+1 fewer garlicky chips (6, more lumpy garlic, 5 "surplus"):
+
+P. 2x 3-garlic [4] 1x 2-garlic [1] 3x 1-garlic
+Q. 1x 3-garlic [2] 3x 2-garlic [3] 2x 1-garlic
+R. 0x 3-garlic 5x 2-garlic [5] 1x 1-garlic
+
+1 more garlicky chips (8, less lumpy garlic, 3 "surplus"):
+
+S. 1x 3-garlic [2] 1x 2-garlic [1] 6x 1-garlic
+T. 0x 3-garlic 3x 2-garlic [3] 5x 1-garlic
+
+With assumption 2 removed entirely, there are many other possiblities
+but they seem very strange.
+
+
+# Suggested Variant Garlic A
+
+Prepare, for each player, one each of X Y Z. Choose secretly and
+randomly.
+
+# Suggested Variant Garlic B
+
+Same, but include P..T. Players are entitled to count their
+chips so they will know the difference between XYZ/PQR/ST.
+
+
+II. Non-garlic
+--------------
+
+Standard load is 1 pumpkin (¤3) plus 1 green (about ¤5).
+
+Options would seem to include:
+
+ * Replace the green with a red, blue or locoweed. These are
+ available at this point and relativity uncontroversial (although
+ probably of higher value, albeit that the locoweed is not very
+ interesting).
+
+ * Replace the green with a black. This is interesting because
+
+
+
+
+III. Mechanics
+--------------
+
+# Random choice
+
+To choose secretly and randomly between identical-looking sets of
+containers (let's call them bags, supposing they're XYZ from "Garlic",
+above).
+
+A computer will be involved. It will produce two sets of secret
+output, one for player 1 and one for player 2 (WLOG there are at least
+2 players)
+
+All players collaborate to parepare bags, one for each player,
+according to each of XYZ. So now (for eg two players) we have bags:
+ X0 Y0 Z0
+ X1 Y1 Z1
+Place these in pairs in locations labelled X, Y, Z. So we ahve:
+ X: bag, bag, ...
+ Y: bag, bag, ...
+ Z: bag, bag, ...
+
+Everyone but player 1 leaves the room. Player 1's secret output
+instructs them to move bag-pairs X Y Z to new locations alpha, beta,
+gamma, according to a random permutation. Now we have:
+ alpha: bag, bag, ...
+ beta: bag, bag, ...
+ gamma: bag, bag, ...