The Timeout Paradox

From BlogNomic Wiki
Revision as of 14:32, 21 November 2024 by Kevan (talk | contribs) (→‎Case studies)
Jump to navigation Jump to search
by Kevan, September 2024

Whenever a dynasty runs some kind of turn-based or round-based gameplay that requires every player to take part, there's often a question of what to do if a player fails to take their turn.

The instinctive solution is often to add a timeout - to say that if the group has been waiting X number of hours for a player to take their turn, that player is skipped and the game continues.

But in practice this can actually slow the game down.

Minimum required activity

A big thing that timeouts do is lower the minimum required activity that's expected of players.

Under a simple secret-order mechanic where every player submits an instruction to the Emperor and they all get processed at the same time once everyone has submitted, the minimum activity for each round is clear: you send in a secret order.

To imagine some thought bubbles for different player types:

Engaged player Busy player Semi-active player Inactive player
"I'll send my order quickly to keep the game moving." "I'll send my order as soon as I can." "I might send an order at some point." "I'm not going to send any orders."

Our first two players are fine, and the game will run as fast as the slowest busy player. (If most players take six hours and a busy outlier takes 72, then the group can discuss what to do about that.)

If the third player hasn't submitted orders for a round, they'll realise that everyone is waiting for them and feel some social pressure to either engage with the gameplay or (if they decide they really don't have enough time and interest after all) step back from it. The fourth player also.

If you add a 48-hour timeout to the ruleset, the minimum required activity is now: nothing, you don't have to do anything. This would change those thoughts to:

Engaged player Busy player Semi-active player Inactive player
"I'll send my order quickly to keep the game moving." "I'll send my order as soon as I can, hopefully within 48 hours." "I might send an order within 48 hours, but it's okay if I don't." "I'm not going send any orders, and that's okay."

With any of the third type playing, some rounds will take 48 hours - and other players may feel less reason to submit orders quickly, if they know that they generally have two days. With any of the fourth type present, every round will take 48 hours.

Even with just the first two types of player, the game will still slow down if any of those players regard the timeout as being a deadline: a busyish player deciding whether to take their turn at the 12 hour mark may feel that the game is allowing them another 36 hours if they need it, and that there's no rush.

Other issues

If there's a timeout mechanic, players can use it tactically like any other part of the game. If the leading player is about to time out and lose their turn, keep quiet or distract them. If the group are discussing an action plan you don't like, try to prolong or confuse their conversation so that they won't have time to agree on it. All of this slows the game down. Without a timeout, there's no reason not to remind people, even your opponents, to take their turns; this keeps the game moving.

A gameplay loop with timeouts is also less reactive to other game events. If there's a contentious proposal or CfJ that would change the gameplay and is generating a lot of engaged discussion, then under a simple gameplay loop players can hold off on taking their turns until the issue has been resolved. If one engaged player is unavoidably offline for longer than they expected, the game would still wait for them. Under a timeout, the game rolls on.

Handling inactive players

It's possible to allow players to opt out of the main game loop, so that the game knows it doesn't have to wait for them. This can be a status that players can voluntarily adopt and abandon, or one that's forced upon them by other players. (The Assassin Dynasty ran on the principle that if a player failed to submit orders for round N, then round N+1 wouldn't wait for them.)

From the minimum gameplay angle, this allows a minimum player activity level of "take no dynastic actions" without it getting in the way of a game loop that waits for players to act. And if players default to being out of the loop and have to opt-in, then players who've never intended to play the dynastic game won't slow anything down.

(The value to the group of a player who takes no dynastic actions yet casts votes on the rules for those actions is a separate question.)

Case studies

  • The Atlantean City Dynasty had players taking turns in a queue, and being removed from the queue if they failed to take their turn within 48 hours. But a couple of people weren't actually playing, and the game had to waste at least 96 hours (13% of the dynasty duration) waiting for them to time out. In retrospect that dynasty should either have required players to opt-in to the queue, or simply had no timeout.
  • The Great Machine Dynasty had a two-act structure where players spent the first act crafting complex "Agendas" for the second act. The ruleset required every player to submit five Agendas before the game could progress, rather than timing out and continuing anyway. With a clear minimum bar for expected activity, the players who did not intend to play the dynasty voluntarily idled out.
  • The Wizard Duel Dynasty was a dynasty that stuck to having no timeout (after adding one and repealing it a week later), and it worked. Players submitted secret orders to the Emperor that were only processed when everyone's were in. The fact that orders could be retracted also allowed players to hold back and see whether a particular proposal would enact, before progressing the game. The Emperor publicly reminded players on Discord if they hadn't submitted orders within a day or two.
  • The Thief Dynasty had "Haul" posts which could only be resolved if a quorum of players agreed on how to split them. The first four took less than a day each to resolve, but when the fifth was more contentious and discussion ran to 35 hours, the group proposed to have future Hauls time out at 48 hours. Some Hauls then started taking 48 hours to resolve (in part, I suspect, because players who could make no valid personal claim on a Haul had an incentive to keep quiet in the hope that it would time out).
  • The Storytelling Dynasty had an "all players submit actions privately" gameplay loop with 96-hour timeouts, which I believe broadly ran to those timeouts.
  • The Alien DNA Dynasty had an all-players-submit-orders system with no timeout, but ended with the group giving up on the game after some rounds were taking six or seven days to process. It possibly wasn't clear to the slower players (who didn't have much to say generally) that others were submitting orders within a day or two.