m

Ratio of field length to player: 100x to 110x

Width of Goal: 7.32m

The footprint of robots in the SSL is 0.18m. Keeping the ratio of the above field dimensions w.r.t. the players, a length-scaled field would have the following dimensions:

Length of field: 18 to 19.8m (100x to 110x 0.18m)

Width of Goal: 1.32m

Our current ratio of field length to robot size is 6.05 : 0.18 = 33.6x

We must commit to keep increasing the field size with concrete deadlines until we reach the final ratio of ~100x in order to be able to reach the 2050 goal.

Here is the proposed double-sized field layout:

Pros:

1. Ensured Existence of the SSL
   If progress in the SSL stagnates, or becomes purely incremental, the RoboCup Federation has warned that the SSL will be abolished. Exploring big changes, like the double sized field, will ensure that the SSL keeps making progress, and exists in the future.
2. Promoting more interesting gameplay
   Increasing the field size will actually help with this. Currently the field is so crowded there isn’t much room to maneuver for intelligent passes. Increasing the size of the field will promote more intelligent placement and attack coordination, instead of being blocked for space and colliding amongst other robots. Specific benefits include:

1. Robots aren’t as crowded, so they have more chance of making interesting choices about where to go on the field and then being able to execute those choices.
2. Ball has further to travel, therefore is less likely to go out of bounds, so stays in play more.
3. Without reducing the kick speed limit, the percentage of field size over which shots and passes can be made without being easily intercepted is reduced. Thus, silliness like chipping or kicking the entire length of the field should be reduced.
4. There may be room for more robots in future years.
5. It will be significantly harder to score from the own defense area. Currently you can score with your goalkeeper.
6. The larger field gives incentive to develop more complex plays. Currently most teams develop strategies that involve maybe 2-3 robots. On a large field strategies like multiple pass shots can become feasible.
7. Games are most interesting if attack and defense are well balanced. Currently a defender can move faster to the ball than an attacker can receive and control a ball. A larger field will provide better balance between attack and defense.

3. Positive effects on the Small Size league in general

1. Teams need to implement approaches that scale well. This should make the code developed in the Small Size community more transferable to other leagues.
2. The hardware requirements do not really change with moving to a large field. Nevertheless, the large field opens up a lot of interesting possibilities for AI development, such as cooperative plays between multiple robots. Teams will need to invest more time in analyzing the intent of their opponent and developing strategies that are not possible to carry out in the current setting.

Cons:

1. Availability of full fields for testing
   Not all teams will have access to a full field for testing. This gives an unfair advantage to teams with a lot of space.
   Counter-argument: The middle size league has already faced this issue successfully, since very few (if any) teams in the middle size have access to a full field, but they still participate successfully. Furthermore, the things that actually need to be tested on real robots, like one-touch passes, chip-kicks, or receiving chip kicks, do not require a full field for testing. The things that do need to be tested on the full field, like role assignments and high-level coordination are equally well testable in simulation.
   Potential resolution, additional Field Access Time
   We could explore the possibility of a separate non-competition double-sized field for open testing for all teams during RC’2014.
2. Inequality of Hardware
   Teams with access to better hardware might initially have an advantage, in terms of max speed of the robots, etc. This can be addressed by setting a reasonable max speed limit (say, 3m/s) to even the competition for the first year or so with the double-sized field. As hardware of all teams catch up, and as we start adopting mixed teams and mixed hardware in the future, these limits will later be revoked.
3. Potential Adverse Effect on Gameplay

1. Reduces the need for good quality navigation, as there’s less probability of hitting another robot with a naïve planner
2. If the ball still goes out of play a lot, the large field may actually *reduce* the amount of time the ball is in play by making the referee walk around more to position it before resuming play.
3. Larger open spaces may result in robots driving at higher speeds, possibly resulting in fewer collisions but more serious ones
4. Good lighting may be harder over the larger area.
5. The field space will be less crowded. Therefore direct shots at the goal might be more successful. This could mean less investment in interesting pass plays until defense development catches up.
6. Tracking gets more difficult. Overlap of the cameras might be in critical areas, e.g., the penalty kick spot.
7. The perceived pace of the game might be slowed down due to the larger length scales. Small Size’s main attraction lies in the fast paced games.
8. On the large field robots will need to travel larger distances. Some team’s robots might not have the stamina for that (overheating, battery capacity, etc.)
9. Currently hardware issues are holding some of us back to start solving interesting AI problems that could contribute to the main goal of RoboCup. The adoption of a significant larger field could have the same effect. Not all code scales seamlessly with the field size.