10

This is a spin-off motivated by Lopsy's interesting variant of Gamow's lion and zebras puzzle. It arose from a line of enquiry that tried to extend the vertical run past 5000km (or characterise the distribution of zebras to limit the run to within 5000km). That line of inquiry was eventually discarded but the spin-off puzzle seemed interesting in its own right.


Imagine a one-dimensional universe inhabited by 101 spots. Every so often, they play their own friendly variant of the 'lion and zebras' game of precision tag.

One spot is selected as the lion and the other spots (zebras) scatter to wherever they wish.

Goal At the start of the game, the lion nominates a position for each zebra. The lion wins if the zebras occupy those positions simultaneously. The zebras win if they can force the game to continue indefinitely without the lion winning.

Target Zebra One zebra is selected as the target to start the game. At any time, the lion may change the target to any zebra that is between 100m and 200m away, inclusive. For example, it may target a zebra 100m or 150m away, but not one 50m, 99m, 201m or 1km away. There is exactly one target at any given time.

Moves The spots take turns to move up to 100m per turn. On the odd turns, only the lion may move. On the even turns, only the zebra that was the target at the start of the turn may move.

Zebra Move Constraint 1: Non-crossing The target must use its turn to ensure its own position never crosses to the other side of the lion during that turn and during the lion's next turn. For example, if the lion is 50m away and the target starts moving towards it and moves for 80m, their positions cross during that move. Other than this restriction, spots may pass through each other with no consequence. Note that it is permissible for the target to end up at the same position as the lion - this constraint just prohibits crossing to the other side of the lion.

Zebra Move Constraint 2: Attractor The zebras are friendly. Within the positions allowed by the non-crossing constraint, the target will move to the position that minimises its distance to the last zebra it passed.

Can the lion win? If so, how? If not, why not?

Does it make any difference if the targeting distance was slightly modified so that the lion cannot tag zebras exactly 100m or 200m away (that is, change it from a closed interval to an open interval)?


Here is an example of placing one zebra at its target point. The game starts with lion $L$ and zebras $X,Y,Z$, with initial target $X$ and initial positions $[L,X,Y,Z] = [-150,0,-100,120]$. All other zebras are far away. The nominated position for $X$ is 140.

Use the notation $L+x$ and $L-x$ to mean $L$ moves $x$ meters towards the positive or negative end respectively. Then we play this out as follows:

$L+100,X+100,L+100$ takes $[L,X]$ to $[50,100]$.

$X$ must move to somewhere between 150 $(=L+100)$ and 200 $(=X+100)$ and therefore passes $Z$. The nearest position to $Z$ in that range is 150, so $X$ only moves $X+50$ to 150. The spots are now at $[L,X,Y,Z] = [50,150,-100,120]$ and $Z$ is the last zebra seen by $X$.

$L-10$ allows $X$ to get closer to $Z$, which it does with $X-10$, bringing $X$ to its target position 140. $L$ is now at 40, and it changes the target to $Y$, which is now in range. This freezes the position of $X$.

In this example, $Y$ and $Z$ are at convenient positions. In the general case, each zebra starts at a random position.

Lawrence
  • 7,919
  • 2
  • 22
  • 56
  • Are the target positions single points? (And if not, what's stopping the zebras from moving $\epsilon$ away from a target position whenever it would otherwise move onto that position?) – Lopsy Mar 14 '15 at 14:40
  • @Lopsy Yes, the target positions are single points. The primary tool provided for the lion here is the non-coincide restriction. One consequence is that it requires the tagged zebra to remain 100m away from the lion. Together with the 100m limit of motion per turn, this allows the lion to fine-tune the position of zebras. – Lawrence Mar 14 '15 at 21:13
  • I'm being stupid today. Suppose the tagged zebra is at Z and the lion is at L on the number line. Then, as I understand the rules, the set of legal moves for the tagged zebra is the interval [Z-100,Z+100], minus the set [L-100,L+100]. This set is either empty (which I think can never happen) or infinite. So how can the lion ever set up a situation where a zebra is forced to make one particular move? – Lopsy Mar 14 '15 at 22:59
  • @Lopsy Thanks. I had a solution in mind when posting the puzzle, but something must have gotten lost in the process. I've added a new constraint that should help the lion with its task. – Lawrence Mar 15 '15 at 10:44
  • 1
    i didnt understand tbh , what s the purpose , to be not at respective prespecified target or to be always a safe distance from lion's claws ? and can a lion be harmless when zebra passes thru or hav i not contained this part well enugh: "spots can pass through each other with no consequence." – Abr001am Mar 16 '15 at 14:52
  • @Agawa001 This is a lion-centric puzzle. The lion's goal in this version is to position the zebras at predetermined locations, given the 2 restrictions on zebra movements. There is no capture goal in this puzzle. This puzzle only has a superficial resemblance to the original L&Z puzzle. I retained the reference to L&Z because it is an abstracted form of a portion of one attempt to solve Lopsy's L&Z variant. – Lawrence Mar 16 '15 at 15:14
  • in this case zebras can win eventually , there s noway a lion can force a zebra to be at a level of a spesific spot when a zebra is free to go wherever it wants both directions. when a zebra feels close to a coming target , it changes drection immediatly , or is there other constraint iwasnt aware of ? – Abr001am Mar 16 '15 at 16:25
  • @Agawa001 The zebra is not completely free to choose - it can only move if tagged, but must stop if a different zebra is tagged. It must avoid the lion across its own and the lion's next move, but within that range of motions must select the one that moves it near the zebra it last saw. These give the lion quite a lot of control. But is it enough? – Lawrence Mar 16 '15 at 23:34
  • I have made a second change to the rules and added an example of how the lion can position one zebra. Both rule changes are highlighted at the bottom of the puzzle description. – Lawrence Mar 17 '15 at 02:04
  • At this rate, we'll need a lion-and-zebra tag. This is the 4th question I'm seeing. – ghosts_in_the_code Apr 16 '15 at 12:33
  • @ghosts_in_the_code It's an interesting field :P . But more seriously, perhaps an "infinite plane" tag would be useful. – Lawrence Apr 16 '15 at 12:42
  • Do the zebras know the positions and try to avoid them within the two constraints given to them? 2) When does "tagging" happen? Is it at the beginning, end, or during a lion's move (e.g. when zebra $X = -150$ enters into a 150-200 m distance of the lion while $L$ = 10 passes through $L = 0$ while chasing zebra $Y = +140$)? 3) What happens when a zebra has no legal move?
  • – Hackiisan Sep 26 '15 at 06:35
  • @Hackiisan In answer to your questions: 1) Yes, all positions are public knowledge. 2) Tagging happens at any time. I should really have used a different word - it means the zebra selected as the target, not the zebra that is caught. One zebra is selected at the start, and the lion can change the target (and hence the movable zebra) at any time. 3) Zebras always have at least one legal move - it can always move away from the lion (up to 100m at a time). The intention of the zebra-attractor constraint is to allow the lion a mechanism to 'pull' zebras, not just push them. – Lawrence Sep 26 '15 at 09:47