Bounties are a new type of fractal competition, taking place as occasional standalone events. They require a less rigid time allocation from participants, than a classic (live) tournament format would. At the start of each event, both a T4 Fractal and a Challenge Mode boss will be drawn (based on the instabilities for the event duration) and teams will compete to get the fastest clear times on both encounters. An announcement will be made on the Fractal Bounties Discord Server and the Guild Wars 2 subreddit, after which teams will have two weeks to practice and record their best runs following the Fractal Bounties Ruleset. Once the two-week event is up, teams will have two days to submit their best runs via the Submission Page according to the submission information found below.
Once the submission deadline is up, points and leaderboard standings will be calculated and a results stream will take place on Guild MMs Twitch channel. The results will also be posted on the Fractal Bounties Discord Server, the Guild Wars 2 subreddit and the Fractal Bounties Leaderboard will also be updated with the standings, including clear times for each team alongside their submitted PoVs.
We plan to put up these bounties on a regular basis, potentially including some additional restrictions to keep the format fresh (no Mistlock Singularity or GG usage, class restrictions, low man). We hope these events may help reinvigorate the competitive fractal scene, while also encouraging new groups to form in order to play some of the best content in Guild Wars 2!
To submit a run please visit the Submission Page and fill in the form. We do not require teams to pre-register for this event, it is open to everyone! Please read through the bullet points below to check your submission will match our requirements:
If something is unclear, please contact an admin before submitting your runs, otherwise you might risk getting your submission disqualified. You can contact an admin through the Fractal Bounties Discord Server.
The final prize pool will be announced closer to the start of the first event. So far the community collected 3,20000 for the prizepool (largest chunk of 2,00000 by Guild MM). This will then be split between 1st, 2nd and 3rd place. We also want to find a way to distribute some rewards to some less experienced teams, so currently we plan to distribute the prizepool as follows:
Donations are very welcome! Drop us a message or DM in discord or alternatively contact us in game:
Please note, the final prize pool and distribution are subject to change. We are still discussing the best way to distribute it. The split above is in no way final!
Scoring for the event will take into account both the clear time of the T4 Fractal and Challenge Mode boss kill time for each team. Both times will be taken and converted into points out of 100 and added together to gain an overall standing on the leaderboard. The overall leaderboard is what the prizepool distribution will be based upon! In the event that a team fails to submit a T4 time, Challenge Mode boss time or fails to follow the ruleset or submission information, they will be given 0 points for the respective encounter.
Due to the nature of fractals and the huge variance in encounter times, we decided to use Min Max Scaling, which will mean that regardless of the timer, T4s and Challenge Mode boss kills will be worth around the same amount of points. For more information on this process, read on.
Min Max Scaling is a form of data normalization that is used to compare datasets that fall over a large range. It does this by scaling the data down to values between 0 and 1, which allows you to accurately compare them (in our case, T4 times and single boss kill times). To do this we use the formula shown below, where x is a teams kill time, min(x) is the fastest overall kill time and max(x) is the slowest overall kill time. We then multiply the scaled time by 100 to get a final score.
This method poses two issues, the first being simple to fix. If you take the encounter time directly and scale it, you will end up in a situation where a slower kill will gain you more points, to solve this we apply the formula
1 / encounter time before scaling. The second issue is a bit more complex to solve. When using Min Max Scaling, having a significantly different value from the rest of the dataset (an outlier) will skew the end result. In our case this would mean that the top end of the leaderboard might become inaccurate (the scores would move very close to each other) if someone were to e.g. submit a run of 1h.
As mentioned, any outliers have the potential to skew the data causing an inaccurate leaderboard. Unfortunately this means that any outliers will need to be removed from the scaling process and manually given a score of 0. While this isn't ideal, it is the best method to ensure an accurate overall leaderboard.
Any time that is 1.5 times the interquartile range (time between the 25th and 75th percentiles) slower than the 75th percentile (Q3) will be given a score of 0 and not be taken into account during the scaling process. Outlier removal is done before any other calculation takes place.