Refining Coelestium!

In Coelestium, players are scientists who gather and publish discoveries to affirm (or refute) claims behind the current reigning paradigm. Their goal is to propose or back competing theories to earn the most “renown”. This week, we put our game through three rounds of playtesting: one in-class playtest and two internal playtests. Each playtest builds on the last, and we’re still planning some major changes for our external playtest on Tuesday. 

In-class playtest (11/11)

Playtesters: our CS377G classmates (Michael, Rachel, and Noé)

In-class playtest

We tested our first full draft of rules. Our playtesters spun their own digital spinner to earn data tokens and cards, published discovery cards to affirm or refute claims, and proposed or supported theories about which claims were true.

Observations

  • Turns early on in the game were slow. All players spent many turns to “recruit” and upgrade their spinner, which starts at zero. 
  • There was room for improvement on assets. There was some confusion around “claim” cards vs. “discovery” cards, the cards and board was not consistent with which side was TRUE vs. FALSE, and there was an ambiguous meaning for a claim to be at zero – is it leaning positive, or negative?
  • Missing pieces: since there is no hand limit, players hoarded many cards that were unplayable. The rule book was missing an example turn to help first-time players with understanding the game.
Our game setup for in-class playtesting

Playtesters did seem to pick up on the system of science – Michael commented on “cherry-picking” data to support a claim, even if he drew discovery cards that were unsupportive of his theory. Rachel said it “mimics how researchers work”, and Noé commented on building up an “engine” so that he has more influence later on.

After discussing our observations, we made some changes to combat these problems. Players now start with data tokens (3, 4, and 5 by turn order) and one discovery card with a hand limit of three. Since making the deck took too much time, we now have two sealed envelopes per claim, which are chosen at random per claim – this way, each claim still secretly leans one way for an “objectively true” theory, but makes the setup way easier. We clarified that 0 leans towards supporting the claim. 

Internal playtest (11/13)

Playtesters: our team (Sabrina, Jeffrey, and Aiden)

Observations

  • Starting with data tokens was good. Now the first player has the option to immediately publish a theory, if they wanted to. 
  • The hand limit worked. Keeping the hand to three made playtesters play or discard unnecessary cards instead of hoarding them.
  • Some claims settled too fast. A claim was decided at a margin of 4, so two strong cards could end the debate almost immediately.
  • Spinners could grow forever. As one of our in-class playtesters had commented earlier, infinite scaling on recruiting researchers doesn’t really make sense in real life.

After some synthesizing, we tuned our game rules to attempt a more balanced game. For our next playtest, we would test a smoother card distribution and clearer “objectively true” claim per era; each claim’s winning side will have cards +/- [1, 1, 1, 2, 2, 3, 3, 3] and its losing side +/- [1, 1, 2, 3] in the deck.

Claims now settle at a difference of 5, and +/- 3 cards cost more. Discarded cards are shuffled back into the deck. Spinners are capped to reflect the real science discovery system. Moreover, we decided to test a new addition to the spinner — now, publish tokens let players bank actions and play several cards in one turn.

Our most recent digital spinner prototype, with publish tokens

Internal playtest (11/14)

Playtesters: our team (Game 1: Yuina, Jeffrey and Aiden. Game 2: Yuina, Sabrina and Aiden)

Across two internal playtests, we refined rules about theorizing. When you propose a theory, your player token goes on it as its first supporter. When other players support your theory, their player tokens go on it as second and third supporters. At the end, supporters of the new paradigm score 4, 3, and 2 points in the order they joined (hence, it’s beneficial to support a theory earlier but still beneficial to support a winning theory later, too). The first supporter also gets +2 per additional supporter, as a benefit of proposing the winning theory. This way, players can juggle 1. risking losing their theory to the winning theory, or 2. giving in to the winning theory and giving some points to their opponent, while still getting some renown for themselves, too. When a theory meets its win conditions, everyone gets one more turn before the era ends, as a last chance to overthrow the winning theory or to give in to the winning theory.

Observations

  • Most changes worked well – publish tokens and the new theorizing rules worked well in terms of balancing out points. 
  • +/- 1 cards feel weak. Players don’t have much choice in deciding which cards to play – playing +/- 1 cards is too weak, and playing into a side that we know is unfavored feels bad. This was the main frustration of playtesters.

After the 11/14 playtest, we stepped back to look at scope. We realistically have about three playtests left, and we haven’t tested era 2 at all. In a short in-class playtest, playtesters would be unlikely to have enough time to get to it. So, we’re planning to cut the game down to a single era. We’re sad to lose the two-era arc, but we think it’s the right call for the scope of this project.

With only one era, playing cards needs to feel more exciting – right now players often don’t have much of a choice: a 1-value card feels weak, and playing into the side you know is losing feels bad. We’re exploring:

  • Card effects, for example, maybe +/- 1 cards that also let you draw, so weak cards have their own role. We need to be careful not to make every card feel the same.
  • A larger, less obviously skewed deck, possibly by cutting a random chunk of a full deck instead of using our envelope setup.
  • Fixing the 0 rule so an open claim doesn’t secretly favor true.

Both of our most recent playtests were internal, so Tuesday’s external playtest will be our first real test of these changes with fresh players. We’ll be watching whether players have meaningful choices about which cards to play, and whether any single strategy takes over again.

About the author

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.