Celebrated among players at the time, but apparently largely unnoticed by game developers, Zach Gage invented a very special game mode for Pocket-Run Pool in 2018.
The so-called “Insta-Tournaments” build on the established concept of “Daily Challenges”, where all players of a single-player game are given one attempt a day to play the daily “seed” — a run featuring the same random generation for all players — to be able to compare their results. At the end of the day, the leaderboard for the previous day is saved and the next run is generated.
Similarly, in Pocket-Run Pool players try to clear the pool table after a random opening. But with his instant tournaments, Gage pushed the concept of seeds even further. Players didn’t just get one chance to compare their scores on a daily basis, but — depending on overall activity — basically whenever they wanted to. Every couple of minutes, a new tournament was generated for small groups of sometimes just a few players.
A Different Kind of Dynamic Difficulty
One might think that this conceptual twist would not only increase the frequency of feedback, but also open up new design possibilities.
However, it took a year and a half for the idea to be picked up again by Arnold Rauers in Maze Machina and its “Challenge” mode. This time around, it’s not about generating billiard tables but a series of tactically demanding, tile-based roguelike levels, but the fundamental idea is the same.
However, the mode introduces a ranking system on top and thereby opens up a meta layer above the tournaments themselves. For each tournament, players gain or lose ranking points depending on how well they did. Based on their overall rank, they are then placed on a global leaderboard. At first sight, this isn’t a terribly complex structure, but it hints are the potential of the concept.
For example, instant tournaments could be a valid alternative to single-player matchmaking, which usually gets quite complicated in terms of design and the details of its implementation. Dividing players into tournament groups not only at certain time intervals but also based on their current meta rank, will challenge them according to their abilities and ensures that their skill level is accurately reflected in the tournament table.
Advantages: Accuracy & Granularity
A significant advantage of this approach: The level of difficulty not only adapts to a player’s rank, but also to the variance of the randomly generated setup. Generating an “outlier match” every now and then is perfectly fine, since the other tournament participants are facing the exact same situation. For example, if a particular run is very difficult, many players will not do too well and thus comparatively “weak” results will suffice to top the rankings.
Matchmaking using a pure single-player Elo system has its weaknesses in this area in particular. Basically the game would have to recognize said difficulty fluctuations between the matches and move its goal posts accordingly. This is such a complex problem — especially for “just” a meta-game feature — that it’s usually not even being tried. Now, restricting the random generator to only ever create fair challenges is an option of course, but also results in less varied gameplay overall. Quite the balancing act.
Another advantage of instant tournaments lies in their granularity. Of course, they are somewhat dependent on an active player base. If one exists however, players can essentially be grouped together by any skill level, which means the game can account for the tiniest differences in skill. This results in much more exact levels of difficulty than those you would find on a rather rigid single-player ladder, where ranks do not only have to be accurately predicted, but also implemented manually.
A High-Potential Shortcut
In a way, “Insta-tournaments” can therefore be considered a design shortcut: single-player matchmaking based on multiplayer data. The unbeatable swarm intelligence of the player base provides the input to find the proper challenge for every single player. Clever!
A possible downside, once again, lies in the act of qualitatively sorting people into tables. Perhaps, one could add a layer of abstraction and disconnect the collected data from actual human players. For any individual player, things would then simply look like a “normal” solitaire game. On the other hand, a competition can hardly be framed more neutrally and indirectly than in the examples mentioned above. Fortunately, we’re far from the “megakill” here.
In any case, it would be nice if it didn’t take years again for game designers to continue their experiments in this direction. There’s quite a bit of untapped potential here.