Deep Dive: Ranked System & Matchmaking

One of the most complex issues we had on Paladins during my tenure was preserving the viability of the Ranked queue. As the game aged, the population of the mode began to suffer and cracks in its design began to show that needed addressing for the mode to continue into the future. Even before my official move to design, I was heavily involved in various attempts to bring the mode back up in relevancy and ensure a competitive space was still offered to players who sought it, ideally more. This began with a focused list of changes to relieve temporary frustrations with queue timers & match quality:

These initial changes went over well with the existing player base, especially adjustments to controller aim assist as the Ranked queue required Crossplay (a previous compromise to improve population), however these fixes only made those playing the queue enjoy it more and did little to resolve its retention issues. The data gathered from the following 2-3 months gave us two potential routes for improvement: restrict how frequently players could queue for Ranked (curate it as an Event mode) or make the mode significantly more approachable. We decided to first design and test the more intense option and share it with our Partner community, creators and community insider who allowed us to test early in development. This was the GDD I created for this vision, rolling out in phases:

When we showcased the above design, feedback was quite harsh from the community cohort we had previewed it to. Their main concerns were that the risk involved in this transition was too high and our most dedicated cohort who tended to play the queue daily would also be the most vocal against this type of shift. With that in mind, many aspects of this plan were scrapped or rolled into smaller Ranked support for the remainder of Season 7. This led us back to that earlier fork in the road, where I now took the other path and designed a much more approachable system for Season 8:

Due to tech limitations, our final implementation varied from this GDD, with Bans instead scaling in amount based on the Ranks of the players in a given lobby, with lower skill players having less ban time spent and more options to improve on. We also made more significant adjustments to our visual Rank system, improving upward mobility through the Ranks until you hit the top 25% of players in ELO. Then, the system would ramp back down to more finetuned gains to allow high-level players to eke out specific standings. In my final weeks on Paladins, we saw the average waiting time for Ranked reduced by up to 12% and the population of Ranked in the launch month was 17% higher compared to the % of DAU the prior season! This was accomplished without increasing our deviation window on what is considered a good match or increasing the # of bad matches, the later being crucial due to low population.