Mare Nostrum: Empires (2016)

On the one hand, the game is pretty fun to play because it’s immersive and most of the game mechanics involved in how you gather resources, how you trade resources, and the various pathways in which you can ensure victory are nuanced in a way that’s simultaneously easy to grasp but full of strategy depth.  On the other hand, there are a couple of huge imbalances in the game that bias the result in favor of a particular kind of gameplay, and since each player has a different starting special ability, the players that specialize in that gameplay are more likely to win the game enough that it makes playing the game at all feel not worthwhile.

I feel like the best way to explain my main gripe with the game is to deal with it from a “top-down perspective” if you will – explain the fundamental exploit and then go into details as to why that’s an exploit in the first place.  This assumes that you understand the basic gameplay; if you don’t know it or need a refresher, here’s a video that explains all of the rules:


The Primary Exploit: Pyramids Win Condition

One of the four ways that a player can win the game is to “build the pyramids” which is a fancy term for “Have 12 different goods or have 12 coins”. Goods and coins are gained by having caravans in territories that you control.  Those goods are then traded amongst all of the players in the the “trading phase” of the game, spent to purchase either more game pieces on the board (that are tokens that contribute to good production or tokens that contribute to military offense or defense) or special abilities, and are purged at the end of the turn with the one exception that you can keep two coins from a previous round.

But one player (yellow) starts with the special ability of holding on to two goods from a previous round.  In both the short game and long game that’s huge because purchasing more game pieces or special abilities is based purely on diversity of goods – you purchase basic things with any 3 different goods and more powerful things with any 6 different goods.  This makes the ability to hoard two goods between rounds a significant advantage – Yellow can expand their production engine more quickly than others and thus get more diverse goods and military on the board more easily.

The three ways to stunt that advantage are to a) prevent Yellow from expanding thus limiting good production, b) tailor the Trading Phase to specifically be about preventing Yellow from getting diverse goods, and c) take over Yellow’s territory via combat.  While preventing expansion can be done without initiating combat, only two of the other players are close to Yellow’s territory and strategically focusing solely on Yellow is not a great path to success (more on that later).  Similarly, focusing too heavily on preventing Yellow from getting the resources that they need during Trade means potentially not choosing something that is optimal for your own purchase needs.

Most importantly, however is combat.  Because while combat is of the primary game mechanics, it’s almost always never worth it to actually attack someone.

The Problem with Combat

The main problem with combat has to do with one particular kind of military structure.

The rules for “attacking” versus “defending” are identical – both players roll dice according to how much military is represented in the battle and damage is dealt to both sides based on the dice roll in isolation.  For example, if Black is attacking with 3 legions and Yellow is defending with 3 legions, both players roll 3 die.  If Black rolls for a total of 12 damage points, then two of Yellow’s legions die (one legion is killed for every 5 damage points), and if Yellow rolls for 11 damage points, then two of Jennifer’s legions die.  That leaves both sides with one legion left.  Fair enough.

But there’s one type of military structure that gives defenders a strong advantage over any attack: a fortress.  Fortresses can’t be moved once they’re placed, but during a battle, a fortress automatically subtracts 5 from the attacker’s roll (so it can absorb one kill’s worth of damage) AND it automatically adds 5 to its own attack roll (so instantly kills one enemy legion).  This is such a huge odds differential that a strategy of turtling and building fortifications is almost always better than invasion, even with the Red player who has an attack advantage of +2 to every roll.  There are special abilities that can be added to potentially counter that imbalance, but by the time they can come into play it’s usually too late because Yellow has such a strong lead in production from the get-go.

I’ve only played the game three times, but in all three cases even once as Red, any time I ended up attacking it decimated my forces enough that I could not recover for the remainder of the game compared to those that opted not to attack, and I’m also pretty positive that Yellow won the game all three times either by building the pyramids or by completing their “God” set which similarly is fast-trackable based on Yellow’s special ability.  It’s possible that more gameplays with Mare Nostrum regulars could balance this out if we’re all aware of this imbalance at the beginning, but given that the game is a 4+ hour investment and its complexities, it would take a particular group of people and a particular kind of group mindset to address it quickly.

Counteracting the Exploit

The best way to counteract Yellow’s advantage is to aggressively expand in uncontested areas close to Yellow before Yellow does and then build fortresses there so that Yellow is boxed in to a limited set of resources.  The problem with that is any individual that chooses to do so isn’t fortifying in other areas and that creates exploitable weaknesses that other players can pounce on, so the choice to stunt Yellow almost has to be done under a flag of truce of all other players so it’s many against one, and that feels like a broken game mechanic.

There’s also some expansion stuff that can help balance this out by immediately obliterating fortresses, but those powers are hard to come by and require too much luck to be able to count on strategically.

Rule Changes That Could Correct the Imbalance

It’s been suggested by the group that i’ve played with that eliminating the Pyramids victory condition could help balance the game, but upon further reflection that feels like a band-aid to the issue that would mask it and make the imbalance less obvious rather than truly fix it.

Apparently in the original version of the game before the “Empires” rules update, fortresses did not have the ability to absorb damage, they were merely immobile attack units that would automatically add +5 to the damage roll.  I think i would prefer to see the opposite, where fortresses absorb damage but do not attack.

The other potential solution is to have another distinct type of military unit that is a fortress specialist, a special legion that adds +5 to their damage roll or absorbs +5 of damage when a fortress is present.

time travel chess adjustments – addendum

In rereading my previous post about an alternative ruleset to time travel chess, I realized that there’s a flaw in the following rule adjustment regarding a piece that returns to the board after jumping forward in time:

If a piece opts to jump, they have to move in a way that’s legal for the board state at the time and then place a marker on that tile.  That square is then “protected” at that move, meaning that no piece can occupy that square at the time when the piece would drop, and that square cannot be in a state where in the turn after the piece drops the opposing player can immediately capture it.  Therefore any piece on that square would be forced to move off of it as a consequence of a time traveling piece about to materialize on it.

The primary problem is that all pieces aside from pawns have the ability to move back to a spot that they previously occupied – so if the “Arrival Point” has a piece on it and a time-jumped piece is going to land on it, then a piece that moves off of it would have the ability to capture the time-jumped piece immediately after it lands.  Also, there’s a potential problem of: what if multiple pieces have the ability to attack the Arrival Point when a time-jumped piece lands?

Therefore I think there has to be three addendums/revisions to that rule:

  1. An opponent’s piece is allowed to occupy an Arrival Point at the point of the time-jump drop – that results in that piece being captured.
  2. A piece that lands on an Arrival Point gets a one-move grace period where they can’t be captured.
  3. A player may not place pieces in their own Arrival Points in a manner that would occupy Arrival Points at the moment of a time jump.

The last rule is essentially “you can’t have a piece at an Arrival Point when you’re about to time jump there”, but it reaches beyond a single instance because what if a player has pieces in more than one Arrival Point and those Arrival Points happen in consecutive moves?  That needs to be accounted for when considering what moves are legal for all moves leading up to the last Arrival Point.

Regarding the idea that “a time-jumped move means that the opponent isn’t allowed to put you in check”, a part of me feels like that could be an effective defensive strategy, but another part of me feels like that’s an exploitable gimmick – a person could opt to try to place a time jump in every future move and effectively make it so that the opponent could never put them in check.  Then again, multiple time-jumps in a row makes a player incredibly vulnerable.  Say that a queen drops in move 5 and the next four moves are time-jump drops.  That gives your opponent four free moves to capture the queen and then escape capture while pieces are being time-jumped on the board.

Regardless of that vulnerability, an easy fix to address that problem is to create some sort of limitation on time travel – the straightforward approach would be something like “you can only have five pieces per multiverse that’s time-jumping at any given point”, but maybe a better approach would be “you can only have x pieces per multiverse that’s time-jumping at any given point where x is some fraction of the number of pieces you have remaining on the board.”  Off the top of my head, I’d say time-jump power would be limited to 4 for 12-16 pieces on the board, 3 for 9-11 pieces on the board, 2 for 6-8 pieces on the board, and 1 for 7 and below.


A discussion with my brother has presented an elegant solution to some of the lingering problems of this revision: consider a time jump piece’s arrival as not being its own move.  Instead it’s considered a “delayed end to the previous move” so that the Arrival Point creates a “start state” for the player to make its true move – which could be that arrival piece or not.

That fixes the fussy “one-move grace period” mechanic that is outside of the bounds of normal chess mechanics and also addresses the issue of the restricting the opponent from being able to place the player in check because now they can.

time travel chess adjustments

I know that i haven’t really talked about my trip to London or posted any of the pics, but the reality is that a lot of what I want to say about that trip is featured in the Can’t Stop episode of my Babbling Rooks series, so writing about it feels redundant.

So instead we’re going to talk about Time Travel Chess – and some adjustments that I feel helps the game make more sense from a spacetime perspective and is also better balanced from a gameplay perspective.

The original Time Travel Chess (or at least the most popular definitive variant i’ve found) was developed by Gary Gifford in December of 2003.  There are detailed rules on chessvariants dot com, but the TL;DR is:

  1. Any piece touching the king can travel forward in time up to 10 moves.  No piece can time travel if the king is in check.  Execution of the time travel is by assigning the designated piece a future move number limited to one piece per ‘future move’.
  2. At the assigned move number, the player must place the time traveling piece on the board on any legal square.  If placing that piece would be an illegal move (because the king is in check and the only way to resolve it is to move the king), then that piece is ‘lost in time’ and can’t come back on to the board.  The king(s) can also move forward in time but if they get ‘lost in time’ that’s considered checkmate.
  3. The king can also move backward in time up to five moves.  They can only do that twice.  If they elect to do that, they choose the move that they’re moving backwards to and the board is restored to that state – with them placing the time traveling king as a (likely) second king on the board on to any legal square in addition to the one from the previous move state.  If they go backwards again, they have three kings on the board.  Only one king needs to be in checkmate for it to be considered a win.

It’s an interesting concept, but i think there are some fundamental issues both practically and philosophically.  Someone else thought so too and wrote about it, and that’s spurned me into thinking of my own ideas about it that I may attempt at some point in a Rooks episode or for fun otherwise.

Here’s a rundown of the issues from my point of view:

Issue #1: Restricting time travel to pieces touching the king changes the nature of the king too drastically.

In typical chess, the king doesn’t move around a whole lot and for good reason – it has the least flexible move mechanic and is what determines the fate of a win versus loss.  Entire game strategies are built around the mutual “i’m protecting the king and the king is protecting me” approach.  Restricting time travel around pieces that are touching the king thus feels counterintuitive because it forces the king to move around much more and then also poke holes in its own defense for the sake of time travel.  Maybe that means that the time travel mechanic would be something that people would choose to use sparingly rather than abundantly, but what i want out of a chess variant such as this would be for the mechanic to be celebrated and exploited for multiple purposes rather than just used as a ‘break glass in case of emergency’ situation.

Issue #2: Letting time traveling pieces pick any legal square on the board to drop seems too arbitrary and too powerful.

Sci-fi shows and movies tend to lump space and time travel as being things that go hand in hand (the TARDIS isn’t restricted to having to land in the same spot when it travels in time), but in the context of chess play it feels strange that a piece could ignore its normal move mechanics entirely by time traveling, especially if they only travel one move into the future.  The power of unlimited teleportation is tempered somewhat by the “has to be touching the king” (but see issue #1), “can’t time travel whist the king is in check”, and the “lost in time” risk.  But i think i would rather see the pieces still act like themselves in some way when it comes to time travel because that restriction feels more interesting.

Issue #3: A piece getting “lost in time” is a band-aid fix to the potential paradox problem.

The MathematRec guy stated this about getting lost in time:

Getting “lost in time” isn’t really a paradox, but it is untidy… consider that time teleportation from time t0 to a later time t1 might or might not violate some conservation law, i.e. it might or might not be impossible, and it might not be evident at time t0 whether conditions at t1 will allow the trip. So what happens if you try? Do you find you can’t, in which case you’ve received information from the future? Do you get “lost in time”, whatever that means? Do you teleport to a different time?…

In chess, a piece cannot space teleport (move) to a square occupied by a friendly piece. It doesn’t get “lost in space”, and it doesn’t end up on a square other than its intended destination; it simply can’t attempt the move at all. Information passes from the destination to the source to let the knight know it can’t move there, if you like. Likewise it seems to me time teleporting must be impossible if the result is an illegal position; in that case information passes from the future to the past, telling the past that the game situation in the future doesn’t allow the teleport.

In other words, a piece should never get “lost in time” because a time traveler who hops wouldn’t “feel” that sense of time – it would be instant travel from one to the next, and so from his perspective, he would never put himself in a position that would make the move illegal.  But the conditions to know whether that move is “illegal” or not in Gifford’s conception is not something that can be predicted in the normal timestream of events.

Issue #4: If the king moves backwards in time and gets inserted as a second or third king into the “new” original timeline, it creates a grandfather paradox.

Even without the rule of adding a second or third king on to the board, going back in time and making different moves from what happened in the original timeline eliminates the moment that motivated the king to travel back in time in the first place.  To ignore that paradox is to ignore the fundamental principle that our perception of time is an illusion and that time is actually a dimension in the same way that height and depth are.


So given those fundamental issues, how are they fixable in a way that still makes the game playable?

Rule Change 1: Allow any piece the ability to travel forward in time but place restrictions on its movement in space.

My initial thought about that was that a piece should be able to travel either in time *or* in space and not both, so if you were going to move forward in time, you’d have to reappear on the same board space.  That feels like a practical mechanic, but it doesn’t feel theoretically sound.  Another idea would to make the mechanic similar to Alice Chess.  If a piece opts to jump, they have to move in a way that’s legal for the board state at the time and then place a marker on that tile.  That square is then “protected” at that move, meaning that no piece can occupy that square at the time when the piece would drop, and that square cannot be in a state where in the turn after the piece drops the opposing player can immediately capture it.  Therefore any piece on that square would be forced to move off of it as a consequence of a time traveling piece about to materialize on it.

That feels sound to me because the knowledge of when a piece is going to drop does affect the strategy and the outcome around that time jump and therefore treats the time jump in a manner closer to the singular spacetime dimension instead of something separate, but it has one potential exploitable problem in that if a player is forced to make a move that would make that time drop legal, that means that they also would not be allowed to put their opponent in check.

[EDIT: I’ve addressed a big flaw in this rule change in this post.]

Rule Change 2: A king going backward in time reverts to that ‘save state’, but that state is an independent timestream from the original timestream.

Going back in time is not about changing the present circumstances but is instead about creating an alternate timeline.  MathematRec’s methodology of “you can only move on that reality if it’s your turn in that reality” is fine for the most part, but I also think in situations where a board’s been neglected by a player because their situation is unfavorable, it might be an interesting mechanic to introduce a “time rip” where the same player can then make a second move to put more pressure on the other player who is content with letting that board sit after x amount of moves.

Rule Change 2 Addendum: One of the two “backwards in time” uses can be used to fracture the timestream at its current point.

In other words, the king could elect to duplicate its current state to an alternate timeline.

all of this is pure intellectual speculation at this point – i’m not sure how the game will actually work in practice.  we’ll see what happens if i ever decide to rope someone into playing this with me.  it’d be a fun experiment in any case.