Set 17 Meta Comps

Update #4

I spent some time working with the mana lock and cast time conditions for champions. It’s complicated enough that I don’t feel I can get a highly accurate representation of the data in a short enough time frame. As such, I’ve decided to label the feature “Experimental” and will continue working on making it more accurate over time, especially as users point out specific flaws. However, it does still provide some useful information which is why I’ve decided it’s still worth shipping with the Experimental tag. The reality is that to make it accurate, I would need actual frame data for exactly how long a champion is mana locked for (ideally with a decent sample size) for every champion in the game with an ability. Then I could compare against the numbers that I have found in the time graphs and use it to figure out what values I’m looking for.

Anyway, I moved on to working on tables for augments and traits primarily. Think Coven Cashouts, Call to Chaos, A Slightly Roll, Magic Roll, etc. I’ve made strong progress on this, to the point where I have fully functioning Coven table, Booster Pack augment table, and a number of other cool tables. But there’s still some minor issues that I’m trying to solve with that and I’m also figuring out how I want to present that data visually in typical TFT Flow-style.

Yesterday, I spent quite a bit of time just refining the work I’ve already done and figuring out small kinks and bugs to make it more resilient for when Set 19 comes out. My goal is to try and get as strong of an Unreal data pipeline as possible right now so that I can spend development time during Set 19 focused on site features and an overlay rather than figuring out how to grab data.

Today I am going to continue to refine what I’ve already done. For example, I spent some hours trying to figure out the best long-term system for serializing “mods” or special items (think Void mutations from Set 16 or Sona mods from Set 17) into the URL since I tend to use that slot for other pieces of data as well such as selectors (e.g: Nova selector from Set 17, Riftbeast selector from Set 18) and armory choices like Miss Fortune’s trait in Set 17 or Kha’Zix evolutions in Set 18. In general, the reason TFT Flow is taking so much longer to update is I really want to build a pipeline that has ALL the details fully automated. I want it so that if there is ever a trait like “Eclipse” again (which activates based on a condition rather than champions on board), my team builder will update without any manual code. Or that if there are units that contribute their trait twice (e.g: Lux or Elder Dragon) or take up multiple slots (e.g: Elder Dragon), I don’t have to make exceptions in my team builder code like I have until now. The Unreal data allows the automation of so many features of the website that previously I had to manually add exceptions for because the Hextech data was so much harder to parse and also frequently omitted important data.

I’d like to finish up data tables for augments and traits and then I’m thinking about adding a “Database” portion of TFT Flow where you can navigate and find information about every champion, item, augment, trait, etc. in the game. It might seem pointless since there are already sites that provide this, but I need to maintain the data for making comps anyway so it would be a one-time job to create a frontend visual display. That might benefit the site long-term with SEO, and I also think it would just generally be cool to have.

The last wishlist item for me would be to support Riot’s localization for their tooltips, but I don’t expect I’ll get to that.

August 16th is the deadline I’m looking at for moving on from technical work to content for Set 18. That gives me 10 days to fully figure out the meta and build all the comps, which is more than enough time. I might push for earlier than that, but the 16th is my absolute “must push” my development environment to production deadline. I will post more updates.