The idea of DCL was hatched during the Rebirth-Rangers era, mid-2013, as recognition that there were many returning players for whom D1 1v1 was the pinnacle of the Descent experience. Rangers was really the community gathering place, and the Rangers system is designed around participation, inclusion, and a focus on larger games. This is fine for what it is, it just didn't really fill the desire that players like MD-1224, Jediluke, DKH, and Djcjr had. So a small group formed and started discussing what we'd like to see in a 1v1 ladder -- beginning with Jediluke, LotharBot, and Drakona, and then seeking additional feedback from Djcjr, JinX, Blarget2, KoolBear, Knopi, DKH, and possibly some people whose names didn't get written down in any of my notes. The goal was not to compete with Rangers, but rather to complement what Rangers was doing with a system designed to meet different goals -- which is why DCL originally remained on the Rangers mumble server, by mutual agreement with DCL admins and Verran and the Rangers council.
Ladder design was partly done based on our own past ladder experience on IDL, UDL, clan ladders like UF and Wolfpack, and Case's. Some ideas came out of looking at other games' ranked systems, like Starcraft 2's ladder play (and Day9's videos about it.) We also looked back at discussion other pilots had in prior years (2011 on the DescentBB) to inform our design.
The central ideas that came out of early ladder discussion:
- We acknowledge that not all players are equally skilled or driven, and therefore that not every win or loss has the same type of meaning. So we divide the ladder into tiers, based on groups of players who are mutually competitive with each other (presumption based on past experience: averaging 13+ means that winning might be an upset, but it's not a huge upset. So we wanted all pilots in the same tier to be expected to be able to average 13+ against others in the same tier.) This gives players all over the spectrum things to play for, whether it's the top bronze spot or a giant gold streak or just getting their first win overall.
- We want as much as possible to be proven in the mines, not by who is the best at arguing or who is socially better connected or more liked. There is some need for public discussion of community standards, but tiers should come entirely from what happens in game, using scoring data from matches. We went through several rating algorithm ideas and tested them on IDL data, UDL data, and UF clan ladder data, at different time periods, and eventually built an algorithm that was pretty stable and seemed to reasonably put pilots who threatened each other into the same tier. We particularly made it a point that the algorithm only uses recent matches, so that a pilot with a long history who gets better over time is not penalized for their past, nor is a pilot who leaves and becomes rusty forced to defend a rating they no longer feel they can defend. (It's far too complex to describe in a retrospective post like this. Short tech details: it's a differentially time- and ratio- weighted graph-based simulated annealing algorithm based on minimizing logarithmic error.)
- We wanted to make sure there was common ground, specifically to avoid pilots taking high spots and then only accepting challenges in weird circumstances. Here's snippet I wrote at the time: "It sucks if you want to play for the top ladder spot, but the top player only plays D3, or only plays shakers, or only plays in a level they designed that has a ridiculous secret that allows them to fly around with 20,000 shields. It's OK to play those things if both players agree to play that, but if players can't agree on a subgame, one player should be able to insist on a 'D1 core' match. That's the one type of game that every player on the ladder must be willing to play, which means every player on the ladder can play every other player in a normal and fair circumstance.... stuff like Athena, Vamped, Manes 2, Black Rose, Three Sisters, Fuzed, and Io" (yes, that was really what I described as "core"! Not a single DKH map to be found!) There was an extended discussion about attributes of core levels, largely amounting to core levels being a reasonable size and architecture, with at least a couple different primaries per player, not excessive numbers of missiles (NYSA was OK; Athena with its 16 homers was pushing it), and no megas, cloaks, invulns, or things like shield-boost puzzles.
- Recognize cool stuff. Awards for streaks against each tier and for breaking streaks, for courage, for conquering someone else's home or defending your own, for defeating a close rival you play a bunch, for defeating a pilot who previously badly beat you, and so on. Ultimately the goal was to provide a place for players to do what they think is cool in 1v1 Descent, to set their own goals and to recognize each others' accomplishments, with players deciding what was worthy of respect and what was worth pursuing or celebrating. With it, the ladder was designed to follow the playerbase -- trophies for subgames people already were playing rather than to drive new subgames, and core levels based on what was already played rather than a new-level showcase.
There were certainly mistakes and adjustments along the way (for example, we didn't have promotion challenges until we had pilots feel like they'd been pushed into a tier they weren't ready for, at which point we created challenges rather than having purely-automated promotions and demotions.) And it wasn't for everyone; there were pilots who joined and decided they didn't like it, pilots who never joined, and pilots who stuck with Rangers or other systems. But I'd say over the DCL's seven year run, we saw a lot of awesomeness being proved in the mines, and built some great memories and even better friendships. Thanks to all who were a part of its origins, and thanks to all who made it what it was over the years!