𝐌𝐄𝐄𝐓𝐈𝐆 𝐅𝐎𝐔𝐑
Updated 06.01.24

𝟎𝟏. 𝐂𝐇𝐀𝐍𝐆𝐄𝐒 𝐓𝐎 𝐓𝐇𝐄 𝐓𝐈𝐌𝐄𝐋𝐈𝐍𝐄
Some of you may notice that some category phases are being judged faster. Not many know this, but I'm hoping to finish this award before I go abroad for studies where the first few days or weeks will be very hectic for me. Therefore, if I finish judging a phase faster, then I'll just publish the results first and adjust the timeline accordingly.

𝟎𝟐. 𝐂𝐎𝐋𝐋𝐄𝐂𝐓𝐈𝐎𝐍 𝐎𝐅 𝐏𝐑𝐈𝐙𝐄𝐒
Please reach out for your prizes within a week after the results are out as the stickers take up storage in my gallery and I don't want to keep downloading the stickers from Canva every other day. Take note that I will stop prize collection a week from the saviour of the seven seas results, which will be around 3 February or earlier. After that, I won't entertain any other collection requests as you all have been given sufficient time. 

𝐌𝐄𝐄𝐓𝐈𝐆 𝐓𝐇𝐑𝐄𝐄
Updated 13.12.24 

𝟎𝟏. 𝐀𝐃𝐃𝐈𝐓𝐈𝐎𝐍 𝐎𝐅 𝐀 𝐏𝐑𝐈𝐙𝐄
The grand winner, first place, second place and third place winners will now be featured on Instagram too! If you've got Instagram, you can give me your handle when you claim your prizes and I can tag you in the post that I'll make after the awards end.  

𝟎𝟐. 𝐂𝐇𝐀𝐍𝐆𝐄𝐒 𝐓𝐎 𝐓𝐇𝐄 𝐓𝐈𝐌𝐄𝐋𝐈𝐍𝐄
With some categories having no participants at all and some filling up faster than anticipated, I've done some reshuffling of the schedule and brought up the judging dates. The awards will now be anticipated to end in March instead of May 2025! 

𝟎𝟑. 𝐂𝐎𝐌𝐁𝐈𝐍𝐀𝐓𝐈𝐎𝐍 𝐎𝐅 𝐂𝐀𝐓𝐄𝐆𝐎𝐑𝐈𝐄𝐒
After a lot of thinking, I've decided to combine a few categories together to ensure that everyone's efforts don't go to waste and minimise categories that won't have placings due to having less than three participants. 

Best prologue and best first chapter will be combined as both categories have two participants each. 

All main genre categories with less than three participants will merge under general fiction. 

𝐌𝐄𝐄𝐓𝐈𝐍𝐆 𝐓𝐖𝐎
Updated 14.10.24

𝐅𝐀𝐍𝐃𝐎𝐌 𝐋𝐈𝐒𝐓
Here is the list of fandoms I'm familiar with and the books I've read / shows I've watched in each fandom

A Throne of Swans*
Classicals
Crazy Rich Asians
Ever After High*

Grishaverse
Lemony Snicket
Miss Peregrine*
One of Us
Hercule Poirot*
Riordanverse*
Run, Hide, Seek, Ever, After
Thea Stilton*
The Bargainer*
The Girl from the Well
The Girl Who Could Fly
The Hunger Games
The Land of Stories
The School for Good and Evil
Throne of Glass
Twisted Tales*
Wizarding World*

*didn't finish all the books or shows in the series / world

𝐌𝐄𝐄𝐓𝐈𝐍𝐆 𝐎𝐍𝐄
Updated: 14.10.24

𝐏𝐑𝐎𝐂𝐄𝐒𝐒 𝐎𝐅 𝐃𝐄𝐓𝐄𝐂𝐓𝐈𝐍𝐆 𝐀𝐈 𝐈𝐍 𝐄𝐍𝐓𝐑𝐈𝐄𝐒
If your story is suspected to have AI content, the following process will be carried out before confrontation

𝟎𝟏. 𝐓𝐇𝐄 𝐀𝐈 𝐃𝐄𝐓𝐄𝐂𝐓𝐎𝐑𝐒
QuillBot
Link: https://quillbot.com
Limit per scan: 1200/1.2k words at a time
Results: shows the percentage of likely AI-generated text with the following breakdown:
• AI-generated
• AI-generated & AI-refined
• Human-written & AI-refined
• Human-written

Scribbr
Link: https://www.scribbr.com/ai-detector/
Limit per scan: 1200/1.2k words at a time
Results: shows the percentage of likely AI-generated text with the following breakdown:
• AI-generated
• AI-generated & AI-refined

• Human-written & AI-refined
• Human-written

ZeroGPT
Link: https://www.zerogpt.com/
Limit per scan: 15 000/15k characters at a time
Results: shows the percentage of likely AI-generated text

Undetectable AI
Link: https://undetectable.ai/
Limit per scan: 15 000/15k characters at a time
Results: shows three percentages only 
• 100% human
• 50% human
• 0% human

Writer
Link: https://writer.com/ai-content-detector/
Limit per scan: 5000/5k words at a time
Results: shows the percentage of the text likely being human written

────────────────

𝟎𝟐. 𝐇𝐎𝐖 𝐓𝐇𝐄 𝐃𝐄𝐓𝐄𝐂𝐓𝐈𝐎𝐍 𝐖𝐈𝐋𝐋 𝐁𝐄 𝐂𝐀𝐑𝐑𝐈𝐄𝐒 𝐎𝐔𝐓
Excerpts of the first seven chapters of your story will be run through all the AI detectors to obtain an average percentage of your story

────────────────

𝟎𝟑. 𝐀𝐂𝐂𝐄𝐏𝐓𝐀𝐁𝐋𝐄 𝐓𝐇𝐑𝐄𝐒𝐇𝐎𝐋𝐃
Phase 1: Average percentages of each text being human-written will be calculated using the results from the five detectors
Ave % of an excerpt = (QuillBot% + Scribbr% + ZeroGPT% + Undetectable AI% + Writer%) ÷ 5

Phase 2: Average percentage of the overall story being human-written will be calculated using the average percentages of all five texts
Ave % of overall story = (text 1% + text 2% + text 3% + text 4% + text 5%) ÷ 5

According to Google, the average acceptability AI use percentage is equal to or less than 25% so anything below 75% human-written won't be accepted

────────────────

𝐃𝐈𝐒𝐂𝐋𝐀𝐈𝐌𝐄𝐑
AI detectors are not 100% accurate but the accuracy of determining whether your work has AI influence or not can be increased when cross checking with multiple detectors that shows high thresholds. For more information on this, you can read the AI info page on any of Ravendipity's awards