[robocup-small] RoboCup 2018: Tournament Format and Schedule

Tue Apr 17 11:01:58 EDT 2018

Hi teams,

the schedule for the RoboCup has been added to the official RoboCup 2018
website <http://www.robocup2018.org/?page=schedules&lang=en>. Compared to
the tentative schedule
<http://www.robocup2018.org/?page=competition-schedule&lang=en>, the venue
opens one hour earlier and closes half an hour later. Additionally, the
closing ceremony is more than one hour later than expected. This allows us
to fit in two additional games for division A for the teams that would
otherwise play only 3 games. Like last year, these additional games (called
Drop Out 1 and 2) are not a part of the play-off stage.

The on the Schedule and Results page
<http://wiki.robocup.org/Small_Size_League/RoboCup_2018/Schedule_and_Results>
has been updated accordingly.

Best,
Christian

On Mon, Apr 16, 2018 at 3:39 AM, Christian Lobmeier <pardinensis at gmail.com>
wrote:

> Hi teams,
>
> I want to add something (mainly numbers) to the tournament format for
> division A.
>
> *4 teams will only play 3 games in total*. This is definitely a downgrade
> from previous tournaments, where every team got to play 4 games in the
> group stage alone. The reason for this is that instead of 20 teams on 4
> fields, we have to play a tournament with 12 teams on only 1 field. We
> tried our best to schedule as much games as possible during the RoboCup
> (resulting in 26 games for division A), but this means that each team will
> play only 4.33 games in average. Having a play-off stage in division A is
> necessary, since it increases the competitive value of the games by a lot,
> not only for the teams but also for the spectators. The downside is that a
> play-off stage also increases the variance of the number of games per team.
> This means that even though the average number of games is above 4, the
> minimum number of games is only 3. We are aware that this means that 4
> teams will most likely not be satisfied with their number of games in this
> tournament. However, we have spent a lot of time going through different
> possible formats and we concluded that this one balances out the different
> aims better than any other.
>
> The best solution would be to increase the number of fields. We will
> definitely try again to get more space for the SSL by 2019 and with this
> year's RoboCup as a foundation for discussions, we should have a better
> chance. Nevertheless, for Montreal we have to deal with having only one
> large field.
>
> Best,
> Christian
>
> On Sun, Apr 15, 2018 at 9:05 PM, Christian Lobmeier <pardinensis at gmail.com
> > wrote:
>
>> Hi teams,
>>
>> this mail is all about the field configuration, the tournament formats,
>> the seeding and the schedule of the upcoming RoboCup.
>>
>> *Field Configuration*
>> Since there is a large amount of changes this year, the tournament will
>> also look different from past RoboCup events. We will have two separate
>> tournaments running at the same time, both with different field
>> requirements. In division A, 12 teams will play on a 9x12m field and in
>> division B, 9 teams will play on a 6x9m field. Since the space in the venue
>> is limited, we will use the following field configuration:
>>
>> - Field A (9x12m): used for official games for division A only
>> - Field B (6x9m): used for official games for division B only
>> - Field C (6x9m): used as a practice field for both divisions
>>
>> All teams can use field C for testing purposes. The OC will distribute
>> time slots for field C. If there is time left between two games on field A
>> or B, the two teams that will play the next game on these fields can use
>> this time for testing their system. However, it cannot be guaranteed that
>> the teams get testing time on field A or B, especially if there are delays
>> in the schedule.
>>
>>
>>
>> *Tournament Formats*
>> The tournament formats are tailored towards the specific requirements of
>> the two divisions.
>>
>> In division A the focus lies on having an exciting (especially for the
>> viewers) elimination phase while also guaranteeing that the best teams
>> don't eliminate each other early in the tournament. At the RoboCup 2018,
>> the 12 teams in division A will be divided into 4 groups of 3. After the
>> short group stage, all teams will advance into the double elimination
>> stage. The winner of each group will start in the upper bracket while the
>> second and third team of each group will start in the lower bracket. This
>> format guarantees that the two best teams of the SSL will face off against
>> each other in the grand final, regardless of the group stage seeding.
>>
>> In division B, the focus lies on having as much games against similarly
>> skilled teams as possible. At the RoboCup 2018, the 9 teams in division B
>> will play a 6 round swiss format. This means that after each round, the
>> teams are ranked according to their score. In the following round, teams
>> with similar score play against each other. The winner of division B is the
>> team with the highest score after round 6. This format guarantees that
>> every team in division B will play at least 5 official games.
>>
>> A more detailed explanation of the tournament formats can be found here:
>> http://wiki.robocup.org/Small_Size_League/RoboCup_2018/Sched
>> ule_and_Results
>>
>>
>>
>> *Seeding*
>> The seeding of the groups in division A as well as the first round in
>> division B will take place on June 4, 2018, exactly 14 days before the
>> start of the official games.
>>
>> For division A, I will use the ssl-grouping tool
>> <https://github.com/RoboCup-SSL/ssl-grouping>. The teams.txt that will
>> be used is attached to this mail. Please note that not only the results of
>> the RoboCup 2017 have been used, but also the results of previous RoboCups
>> as well as other tournaments (Iran Open, Asia-Pacific, ...). Like last
>> year, to get the random numbers for the seeding I will use
>> https://www.random.org/integers/?mode=advanced with the following
>> parameters:
>> - Part 1: Number of integers to generate: 2
>> - Part 1: Value range: -1,000,000,000 and 1,000,000,000
>> - Part 4: Pregenerated randomization date: 2018-06-04
>>
>> For division B, I will use https://www.random.org/sequenc
>> es/?mode=advanced to generate a random permutation with the following
>> parameters:
>> - Part 1: Smallest Value: 1
>> - Part 1: Largest Value: 9
>> - Part 4: Pregenerated randomization date: 2018-06-04
>> The numbers correspond to the participating teams of division B sorted in
>> alphabetical order, with 1 being AIS and 9 being ULtron (see
>> http://wiki.robocup.org/Small_Size_League/RoboCup_2018/Teams). The
>> website will generate 9 random numbers. The first number defines the team
>> that skips the first round. The next two numbers set the first match, the
>> next two the second match and so on.
>>
>>
>>
>> *Schedule*
>> The preliminary schedule can be found at the bottom of the "schedule and
>> results" page:
>> http://wiki.robocup.org/Small_Size_League/RoboCup_2018/Sched
>> ule_and_Results
>>
>>
>>
>>
>> If there are any questions, feel free to ask.
>>
>> Best,
>> Christian
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/robocup-small/attachments/20180417/0aebbffe/attachment.html>