(concept) dig race ffa

[kevincentius]

• infinite dig, shared garbage map
• players who are behind will rise
• players who are ahead will descent, capped at 7 visible garbage lines to allow planning ahead

rise / descent mechanic:

• target garbage height is 9 - d*f
• d = delta, i. e. garbage cleared minus average garbage cleared by other players
• f = factor (constant), e. g. 1

Garbage shouldn't rise immediately but gradually instead:

• if < 7 garbage lines, rise speed is fast and constant
• if >= 7 garbage lines, rise speed = min(riseSpeedCap, (targetVisibleGarbageLines - currentVisibleGarbageLines) * multiplierConstant)

This means the further behind someone is, the harder it is to keep the garbage lower than the target height. When too far behind, the target height is way beyond the field's height. This means no matter how far behind, one can survive indefinitely if digging very fast (should be unsustainable).

Scenario thought experiments

• 3 players: 1 player is falling behind
• the slow player eventually dies
• the other 2 players both have low garbage lines
• now what? Do we include the dead player in the average digCount?
• scenario 1: no - then players will get a spike in rise speed when players slower than them dies
• scenario 2: yes - then everyone will stay low and game doesn't end.
• scenario 3: yes but after dying the dead player's digCount increases at average of other player's speed (e.g. 5 players remaining --> each line clear by anyone adds 0.2 to the dead player's digCount). Hmm what will be the consequence of this rule?

Scenario 3 thought experiment 1: say there's 3 players, 1 slow and dies. 1 fast and should eventually win.

• When the slow player dies, his digCount will catch up with the middle player eventually. Besides, the fast player will get further ahead anyway. If P1 is 10% faster than P2, then every 10 lines cleared by P2, P1 clears 11. So P3's digCount will increase by 10.5. This means the average will increase by 10.5, P2 falls behind by 0.5 line. instead of 1! In other words, P1's progress is watered down by dead players. Not a good solution.

Hourglass problem? Average should be constant but when a player dies with high garbage count the average decreases. Maybe solution is just to re-add garbage target height to everyone while the average is still below the initial value. This still causes a spike when slower players die though.

But maybe that's not a problem. It can be abused e. g. wait for 3rd player to die then spike clear many lines at once. But it is generally the case for FFA anyways.

So mechanic now is:

• each player starts with a H = hInit (H = target garbage height)
• clearing a line transfers F <= 1 of the H to all other players (averaged out)
• each frame H of everyone increases by hRebound if average H of living players is below hInit

And rising mechanic for each player:

• if there are less than hMin garbageLines then garbage rises at the speed of riseSpeedReplacement
• otherwise, rises at the speed of min(riseSpeedCap, (H - currentVisibleGarbageLines) * riseSpeedFactor)

[kevincentius]

So mechanic now is:

• each player starts with a H = hInit (H = target garbage height)
• clearing a line transfers F <= 1 of the H to all other players (averaged out)
• each frame H of everyone increases by hRebound if average H of living players is below hInit

And rising mechanic for each player:

• if there are less than hMin garbageLines then garbage rises at the speed of riseSpeedReplacement
• otherwise, rises at the speed of min(riseSpeedCap, (H - currentVisibleGarbageLines) * riseSpeedFactor)