diff --git a/board.gd b/board.gd
index 9341a87..403e3f0 100644
--- a/board.gd
+++ b/board.gd
@@ -1,6 +1,9 @@
 class_name Board
 extends Node3D
 
+## squares per second
+@export_range(0.0, 1, 0.01) var anim_time := 0.15
+
 var dims: Vector2i
 
 func _ready():
@@ -75,7 +78,12 @@ func finish_tween():
 # diagonal velocity
 # uhh that's it lol?
 # well, subphases: higher lvel magnitude moves first within a phase
-# TODO: maybe show these phases with the sun
+
+## returns the total time animation *will* take
+func step_anim_time() -> float:
+	var pieces_moving := pieces().filter(func(piece): return piece.lvel != Vector2i.ZERO).size()
+	return pieces_moving * anim_time
+
 func do_step():
 	var pieces_moving := pieces().filter(func(piece): return piece.lvel != Vector2i.ZERO)
 	var pieces_cardinal := pieces_moving.filter(func(piece): return piece.lvel.x == 0 or piece.lvel.y == 0)
diff --git a/main.gd b/main.gd
index e22ffaa..6825638 100644
--- a/main.gd
+++ b/main.gd
@@ -26,11 +26,8 @@ var levels: Array[PackedScene] = [
 ]
 
 # TODO: sun movement takes exactly as much time as all piece movement
-var time := 0:
-	set(new_time):
-		var tween := get_tree().create_tween()
-		tween.tween_property($Sun, "rotation_degrees:y", -15*new_time, 0.1)
-		time = new_time
+var time := 0
+
 var advancing := false
 
 @onready var sound: AudioStreamPlayer = $Sound
@@ -118,11 +115,29 @@ func restart():
 	# scratch that i like that, don't clear the history
 	# you can like replay your steps up to a point
 
+var last_sun_tween: Tween = null
 # time advancing, basically
 func board_step():
+	var anim_time := maxf(board.step_anim_time(), board.anim_time)
 	hist.add_do_method(board.do_step)
 	hist.add_do_property(self, "time", time+1)
+	hist.add_do_method(func():
+		if last_sun_tween:
+			last_sun_tween.custom_step(413)
+			last_sun_tween.kill()
+		var tween := get_tree().create_tween()
+		tween.tween_property($Sun, "rotation_degrees:y", -15*time, anim_time)
+		last_sun_tween = tween
+	)
 	hist.add_undo_property(self, "time", time)
+	hist.add_undo_method(func():
+		if last_sun_tween:
+			last_sun_tween.custom_step(413)
+			last_sun_tween.kill()
+		var tween := get_tree().create_tween()
+		tween.tween_property($Sun, "rotation_degrees:y", -15*time, 0.1)
+		last_sun_tween = tween
+	)
 	hist.add_undo_method(board.undo_step())
 
 func step(move: Vector2i):
diff --git a/piece.gd b/piece.gd
index 4070a2e..5ced269 100644
--- a/piece.gd
+++ b/piece.gd
@@ -16,9 +16,6 @@ enum Type {
 	PlayerBarrier,
 }
 
-## squares per second
-@export_range(0.0, 1, 0.01) var anim_time := 0.1
-
 func pos_of_lpos(pos: Vector2i, y := 0.0) -> Vector3:
 	return Vector3(pos.x, y, pos.y)
 
@@ -26,7 +23,7 @@ func target_pos() -> Vector3:
 	return pos_of_lpos(lpos, position.y) + Vector3(0.5, 0, 0.5)
 
 var last_move_tween: Tween = null
-func tween_to_target(tween := get_tree().create_tween()):
+func tween_to_target(tween := get_tree().create_tween(), anim_time: float = get_parent().anim_time):
 	if last_move_tween:
 		last_move_tween.custom_step(413)
 		last_move_tween.kill()
@@ -113,7 +110,7 @@ func do_bump(move: Vector2i, tween := get_tree().create_tween(), bumpee: Piece =
 	if bumpee:
 		old_lvel_bumpee = bumpee.lvel
 	return func():
-		tween.tween_property(self, "position", target_pos() + pos_of_lpos(move)/2, anim_time/2)
+		tween.tween_property(self, "position", target_pos() + pos_of_lpos(move)/2, get_parent().anim_time/2)
 		if bumpee:
 			tween.tween_callback(func():
 				if delay_bumpee_update:
@@ -122,7 +119,7 @@ func do_bump(move: Vector2i, tween := get_tree().create_tween(), bumpee: Piece =
 					bumpee.lvel_displayed = old_lvel_bumpee
 			)
 		tween.tween_callback(func(): lvel_displayed = lvel)
-		tween_to_target(tween)
+		tween_to_target(tween, get_parent().anim_time/2)
 
 # TODO?: maybe fix? maybe the bump should be more complicated?
 func undo_bump(_move: Vector2i) -> Callable: